concurring in part and dissenting in part:
I agree with the district court that the Duval County School Board (“the Board”) was entitled to a declaration of unitary status on the majority of the Green factors. See Green v. Sch. Bd. of New Kent County, 391 U.S. 430, 88 S.Ct. 1689, 20 L.Ed.2d 716 (1968). However, I do not believe that the Board was entitled to a declaration of unitary status in the area of student assignment, the root of the Duval County’s (the “District’s”) prior unconstitutional de jure segregated school system. Certainly, I too would like to say that the Board finally “got it right,” and end the forty one years of judicial supervision in this case. However, until the Supreme Court simply announces that the time has come to end judicial oversight over school districts, regardless of their status, we must faithfully apply the unitary status standards when determining whether judicial supervision should be terminated. In this case, the Board was required to establish that it made a good faith effort to desegregate the black core city schools created under de jure segregation in order to merit an award of unitary status in the area of student assignment. It failed to do so and, therefore, we are required to retain judicial supervision over this area.
As the majority explains, when a school board that is subject to a desegregation order seeks to bring an end to judicial supervision over the school district it must satisfy the two unitary status requirements: first it must establish that it has complied in good faith with the controlling desegregation order, and second, it must demonstrate that it has eliminated the vestiges of de jure segregation to the “extent practicable.” See Bd. of Educ. of Oklahoma City Pub. Sch. v. Dowell, 498 U.S. 237, 249-50, 111 S.Ct. 630, 112 L.Ed.2d 715 (1991); Lockett v. Bd. of Educ. Muscogee County Sch. Dist., 111 F.3d 839, 842 (11th Cir.1997); Lee v. Etowah County Bd. of Educ., 963 F.2d 1416, 1425 (11th Cir.1992). In this case, the NAACP and the Board agreed that the Board’s desegregation obligations would be outlined in a negotiated settlement agreement, the Corrected Stipulation and Agreement (“CSA”), instead of a court order. However, before the district court approved the agreement, it had to be satisfied that the CSA, at a bare minimum, adequately addressed the Board’s constitutional obligation to desegregate the District’s formerly de jure segregated schools. The district court concluded that the CSA addressed the Board’s constitutional obligations in the area of student assignment by setting target racial enrollment goals for various school groups in the District. However, the parties are fundamentally at odds over how many and which schools were required to reach the CSA’s student enrollment goals in order to demonstrate that the Board substantially complied with the CSA’s student enrollment provisions.
The majority resolves the compliance disputes in this case in favor of the Board and rules that the Board’s magnet program, which achieved a 65% compliance rate district-wide with the CSA’s student enrollment goals, was sufficient to meet the Board’s obligations. It also accepts the Board’s argument that the current 96% segregation rate in the District’s Attachment C former de jure black schools is the result of a new pattern of white flight,1 *978as shown by demographic evidence revealing a 12% relative shift in the ratio of black to white students in the District over the twenty seven years prior to the CSA’s creation. I cannot agree.
I believe that the CSA was crafted to address the core constitutional injury in this case: the high rates of racial segregation in the District’s formerly de jure black core city schools. The record shows that these schools had over 90% black enrollments at the end of de jure segregation, and have remained 90% black during the course of the CSA’s enforcement term. Aso, the record plainly shows that these schools are not segregated as a result of a new pattern of white flight. The Board’s demographic evidence shows that the white flight it complains of preceded the creation of the CSA and, moreover, was adequately accounted for in the CSA’s less stringent student enrollment goals.2 The white flight that did occur during the CSA’s enforcement term, a 4% decrease in the number of white students district-wide, had little or no effect on the Board’s ability to desegregate the core city schools, as these schools are located in black neighborhoods that have been over 90% black since the end of de jure segregation.
Athough the majority ultimately concedes the importance of the Attachment C schools, it excuses the Board’s dismal performance at these schools by showing that the Board met the CSA’s standards for desegregating these schools prior to the CSA’s creation. However, in a proper unitary status inquiry, the court must apply the relevant desegregation order to the period it was designed to address. The proper inquiry in this case is whether the Board met its obligations under the CSA during the CSA enforcement period, which began after 1991. Under this inquiry, the record shows that the Board did not eliminate the vestiges of segregation at its formerly de jure black schools during the CSA enforcement term, and it failed to provide a plausible justification for its failure to do so. Therefore, it was not entitled to a declaration of unitary status in the area of student assignment. Furthermore, by applying the CSA standards to *979assess the Board’s performance prior to the period the CSA was created, the majority runs afoul of our precedent in this case, and improperly lowers the Board’s compliance standard below what was required by the orders governing the pre-CSA period.
A Identifying The Proper Compliance Standard For The Student Enrollment Provisions
As the majority explains, the parties’ disputes about the implementation of the CSA must be resolved under a contract analysis. Therefore, the court must first refer to the CSA’s terms to discern the Board’s desegregation obligations. Jacksonville Branch, NAACP v. Duval County Sch. Bd., 978 F.2d 1574, 1578 (11th Cir.1992) (“NAACP II ”) (citing United States v. Armour & Co., 402 U.S. 678, 91 S.Ct. 1752, 29 L.Ed.2d 256 (1971)). As with all contract disputes, the obligations of the CSA must be discerned “within its four corners, and not by reference to what might satisfy the purposes of one of the parties to it.” NAACP II, 978 F.2d at 1578. Also, the court’s interpretation must give full effect to both the individual obligations outlined in the agreement, and their collective or overall purpose. See Restatement (Second) of Contracts § 203(a) (explaining that in interpreting “a promise or agreement or a term thereof ... an interpretation which gives a reasonable, lawful, and effective meaning to all the terms is preferred to an interpretation which leaves a part unreasonable, unlawful, or of no effect.”). However, when the CSA is silent or ambiguous as to the meaning of a particular term, its meaning must be interpreted based on the constitutional standards that inform all desegregation cases, because the CSA would have been invalid had it allowed the Board to escape its desegregation obligations under the Constitution.
1. The Compliance Standard Suggested By The CSA’s Terms
As explained above, the CSA’s student assignment provisions do not explicitly indicate how to measure compliance under the agreement, apart from providing general guidelines on the enrollment goals for particular school groups. Consequently, the Board argues that we should measure its compliance with the CSA’s student assignment provisions by examining its performance district-wide, that is, with regard to all of the schools in the District. Using this measure, the Board argues that it is entitled to a finding that it “complied in good faith” with the CSA because it achieved a 65% success rate district-wide with the CSA’s student enrollment goals.3 The NAACP argues that the CSA requires the Board’s compliance to be assessed based on its ability to desegregate the Attachment C schools — a subset of 28 schools specially highlighted in the CSA as schools that were “expected to become” racially identifiable black schools at the end of the Mims Injunction forced busing program. (CSA, Attachment C).4 Using *980this measure, the Board’s performance under the CSA was a dismal failure because, in 1998, when the district court granted unitary status in this case, 96% of the Attachment C schools were not in compliance with the CSA’s student enrollment goals.5
In my view, the CSA clearly indicates that the parties’ primary focus was on the Attachment C schools, and any measure of the Board’s success must therefore prioritize this school group. Indeed, the majority concedes this point, as it recognizes that the CSA’s main goal was to desegregate the Attachment C schools. (Maj. Op. at 763-64). See also Dist. Ct. Op. at 11 (recognizing same). This conclusion is evidenced by the fact that the vast majority of the CSA’s student assignment provisions are devoted to this purpose. Specifically, the CSA divided the District into school groups that were rated based on their need for further desegregation efforts. CSA ¶¶ 3, 5-14.6 The CSA then identified three sets of elementary schools: “Attachment B” or “stand alone schools,” which were racially integrated elementary schools that were not in need of immediate assistance, CSA ¶ 6; “Attachment C” or “racially identifiable black schools;” and “Attachment D,” or “racially identifiable white schools,” both of which were in need of immediate desegregation assistance. CSA ¶ 7. The Attachment C and D schools had, or were expected to develop, racially imbalanced enrollments, which the CSA defined for Attachment C schools as a student body composed of 75% or more black students, and defined for Attachment D schools as a student enrollment of 85% or more white students. (CSA, Attachments C & D) (Dist. Ct. Op. at 11). The CSA expressly established as its primary goal that the Board would enroll 20%-55% black students at the Attachment C and D schools, and would maintain this enrollment over a three-year period. CSA ¶¶ 4 & 31.7
Even between the Attachment C and the Attachment D schools, the parties appeared to recognize that the predominately white schools in Attachment D would have *981less difficulty meeting the CSA standards, i.e., less difficulty drawing black students,8 because there are only two CSA provisions specifically directed to the Attachment D schools, ¶¶ 18 & 19, and the more detailed of these two provisions is designed to ensure that the Attachment D schools would not continue to draw white students away from the Attachment C schools. ¶ 19(a)-(c). In contrast, the CSA contemplates greater difficulty in meeting the student enrollment goals at the Attachment C schools, as the agreement contains numerous and detailed planning provisions designed to attract white students to the Attachment C schools. See, e.g., CSA ¶¶ 11-12, (establishing a time line for magnet program proposals for Attachment C schools, and identifying recruitment and funding efforts for Attachment C schools); ¶¶ 15-17 (discussing the role facilities renovation and replacement would play in diversifying the enrollment at the Attachment C schools).
Since the bulk of the CSA’s desegregation efforts focus on the Attachment C schools, the schools in the greatest need, it would ignore the CSA’s purpose to measure the Board’s compliance with the CSA based on district-wide figures. To do so would credit the Board for success in desegregating the “stand alone” Attachment B schools, which as of 1990 did not require help, and “majority white” Attachment D schools, which the Board anticipated would have less difficulty meeting the CSA standards.
2. The Compliance Standard Required By the Board’s Constitutional Obligations
Even if the CSA could be interpreted to assess the Board’s compliance with the student enrollment goals based on the Board’s district-wide performance, the constitutional injury in this case still requires that we focus on the Attachment C schools. A review of the first comprehensive remedial order entered in this case shows that the court stressed the importance of desegregating the eighteen “core city” historically black schools, the majority of which became Attachment C schools under the CSA.9 Mims v. Duval County Sch. Bd., 329 F.Supp. 123 (M.D.Fla.1971). The Mims Court explained, “[n]o one argues that the 18 black core city schools are anything but the vestiges of a de jure segregated dual school system. Under the circumstances ... [Swann v.] Charlotte-Mecklenburg [402 U.S. 1, 91 S.Ct. 1267, 28 L.Ed.2d 554 (1971) ] literally commands the integration of these core-city students with the outlying white students through busing.” Mims, 329 F.Supp. at 130. The Board now asks us to rule in a manner that functionally negates the importance of these core city formerly de jure black schools. However, as we explained in 1971, these schools are at the heart of the constitutional violation in this case, and we cannot simply ignore them. Therefore, without some meaningful showing of success with respect to the core city schools, *982the Board cannot satisfy its constitutional obligations.
B. Identifying the Proper CSA Enforcement Period
1. The Enforcement Period Required By The CSA’s Terms
Two things are clear from the record in this case: (1) the Attachment C schools were the focus of the CSA and are a central part of the constitutional violation recognized in this case and (2) during the CSA enforcement period (1991-1998) the Board failed to desegregate the Attachment C schools. To negate these facts, the Board claims that it satisfied the CSA’s enrollment goals for the Attachment C schools in the years prior to the CSA’s creation. Specifically, the Board cites enrollment statistics from 1971 to 1990, during the Mims Injunction’s forced busing program, and argues that the Attachment C schools were desegregated during this period.10 In support of this claim, it notes that the number of black majority schools fell from 21% to 15% during the period covered by the Mims Injunction. The majority concludes that this evidence is sufficient to show that the Board eliminated the “vestiges of de jure segregation” in the Attachment C schools for a reasonable period, and therefore met its burden. (Maj. Op. at 968-69).
I have found no case, and not surprisingly the Board points to none, in which a court assessed a party’s compliance with an agreement based on a party’s behavior prior to the agreement’s creation. Also, in this case, the plain terms of the CSA show that there is no defensible basis for taking the CSA’s terms and interpreting them to cover a period prior to the time the agreement was formed. The agreement clearly states that it was intended to govern the period after 1991. See CSA ¶4 (“[T]he parties agree that, commencing with the 1991-1992 school year, each school [in Zones I — V] shall have as its desegrega-tive goal the enrollment of at least 20% black students and 45% white students.”).
Furthermore, I find this interpretive maneuver illogical. The majority implicitly concludes that the parties entered into an agreement in 1991 to achieve goals that they had already reached prior to that period. But why would they do that? The only fair reading of the CSA’s terms is to treat the agreement as a guide for the Board’s desegregation efforts and goals subsequent to 1991, as a substitute for the Mims Injunction forced busing program. Therefore, the Board’s performance prior to this period is not relevant to this dispute under a contract inquiry.
2. The Enforcement Period Required Under The Constitutional Framework
In addition to running afoul of the CSA’s plain terms, the majority’s application of the CSA standards to the period between 1971-1990 ignores the earlier precedent in *983this case. All of the earlier decisions covering this time period establish that the Board’s desegregation obligations after 1971 were governed by the Mims Injunction standards, up and until that order was superceded in 1991 by the CSA. Mims, 329 F.Supp. 123; Jacksonville Branch NAACP v. Duvall County Sch. Bd., 883 F.2d 945, 953 (11th Cir.1989) (“NAACP 7”). This decision effectively overrules the standards the Mims Court established to assess the Board’s desegregation efforts between 1971 and 1991 and, more importantly, sub silentio reduces the Board’s obligations during this period. Specifically, the majority asserts the that Board met its desegregation obligations between 1971 and 1990 by measuring the Board’s performance at the Attachment C schools relative to the 20%-55% black student enrollment goal provided for in the CSA, CSA ¶ 6, instead of using the 21%-34% goal for elementary schools established for this time period under Mims. Mims, 329 F.Supp. at 130.11 The effect of the majority’s analysis is to lessen the Board’s desegregation obligations between 1971 and 1990 in an attempt to show that the Board satisfied its constitutional obligations during this period. However, we have already held that the Board did not meet the Mims Injunction enrollment goals prior to 198612 and we denied the Board unitary status on this basis. NAACP I, 883 F.2d at 953. More important to this case, the Board did not meet the Mims enrollment goals (21%-34%) under the remaining portion of the Mims enforcement period (1986-1991), and it failed to meet the CSA enrollment goals (20%-55%) once the CSA became the governing enforcement standard (1991-1998). These facts are made clear in the chart below.13
*984[[Image here]]
In summary, there is no basis for rejecting the Mims Court’s standards and conclusions and, therefore, no basis for revisiting and reinterpreting the Board’s performance during the Mims Injunction period under the CSA standard. The only relevant inquiry in this case is whether the Board successfully desegregated the core city black schools in compliance with the CSA’s student enrollment goals, which required that it maintain a black student enrollment between 20% and 55%, at the Attachment C schools during the CSA enforcement period, from 1991 to 1998. The Board failed to do so, and therefore, it has not met its constitutional obligations in the area of student assignment.
C. Evidence of Good Faith Compliance During the CSA Period
The Board also argues that even if it did not meet the CSA’s student enrollment goals at a sufficient number of schools, its failure can be forgiven given its good faith effort to implement the CSA’s desegregation techniques. See Morgan v. Nucci, 831 F.2d 313, 320 (1st Cir.1987) (recognizing that a school board had discharged its desegregation obligations when it did not entirely achieve its student enrollment goals because of the school board’s good faith efforts in implementing its desegregation plan). In support of this claim, the Board argues that the CSA only required that it establish and fund a network of magnet programs at the Attachment C and D schools and, since it opened these programs and funded them, it met its obligations under the agreement.
The NAACP argues that the Board did not make a good faith effort to utilize the CSA’s desegregation strategies because the agreement required the Board to implement one or more of the supplemental desegregation techniques outlined in the agreement that were designed to trigger parents to make a desegregative magnet school choice for their children. Specifically, the NAACP argues that the Board knew that a bare magnet program would fail to desegregate the racially identifiable black schools because, unless the program included some provision that directed parents towards a desegregative school choice for their children, parents would not *985choose magnet programs in racially identifiable black schools. Recognizing this difficulty, the CSA provided for a variety of means to encourage parents to make a desegregative school choice for their children, including re-drawing attendance zones, capping enrollment at Attachment D schools, and/or limiting curricular offerings at Attachment D schools, implementing a majority-to-minority transfer program and, as a last resort, re-instituting mandatory reassignment. CSA ¶¶ 19(a)-(c) & 15(a)-(b). The NAACP argues that because the Board refused to consider or fairly implement the supplementary desegregation techniques, it did not fulfill its obligations under the CSA.
The majority adopts the Board’s view, as it reads the CSA to require only the establishment of a “free choice” student assignment program between a network of magnet schools. (Maj. Op. at 967-68). The majority also concludes that, although it was not required to do so, the Board made a good faith effort to institute some of the supplementary desegregation techniques, explaining that the Board adopted a majority-to-minority transfer program, redrew attendance boundaries, and capped enrollment at some of the Attachment D schools. (Id.) I do not believe the record supports either of the majority’s assertions in this regard, and I address each issue in turn.
First, I cannot endorse the majority’s restrictive reading of the CSA. We are required to prefer a reading of a contract that “gives a reasonable, lawful, and effective meaning to all the terms ... [over] an interpretation that leaves a part unreasonable, unlawful, or of no effect.” See Restatement (Second) of Contracts § 203(a). The majority’s reading of the contract runs afoul of this principle, as it makes surplus-age of over three-quarters of the CSA, a 25-page contract containing 33 paragraphs that, when it includes the supplemental desegregation techniques, could be read as incorporating up to several hundred individual provisions targeted at achieving desegregation in the District’s 144 schools. By treating the supplemental desegregation methods in the CSA as mere precato-ry references with no binding effect, the majority effectively nullifies most of the CSA’s terms.14
A more reasonable interpretation of the CSA, one that gives effect to all of its terms, is that it was intended to function either as a “best-efforts” contract,15 or an *986“alternative contract.” Black’s Law Dictionary at 319 (1999). A best-efforts contract is “a contract in which a party undertakes to use best efforts to fulfill the promises made ... [and] the adequacy of [that] party’s performance is measured by the party’s ability to fulfill the specified obligations.” Id. An “alternative contract” is “a contract that provides more than one way for a party to complete performance, ... [and usually] permits that party to choose the manner of performance.” Id. Both of these types of agreements allow a party who is obligated to achieve a particular goal under a contract limited discretion to choose what strategy it will use to achieve that goal; however, the agreements also require that, if a party fails to meet a contract goal, and it does not utilize the strategies identified under the contract, it must present valid reasons for rejecting the contract’s suggested strategies. See 17A Am.Jur.2d Contracts § 682 (discussing alternative contracts); 6 Cor-bin On Contracts § 1330 (same); Restatement Contracts § 469 (1921) (same). Under this reading of the agreement, the Board was not necessarily required to pursue all of the supplementary desegregation techniques in the CSA, but it was required to present evidence that it fairly considered adding the supplemental techniques in its effort to desegregate the Attachment C schools, and rejected them because they were impractical.
In my view, the record unequivocally shows that the Board knew that its purely voluntary program was failing to reach the CSA’s student enrollment goals in the Attachment C schools, but it failed to add any of the strategies recommended in the CSA as a more aggressive means to guide parents toward a desegregative school choice for their children. Specifically, the record shows that the Board knew as early as 1992 that the free-standing magnet program was failing to desegregate the racially identifiable black schools. (Tr. Vol. 10 at 162). The record also shows that, Dr. Zenke, the Board’s desegregation advisor, indicated that he was aware during the implementation of the plan that when other districts had used voluntary magnet programs to desegregate their schools, they had included methods to guide parents towards desegregative choices, including capping, balloting measures, and redrawing school attendance zones. Dr. Zenke testified that although he conferred individually with the Board members about the District’s problems in meeting the CSA’s student enrollment goals, he never brought for a full Board vote any CSA desegregation strategy that would be perceived as constraining parents’ choices in selecting schools for their children. Dr. Zenke explained that he made this decision because he did not support the adoption of these strategies and, based on his conversations with the Board, he did not believe that the majority of the Board would ever adopt any of them. (Id. at 161-162, 174). In Dr. Zenke’s own words, the Board’s desegregation effort was “entirely based on voluntary options and choices” and it refused to consider any desegregation strategy that it interpreted as limiting parents’ choices to send their children to a particular school. (Id. at 174).
Consistent with this approach, the Board persistently refused to implement any CSA mechanism that would guide parents’ school choices. It continued with this purely voluntary effort despite substantial evidence that the voluntary program was not achieving the CSA’s goals. Also, the record shows that the Board’s efforts even undermined its purely “voluntary approach” to desegregating the schools. Specifically, the Board held a magnet schools fair, and it sent out a letter that guaranteed parents transportation for their children if they agreed to transfer their children from a school where they *987were in the majority race group to a school where they were in the minority (a “majority-to-minority” school transfer). However, although the magnet fair and the transfer letters were supposed to encourage parents to make a desegregative school choice for their children, the Board assured parents that if they did not want their children to participate in the magnet program or the majority-to-minority transfer initiative, they could simply decline to participate in these programs, and by default end up in a neighborhood school. (Id. at 90). Therefore, it is not surprising that out of the 36,000 letters the Board sent to parents, only 332 families elected to have their children transferred. Parents knew that they were permitted, with Board approval, to avoid a desegregative school assignment for their children simply by declining to respond to the Board’s letters. The Board’s actions were inconsistent with the commitment it made under the CSA. CSA ¶ 19.
Certainly, the CSA was not designed to hamstring the Board into pursuing unworkable solutions in order to achieve the CSA’s goals. However, the record evidence shows that the Board knew that the bare magnet program was failing, and it knew why, but it did not consider in good faith any of the CSA’s options for improving the program, and it did not attempt to identify any other means to achieve the CSA’s student enrollment goals.
As an additional matter, our inquiry into good faith must take into account the Board’s efforts apart from its unitary status obligations; that is, if the Board’s other administrative decisions seem recklessly unmindful of the problems of segregation in the District, such behavior is inconsistent with a finding of good faith. In this case, the Board sought and received federal grant money to improve the Attachment C schools during the period covered by the CSA, and it spent this money in an effort to improve the core city schools. (Id at 109). However, the Board’s choices as to how to spend its other facilities dollars aggravated segregative trends in the District, thus creating well-funded but separate white and black schools. Specifically, the record shows that the Board did not build its new schools in the integrated area immediately surrounding the core city, which, as the record reflects, contained the highest student population, the highest number of portable classrooms, the largest number of overcrowded schools, and declining school facilities.16 Instead, the Board chose to build thirteen new schools over the past ten years in the almost all white outlying periphery of the county.17 In support of the proposition that its actions did not have a discriminatory pur*988pose, the Board pointed to Milan Mueller’s report, which showed that the District was experiencing the highest rates of population growth in the periphery of the county. However, almost all of these new schools were under-enrolled at the beginning of their tenure. In contrast, during the same period, the Board built no new schools in the area immediately surrounding the core city to meet the needs of the naturally integrated and overpopulated schools.
In short, the record evidence in this case commands a finding that the Board failed to achieve the CSA’s goals with regard to the core city’s historically black schools under the magnet program, and that it did not give fair consideration to the CSA’s supplementary methods for desegregating these schools. Simultaneously, the Board started a building program that created a new segregative trend within the District, and ignored the naturally integrating trends that would have complemented its desegregation efforts.
D. Whether The District Was Desegregated To “The Extent Practicable. ”
The majority has also concluded that, in addition to showing that it complied with the CSA in good faith, the Board has satisfied its obligation under the unitary status inquiry to desegregate the district “to the extent practicable.” Again, I would not reach this issue because I do not believe that the Board satisfied its obligation to show that it complied with the CSA in good faith. However, I address this point because I do not agree with the majority’s view that the Board met its burden to show that demographic changes made it impossible for it to achieve the CSA’s student enrollment goals for the Attachment C schools.18
The Supreme Court has held that when a court reviews a school board’s compliance with a desegregation order under the unitary status inquiry, it must presume that any remaining segregation in the school district is tied to the former de jure segregated school system, and therefore must be remedied before the court ends judicial supervision of the district. Keyes v. Sch. Dist. No. 1, Denver Colo., 413 U.S. 189, 208, 93 S.Ct. 2686, 37 L.Ed.2d 548 (1973). A school board may rebut the presumption that the segregation in its district stems from the former de jure *989racially segregated school system by presenting evidence that it temporarily achieved the desegregation decree’s goals, but that demographic changes in the district led to resegregation, unrelated to the original de jure segregated school system. Freeman, 503 U.S. at 494, 112 S.Ct. 1430 (“Where resegregation is a product not of state action but of private choices, it does not have constitutional implications.”); cf. Spangler, 427 U.S. at 435, 96 S.Ct. 2697 (recognizing that a school board that had already achieved a desegregation order’s goals could not be required by the court to continually change its efforts to address the changing demographic composition of the district ).19 Courts have also recognized that a school board may use demographic evidence to show that it cannot accomplish a desegregation decree goal because the demography of the district has dramatically changed, and the desegregation plan is no longer workable. Dowell, 498 U.S.at 242, 111 S.Ct. 630 (discussing a district court order in which the court ruled that demographic changes made the district’s desegregation plan “unworkable”). These arguments, however, are not available to assist those school districts that, because of willfulness or neglect, still contain a large pool of segregated schools after decades of federal court supervision, and these schools are the exact same schools that the court initially identified as the vestiges of the district’s de jure segregated school system, and the high racial concentrations at these schools are not traceable to any independent post-decree demographic changes.
The Board cites demographic evidence to show that the CSA’s desegregation goals are impracticable because the District has experienced “white flight.” See Morgan, 831 F.2d at 322-23 (recognizing white flight as a factor beyond a school board’s control); Ross v. Houston Indep. Sch. Distr., 699 F.2d 218, 225 (5th Cir.1983) (same). Specifically, the Board claims that white students have withdrawn from the Duval County public school system to attend private schools, leaving it with an insufficient white student pool to achieve the CSA’s goals. The Board also argues that whites have moved away from the neighborhoods associated with the core city schools, making it even more difficult to desegregate these schools. The Board offered three pieces of evidence to support its contention of “white flight”: (1) district-wide enrollment statistics showing a 12% decrease in the number of white students over the twenty seven years prior to the CSA’s creation; (2) eyewitness testimony from its former superintendent and current desegregation advisor, Dr. Larry Zenke; and (3) a report on black residential migration patterns, prepared by Dr. Mueller. I cannot agree with the majority that this evidence showed that a new post-decree trend of “white flight” transformed the core city into an area with a school-age population that is 96% black and, thus, made it impracticable to further desegregate the core city schools. (Maj. Op. at 970). Rather, the record shows that the District’s white flight predated the CSA, and was adequately accounted for in the CSA’s student enrollment goals.
The Board’s argument mainly relies on the District’s student enrollment statistics showing that between 1969 and 1990, the proportional number of white students in Duval County’s public schools decreased *990by 12%. They ask that we rule that this decrease, suffered over a twenty seven year period, was significant enough to render the enrollment goals under the CSA for the Attachment C schools impracticable. However, the record indisputably establishes that the majority of this white flight had occurred before the CSA was created. (Dist. Ct. Op. at 10). During the enforcement term of the CSA the Board experienced only a 4% decrease district-wide in the number of white students.20 Because the Board was already aware of the majority of this white flight when it negotiated the agreement, it seems likely that the parties accounted for this change, and agreed to relax the black student enrollment figures to 20%-55% at the core city elementary schools, rather than the more stringent 21%-34% figure required under the Mims Injunction. Indeed, it seems fundamentally unfair to allow the Board now to argue that the more relaxed enrollment standards under the CSA are unreachable based on this earlier larger pattern of white flight, particularly when this problem is adequately accounted for in the CSA’s goals.
Rather, in my view, our inquiry must be confined to the 4% white flight that actually occurred during the term of the agreement and, thus, was not accounted for in the CSA’s goals. I am unconvinced that this 4% district-wide shift could have had any real effect on the Board’s ability to desegregate the core city black schools. (Maj. Op. at 970).21 Since the time of the Mims Injunction, the parties have been on notice that it is impossible to desegregate the formerly de jure black core city schools without mass busing. Mims, 329 F.Supp. at 130. The CSA was simply designed to end forced busing, and instead to encourage parents to volunteer to have their children bused to core city schools, either by subtly limiting parents’ choices, and/or by creating a group of highly attractive core city schools. The CSA represents the parties’ attempt in 1991 to establish a fair and equitable standard to determine whether desegregation had been achieved in light of the current demographic status of the District after the implementation of the CSA desegregation program. By allowing the Board to rely on the pre-CSA pattern of white flight as evidence that it cannot achieve the CSA’s goals, the majority allows the Board to escape its contractual obligations.
CONCLUSION
The majority notes that the Duval County School Board has been embroiled in litigation for forty one years, has been subject to orders from five district courts, and has appeared multiple times before this court seeking review of its decisions. In its view, the time for judicial supervision is at an end. However, it likewise bears noting that the duration of court supervision occurs only as a result of a school board’s failure to satisfy its desegregation obligations. I agree that Duval *991County School Board has made substantial progress in many aspects of its desegregation mandate; however, the record does not support the view that it met its obligations in the area of student assignment. Accordingly, I believe that under the CSA we are required to retain judicial supervision over the District in the area of student assignment.
. This figure was computed based on Attachment C, a CSA Appendix that identifies 28 *978schools that the parties recognized would have 75% black student enrollments after the District ended forced busing. As such, these schools were the primary focus of the CSA. Review of Attachment C shows that, out of the 28 schools listed, only 23 are currently open. Therefore, I exclude from my analysis the five elementary schools that were closed; namely, the Sherwood Forrest, Beal, Lackawanna, Rutherford, and Scott elementary schools. I also exclude Jacksonville Beach from these figures, as the school is located in Zone VII, and Attachment C and the CSA clearly indicate that the Attachment C student enrollment goals were only intended to apply to schools in Zones I-V. (CSA Attachment C & CSA ¶ 4). Out of the remaining 22 Attachment C schools, only one (Axson Elementary School) achieved the CSA student enrollment goals in 1998-99, the year the District was granted unitary status. Therefore, 96% of the Attachment C schools were not in compliance with the CSA's student enrollment goals.
. The Supreme Court has indicated that it is improper for the court to alter a school board's desegregation obligations in light of demographic shifts in the district, explaining that once the Board satisfies its original obligations under a desegregation agreement, it cannot be required to continually reorganize its efforts to address the changing composition of the district. See Freeman v. Pitts, 503 U.S. 467, 493-94, 112 S.Ct. 1430, 118 L.Ed.2d 108 (1992); Pasadena Bd. of Educ. v. Spangler, 427 U.S. 424, 434-35, 96 S.Ct. 2697, 49 L.Ed.2d 599 (1976) (same). This rule, however, does not prevent a school board from entering into a desegregation agreement that creates new student enroll*979ment goals that account for demographic changes.
. The majority parses this same data several ways, to show that the Board achieved different rates of success as measured between elementary schools, middle schools and high schools. The distinction between elementary, middle, and high schools however, is not relevant to our assessment. Rather, the focus must be on whether the Board remedied the original constitutional violation in this case— that is, whether it desegregated the former de jure black schools.
. Although the parties describe the Attachment C schools as schools that were "expected to become” racially identifiable black schools, this designation is in some ways a misnomer. Many of the schools in Attachment C were actually the former de jure black schools in existence at the start of desegrega*980tion, which remained identifiably black during Mims Injunction period and were still identifiably black when the CSA was negotiated, as they had not reached their student enrollment goals. The Attachment C group, however, also contains some schools that were expected to become racially identifiably black at the end of the Mims injunction because they were located in or near black neighborhoods created during the era of de jure racial segregation. (Dist. Ct. Op. at 13). The Board was permitted to account for these problems as a contractual matter. Also, it was required to take account of this problem in fulfilling its constitutional obligations. See Freeman, 503 U.S. at 514, 112 S.Ct. 1430 (Blackmun, concurring).
. There is some authority that suggests that a school board may deviate from a desegregation order's student enrollment goals by as much as 15-20% and still be considered in compliance. See Belk v. Charlotte-Mecklenburg Bd. of Educ., 269 F.3d 305, 319 (4th Cir.2001) (citing Manning v. School Bd. of Hillsborough County, 244 F.3d 927, 935 (11th Cir.2001)). I offer no opinion on whether this would be a fair standard for assessing compliance under a desegregation agreement, like the one at issue here. However, even under this more permissive measure, the Board's enrollment statistics show that more than half of the Attachment C schools deviated more than 20% from the CSA's student enrollment goals, and 85% of the Attachment C schools named in the Mims Injunction deviated more than 20% from the CSA's goals.
. Middle and high schools were treated separately, and were not the main focus of the agreement. (Dist. Ct. Op. at 4).
. The CSA was created and negotiated between 1989 and 1990 and, because of the need for minor technical corrections, was reissued in 1991. I use 1991 as the start date of the CSA enforcement period for the purposes of clarity.
. Dr. Larry L. Zenke, the Board's desegregation advisor and the former superintendent of the District, testified that, when the CSA was created, the Board recognized that black students would willingly transfer into white schools, even when there were no magnet programs at the white schools, and even when there were no additional resources available at the white schools. (Vol. 10 Tr. at 87-88).
. Again, the number of elementary schools in Attachment C is higher than the number in the Mims Injunction because the parties agreed that the Board was required to remedy segregation in those schools expected to become identifiably black at the end of mass busing. These schools were "expected to become” majority black because they were either located in or near black neighborhoods that developed as a consequence of de jure segregation.
. The majority concludes that the Board successfully desegregated the core city schools because substantial numbers of black students attended the Attachment C schools at what later became the CSA’s goals. Specifically, the majority states:
All but two of the fifteen schools formerly operated solely for Black students reached what later became the CSA racial balance goals for at least a three-year period before the implementation of the CSA. The Board had broken the pattern of all-Black enrollment at the schools it formerly operated solely for Black students during de jure segregation, and had eliminated the vestiges of de jure segregation in all the Attachment C schools.
(Maj. Op. at 968-69). (emphasis added). As explained above, the Mims standards govern the period prior to 1991 and,, thus, the CSA is *983the wrong standard to apply to this time period.
. The Mims Injunction also set goals for the formerly de jure black middle schools and high schools. However, because Attachment C only covers elementary schools, these figures are not relevant to our analysis.
. As noted above, the majority argues that the Board met the CSA goals for the Attachment C schools for the requisite three year period prior to 1991. However, in NAACP I, this Court clearly indicated that Board's performance prior to 1986 was insufficient to satisfy its burden under the proper measure, the Mims standard. In regard to the Board’s performance after 1986, my review of the Board's student enrollment statistics shows that the Board's performance remained static, and that the number of racially identifiable schools did not substantially change from 1986, the date it was denied unitary status in NAACP I. See Def. Exh. 96.
.This graph only covers the schools that are the “core city” schools identified in the Mims Injunction, and were included in the subset of Attachment C schools that the Board was required to desegregate during the CSA time period. Stated alternatively, the chart only shows those schools that were originally recognized as the source of the constitutional injury in this case, and were specifically designated as subject to the CSA’s student enrollment goals. As shown above, none of the Attachment C schools that were previously identified in the Mims Injunction met the CSA’s student enrollment goals.
. I am also concerned that the majority’s reading of the CSA construes the agreement to provide for a program that is dangerously close to the voluntary choice program the Supreme Court rejected in Green v. New Kent County. See Green, 391 U.S. at 437-40, 88 S.Ct. 1689 (explaining that "in desegregating a dual system a plan utilizing 'freedom of choice' is not an end in itself.”). In Green, the Supreme Court explicitly held that a school board cannot satisfy its obligation to implement a desegregation plan merely by implementing an ineffective "voluntary” or "free choice” attendance program. Id. Therefore, if the contract could be interpreted in a manner that would not bring it in conflict with the Board's constitutional obligations, we are required to interpret it in this manner. In this case, that reading would require that we construe the agreement to provide for additional obligations beyond the establishment of a free choice program. See Restatement (Second) Contracts § 207 ("In choosing among the reasonable meanings of a promise or agreement or a term thereof, a meaning that serves the public interest is generally preferred.”).
. Specifically, "a best efforts contract requires that the obligor must use best efforts to achieve a particular goal, but the risk of failure lies with the obligee. Black's Law Dictionary at 318. To be enforceable, a best-efforts term must generally set some kind of goal or guideline against which the efforts may be measured. Id. See also Murray On Contracts § 58 (explaining how the standard is interpreted in the commercial context).
. The Board’s demography expert, Dr. Milan Mueller conducted a study which documents this problem. Mueller divided Jacksonville into three concentric rings, Area A, the core city, and areas B and C, the successive surrounding rings. Most of the Attachment C schools fall in Area A. Mueller testified at trial that thirteen new schools were built in Area C, a section of Jacksonville in which the student population is 84% white, and one school was built in Area A, the core city, where the student population is 97% black. See Def. Exh. 77, tab. 12. No new schools were built or planned in Area B. This was the case even though Mueller’s demographic research indicated that, throughout the course of the study, the highest concentration of students were enrolled in Area B (approximately 50% of the overall enrollment in the school district), an integrated area which was 55% black and 45% white. See Def. Exh. 77, tab.4. The Board argues that it made a valid administrative choice, as it was building schools in areas experiencing the most significant student population growth. (Tr. Vol. 3 at 115). Although future planning is indeed a laudable goal, it is not laudable when it comes at the expense of remedying the Fourteenth Amendment violation.
. Instead of building new schools in the newly integrated outer core of the city, the Board used 500 portable classrooms in that area of the district. Simultaneously, it built *988new schools in the predominantly white residential neighborhoods in the outlying areas of the city away from the core. The Board knew that the Court was concerned about its distribution of new facilities, separate and apart from the Board's obligation to remedy the poorly funded and inadequate core city schools. Indeed, in NAACP II we held, that the Board's "failure to consider the objective of desegregation in its efforts to alleviate overcrowding violate[d the Board's] affirmative duty to desegregate.” When combined with other factors, we concluded that this counseled against a unitary status finding. Id. at 952-53. Despite this warning, the Board made no effort to consider whether its strategy of using portable classrooms in the core, as opposed to building new schools, conflicted with its affirmative duty to desegregate the District's schools.
. Together the Board's student enrollment statistics, and the demography report prepared by the Board's expert, Dr. Mueller, establish that the core city schools had 90% black enrollments during de jure segregation, and the black pupil enrollments at these schools remained in excess of 90% for the duration of the CSA. The Board submitted raw data for the 1969-1970 time period. Mueller's report provides percentages for the later dates. The report shows that the core city schools were 96.1% black between 1989 and 1990, 92.78% black between 1993 and 1994, and 93.45% between 1996 and 1997. The Board failed to identify any evidence to show that the 90% segregation in the core city schools was traceable to any intervening demographic shift that changed the racial composition of these areas.
. The majority appears to suggest that once a Board presents evidence of a demographic shift within the district, this establishes that there is "no constitutional violation.” However, the proof of demographic shift shows that some segregated schools should not he included in the court's assessment gs to whether the original constitutional violation has been remedied.
. This fact is corroborated by Mueller's report, showing that the black population district-wide increased from 36.5% to 40.2% over the course of the CSA, and the white population decreased from 59.5% to 54%. (Mueller Report at 4).
. Also, a 4% shift in the District's white student enrollment is hardly the kind of dramatic demographic change about which the Supreme Court was concerned. For example, in Freeman v. Pitts, the black student ratio increased from 5.6% in 1969 to 47% in 1986. See 503 U.S. at 476, 112 S.Ct. 1430. Also, in Missouri v. Jenkins, 515 U.S. at 95 n. 6, 115 S.Ct. 2038., the demographic shift that triggered the district's doomed attempt to create an interdistrict remedy was from 18.9% in 1954 to 60% in 1975. Here the record shows no such dramatic change in the racial makeup of the student population in Duval County.