San Francisco NAACP v. San Francisco Unified School District , 413 F. Supp. 2d 1051 ( 2005 )

413 F. Supp. 2d 1051 (2005)

SAN FRANCISCO NAACP, et al., Plaintiffs,
v.
SAN FRANCISCO UNIFIED SCHOOL DISTRICT, et al., Defendants.
Brian Ho, by his parent and next friend, Carl Ho, et al., Plaintiffs,
v.
San Francisco Unified School District, et al., Defendants.

Nos. C 78-01445 WHA, C 94-02418 WHA.

United States District Court, N.D. California.

November 8, 2005.

Brain H. Chun, James L. Hunt, Margaret A. McLetchie, Bingham McCutchen LLP, Diana C. Tate, Robert J. Rubin, Lawyers Committee for Civil Rights of San Francisco Bay Area, Eva J. Paterson, Equal Justice Society, San Francisco, CA, Thomas I. Atkins, Brooklyn, NY, Peter Graham Cohn, Petaluma, CA, Anthony K. Lee, San Francisco, CA, for Plaintiffs.

Aubrey V. McCutcheon, Jr., Aubrey V. McCutcheon Jr. Law Offices, Ypsilanti, MI, Dennis J. Herrera, City Attorney's Office, San Francisco, CA, Jerome B. Falk, Jr., Howard, Rice, Nemerovski, Canady, Falk & Rabin, San Francisco, CA, Maree F. Sneed, Hogan & Hartson, LLP, Washington, DC, Therese M. Stewart, San Francisco, CA, John W. Borkowski, Hogan & Hartson, LLP, New Orleans, LA, Robin Johansen, Thomas A. Willis, Remcho Johansen & Purcell, San Leandro, CA, Stuart Biegel, Los Angeles, CA, Gary Orfield, Cambridge, MA, for Defendants.

Hoover Liddell, San Francisco, CA, pro se.

Robert L. Green, Las Vegas, NV, pro se.

Gwen Stephens, Sacramento, CA, pro se.

Laureen Chew, San Francisco, CA, pro se.

J. David Ramirez, Brooklyn, NY, pro se.

ORDER DENYING PROPOSED EXTENSION OF CONSENT DECREE

ALSUP, District Judge.

INTRODUCTION

A half-century ago, in the landmark decision Brown v. Board of Education, 347 U.S. 483, 74 S. Ct. 686, 98 L. Ed. 873 (1954), the Supreme Court called for an end to de jure school segregation. Commenced in 1978, this class action accused the San Francisco Unified School District of de jure segregation. Although this charge was never proven, a settlement in 1983 resulted in a consent decree.

For twenty-two years, the consent decree has regulated all public school assignments for San Francisco's children. At first, the decree resulted in better racial integration. Later, as a result of a different lawsuit challenging the decree, the parties agreed on a different method for student assignment, one that incorporated a system called the "diversity index." This proved to be a misnomer, for it has not achieved diversity in any meaningful sense. To the contrary, the student-assignment system imposed by the negotiated revision has allowed, if not fostered, resegregation in San Francisco schools.

Since the case was reassigned to the undersigned judge on January 11, 2002, the Court has repeatedly urged the parties to fix the so-called diversity index and substitute a system that will better achieve racial integration. While the parties have discussed alternatives, they have not come close to even proposing a solution. By its own terms (and by prior stipulation of all parties), the consent decree is scheduled to end on December 31, 2005, having previously been extended from December 31, 2002. Earlier this year, plaintiffs' counsel threatened to seek another extension (despite the stipulated termination date). The deadline for filing any such motion was set for August 25, 2005. In lieu of filing a motion, however, the parties submitted a compromise proposal to extend the decree by another eighteen months through June 30, 2007, a proposal that would, among other things, bar any revision to the student-assignment method for another year.

This proposal was the subject of a public hearing at which many San Francisco parents attended and were heard. It is clear that the present system imposes a burden on families. Their sacrifice has been in vain, for the diversity index has not and will not produce the benefit of diversity or racial integration. For this reason and other reasons now set forth, including the failure to prove up the legal prerequisites for prolonging the consent decree, this order DENIES the proposed eighteenmonth extension. The consent decree will end, as earlier agreed, on December 31, 2005.

STATEMENT

In 1978, the San Francisco branch of the National Association for the Advancement of Colored People ("SFNAACP") and a group of black parents filed this classaction lawsuit. They charged the San Francisco Unified School District, its board members and its superintendent, the California State Board of Education and its members, the State Superintendent of Public Instruction and the State Department of Education with engaging in racially discriminatory practices and maintaining a segregated school system in San Francisco, in violation of the constitutions and laws of the United States and California. Plaintiffs sought a declaratory judgment and injunctions guaranteeing them equal educational opportunity and fully desegregated schools under a court-ordered desegregation plan.^[1] The action was assigned to the Honorable William H. Orrick, Jr., who presided over the case for almost all of its nearly thirty-year history.^[2]

During a national era of widespread de facto segregation, if not de jure segregation, San Francisco faced a situation similar to that of many other large urban school districts: Its schools were sharply divided along racial lines. In 1970, 63 of 96 elementary schools were segregated. At those schools, one racial group comprised over fifty percent of the school population at a time when the largest racial group—white students—comprised only 34 percent of the total elementary-school population. There were 29 white schools, 23 black schools, four Hispanic schools and seven Chinese schools (Order Denying Plaintiffs' Motion for Partial Summary Judgment, June 30, 1981, at 11, 42-43).

Schools were also segregated according to faculty and administrators. The racial and ethnic makeup of the district's teachers was consistently disproportionate to that of its students. In 1970, although white students comprised only 35 percent of the district's total student population, white teachers comprised 79 percent of the district's teachers (id. at 58-60). In 1965, more than 73 percent of the district's regular schools had no black teachers (id. at 60). Most black teachers were assigned to predominantly black schools, where teachers tended to be less experienced and paid lower salaries (id. at 60-61). A similar pattern of segregation applied to school administrators. In 1965, there were only two nonwhite administrators in the entire district. The only black principal was assigned to Bret Harte, a school 77 percent black. The only Chinese principal was assigned to Spring Valley, a school 91 percent Chinese (id. at 64).

1. LITIGATION LEADING TO THE 1983 CONSENT DECREE.

In the five years following the filing of the complaint, the parties litigated various issues. On September 18, 1979, the state defendants' motion for dismissal or abstention was denied. Judge Orrick rejected defendants' arguments that: (1) the Court should not decide the "unsettled" issue of state law concerning the apportionment of policy-making powers between state and local authorities, and (2) primary responsibility for implementing education policy should rest with the state legislature and local school districts, not with defendant state agencies.

Because proof of segregative intent was a predicate for demonstrating a violation of equal protection, this issue was also litigated. See, e.g., Personnel Adm'r of Mass. v. Feeney, 442 U.S. 256, 279, 99 S. Ct. 2282, 60 L. Ed. 2d 870 (1979). To establish segregative intent, plaintiffs needed to demonstrate evidence of past purposeful discrimination that would then give rise to liability for subsequent acts which had the effect of maintaining a segregated school system. Absent such a showing, plaintiffs needed to demonstrate ongoing intentional segregation, a higher burden.

On September 24, 1981, the Court denied the school district's motion for a separate trial on the issue of whether, at the time of filing of the original complaint, it was operating a segregated school system. While that motion was still pending, on June 30, 1981, the Court denied plaintiffs' motion for partial summary judgment on the issue of whether, as of 1954 and 1970, the district was intentionally racially segregated.

These two motions sharpened the issues for trial and highlighted the factual strengths and weaknesses of plaintiffs' case. For example, plaintiffs had stronger evidence on racial disparities in faculty and administrative hiring but lacked evidence concerning discrimination against minority groups other than blacks. As a result, the parties entered into settlement negotiations, drawing on the expertise of a Courtappointed settlement team co-chaired by Professor Harold Howe II and Professor Gary Orfield, two of the most well-reputed educators in the country; both had extensive experience in desegregation cases. The parties submitted a consent decree on December 30, 1982. After a fairness hearing held on February 14, 1983, the Court approved the consent decree by written decision on May 20, 1983. San Francisco NAACP v. San Francisco Unified Sch. Dist., 576 F. Supp. 34 (N.D.Cal.1983).

The goal of the decree was "to eliminate racial/ethnic segregation or identifiability in any District school program or class and to achieve throughout the system the broadest practicable distribution of students from nine racial and ethnic groups." Id. at 37; Consent Decree ("CD") ¶ 12. Guidelines were adopted to prevent any racial/ethnic group from exceeding 45 percent of the student body at any regular school or 40 percent at any "alternative" school (CD ¶ 13). All schools and all grades (K-12) were covered. The decree further designated nineteen historically segregated schools for special desegregation treatment and identified certain schools in Bayview-Hunter's Point for reconstitution and conversion into magnet schools (CD ¶¶ 14, 17).^[3]

Many of the policies and practices alleged to have contributed to a racially segregated school system were also addressed. For example, the decree provided that each school's staff should reflect the district-wide racial/ethnic composition of students and that staff would be equitably assigned throughout the district (CD ¶¶ 34-35). The parties were also required to report to the Court on the fairness and consistency of disciplinary actions (CD ¶ 38).

Finally, the decree stated that "the overall goal of this Consent Decree will require continued and accelerated efforts to achieve academic excellence" and required monitoring of student academic progress in an annual report (CD ¶¶ 39-40). Other highlights of the decree included a requirement that both the school district and an independent reviewer (subsequently known as the "consent-decree monitor") submit an annual report regarding the district's compliance with the consent decree and a provision that the state would pay the costs of complying with the decree (CD ¶¶ 44-45).

2. RESULTS OF THE 1983 CONSENT DECREE.

Implementation of the consent decree proceeded successfully, with the exception of a dispute over exactly how much the state was required to reimburse the district for the costs of implementation. See San Francisco NAACP v. San Francisco Unified Sch. Dist., 695 F. Supp. 1033 (N.D.Cal.1988), rev'd, 896 F.2d 412 (9th Cir.1990). After eight years of implementation, the Court commissioned a comprehensive report on the results of the consent decree, appointing Professor Orfield to lead the investigation team. This committee of experts filed its report in June 1992, finding that the district had "largely achieved the Decree's desegregation goals," while also noting the continuing achievement gap for black and Hispanic students. Desegregation and Educational Change in San Francisco, Findings and Recommendations on Consent Decree Implementation 1 (1992).

Based on recommendations made in the report, the Court approved a modified consent decree on November 5, 1993 (hereinafter "CD II"). The modified decree left the basic structure of racial/ethnic distribution the same but incorporated changes in individual school treatment (CD II at ¶¶ 15-16). The district was also required to provide an additional annual report showing the longitudinal impact of the consent decree (CD II ¶ 40).^[4] At the fairness hearing, motions to intervene were heard from Multicultural Education, Training, and Advocacy, Inc., representing the interests of both Latino and Chinese students, and the local teachers union. On July 22, 1993, the Court denied the motions to intervene while permitting two coalitions, representing the interests of Latino and Chinese students respectively, to act as amici curiae.

The goal of the consent decree was to promote desegregation. In doing so, it intended to improve the quality of education all students received by equalizing access, furthering diversity and giving effect to every child's right to equal educational opportunity under Brown v. Board of Education. For a period of time, this goal was realized. The consent decree succeeded in promoting integration and academic excellence throughout the district. Between 1983 and 1997, fifteen schools were reconstituted, five new schools were established and one school was closed (Stuart Biegel, Report 17 of the Consent Decree Monitoring Team 9 (1999-2000) [hereinafter "Biegel Report No. 171"]). At the apex of integration, during the 1997-98 academic year, only thirty regular district schools and six alternative schools (out of 122 schools total) were out of compliance with the 45/40 percent requirement, and only one school enrolled more than fifty percent of a single race/ethnicity. In addition, standardized test scores increased seven years in a row between 1992 and 1999. Reading scores increased for students of every race and ethnicity, with the exception of a slight drop for Korean students; math scores also increased for students of every race and ethnicity, with the exception of a slight drop for Native American students (Biegel Report No. 16 (1998-99) at 45, 116).

3. THE HO ACTION.

Notwithstanding this success, some members of the San Francisco community were not pleased with the results of the 1983 consent decree, particularly the requirement that student assignment to schools adhere to a certain racial/ethnic distribution. After 1983, the demographics of the district changed significantly. Enrollment of black and white students declined as Chinese-American and Latino enrollment grew. While Chinese-American students represented 19.5 percent of the district in 1983, by 1992 they had grown to 24 percent, the largest ethnic subgroup in the district.^[5] The practical implication of this demographic trend was that Chinese-American students were bumping up against the 45/40 percent enrollment ceilings set by the 1983 consent decree, especially at the city's most desirable public schools. At Lowell, the district's renowned academically-selective, alternative high school, for example, Chinese-American students needed a significantly higher admission index score to gain admission than did white, black or Latino students.

On July 11, 1994, several schoolchildren of Chinese descent filed suit against the district and state defendants. The plaintiffs in Ho v. San Francisco Unified School District, a Lowell applicant as well as two elementary school children who had been "capped out" of their neighborhood schools, alleged that the student-assignment plan described in Paragraph 13 of the consent decree—the 45/40 percent ethnic distribution requirement—constituted race discrimination in violation of the Equal Protection Clause. In January 1995, the Ho plaintiffs filed a first amended complaint adding the SFNAACP as a defendant. In March 1996; the Court certified a class comprised of San Francisco schoolchildren of Chinese descent.

The Ho plaintiffs then filed a motion for summary judgment, arguing that the ethnic-distribution requirement of Paragraph 13 was unconstitutional in 1983 and remained so. In the alternative, they argued that because the district had already complied for fourteen years, the decree was no longer, necessary and should be dissolved. In May 1997, the motion for summary judgment was denied. The Court held that the decree was not unconstitutional when entered and that, in any event, res judicata barred the plaintiffs—as members of the original class—from raising the issue. Ho v. San Francisco Unified Sch. Dist., 965 F. Supp. 1316, 1320 (N.D.Cal. 1997). The Court also held that there were issues of fact precluding summary judgment, at least as to: (1) the existence of past discrimination justifying the consent decree; (2) the necessity of Paragraph 13 as a sufficiently narrowly-tailored remedy; and (3) whether continued judicial supervision was necessary to achieve full compliance with the decree. Id. at 1324-25, 1327.

The Ho plaintiffs appealed the summary-judgment ruling. They also sought a writ of mandamus directing the Court not to proceed with trial. In an opinion by Judge John Noonan filed on June 4, 1998, the Ninth Circuit dismissed the appeal for lack of jurisdiction. It noted that "proper resolution of any desegregation case turns on a careful assessment of its facts," which had not yet been presented. Ho v. San Francisco Unified Sch. Dist., 147 F.3d 854, 860 (9th Cir.1998) (emphasis in original). The court also denied the petition for a writ of mandamus, while nevertheless pointing out weaknesses in the evidence justifying the continued implementation of Paragraph 13. After a lengthy discussion of the history of racial discrimination against the Chinese in San Francisco, Judge Noonan identified the two issues that remained for trial: "Do vestiges remain of the racism that justified Paragraph 13 of the consent decree in 1983? Is Paragraph 13 necessary to remove the vestiges if they do remain?" Id. at 865. The Ninth Circuit stressed that defendants' evidence must tie the current vestiges of segregation to the discriminating practices and policies that justified the adoption of the consent decree in 1983. It also held that it was defendants' burden "to demonstrate that Paragraph 13 is still a remedy fitted to a wrong—to show that the racial classifications and quotas employed by Paragraph 13 are tailored to the problems caused by vestiges of the earlier segregation." Ibid.

On December 10, 1998, the Court denied the Ho plaintiffs' motion for a preliminary injunction. That motion sought to enjoin defendants from further implementation of the student-assignment plan described in Paragraph 13. The Court found that plaintiffs had not shown irreparable harm. If they succeeded at trial, there would be ample time to assign students to schools in a race-neutral manner for the following school year.

A. The 1999 Settlement.

On February 16, 1999, the first day of trial, the parties submitted a settlement agreement, which was later approved in an order filed July 2, 1999. Therein, Judge Orrick expressly noted that the Ho "[p]laintiffs were likely to succeed on their claim that the race-based student-assignment plan is no longer constitutional" because defendants were unlikely to be able to meet the burden of "proving that vestiges still exist of the discriminatory policies and practices that justified the Consent Decree in 1983, and that the race-based student-assignment plan in the Consent Decree is necessary to address the problems caused by the vestiges of the earlier segregation." San Francisco NAACP v. San Francisco Unified Sch. Dist., 59 F. Supp. 2d 1021, 1029 (N.D.Cal.1999).

The central feature of the Ho settlement was a modification of Paragraph 13 to allow the creation of a new student-assignment plan that eliminated race as a factor in determining student placement. The Court further required the parties to respond to reports submitted pursuant to Paragraph 44 of the original 1983 consent decree, which mandated an annual independent review of the district's progress. Id. at 1039. Specifically, the parties were required to address issues raised in those reports by members of the consent-decree advisory committee (represented at hearings by Hoover Liddell) and consent-decree monitor Stuart Biegel.^[6]

The Ho settlement agreement also set a termination date for the consent decree of December 31, 2002, subject to Court approval (¶ A); provided for the modification of Paragraph 12a, so that the district could request, but not require, parents and/or students to identify themselves by race at the time of enrollment (¶ D); and stipulated that the parties must meet to discuss additional modifications to Paragraph 13 if the new student-assignment plan resulted in an identifiable racial/ethnic concentration at any particular school (¶ G). On August 11, 1999, the consent decree was amended to include the terms of the Ho settlement (hereinafter "CD III").

During 1999-2001, the reports submitted by the advisory committee and Mr. Biegel reflected continuing problems within the school district and failure to comply with certain terms of the decree, including the development of a race-neutral student-assignment plan (see CD III ¶ 13). The district was ordered to file a comprehensive plan to address these problems, which it did on April 11, 2001. The plan was titled "Excellence, for All: A Five-Year Comprehensive Plan to Achieve Educational Equity in the San Francisco Unified School District for School Years 2001-02 through 2005-06." At the parties' request, Thomas J. Klitgaard was appointed as special master to assist the parties in resolving any disputes.

B. The 2001 Settlement.

On July 11, 2001, the parties submitted an amended stipulated settlement agreement, which was incorporated into the current version of the decree (hereinafter "CD IV"). These changes included further amendments and extended the life of the decree to December 31, 2005, with state funding to continue until June 30, 2006, in order to "facilitate an orderly transition" (CD IV ¶ 48-1). This extension was allegedly necessary due to "unexpected turmoil" in the district. This included the sudden resignation of Superintendent Waldemar Rojas (after which the school district was left without a leader for over a year) and widely-publicized financial mismanagement. The amended settlement expressly stated that by December 31, 2005, the district "shall have taken all practicable steps" to eliminate any vestiges of de jure discrimination, thereby specifically incorporating the legal standard for terminating desegregation consent decrees. See Bd. of Educ. of Oklahoma City Pub. Schs. v. Dowell, 498 U.S. 237, 111 S. Ct. 630, 112 L. Ed. 2d 715 (1991); Freeman v. Pitts, 503 U.S. 467, 112 S. Ct. 1430, 118 L. Ed. 2d 108 (1992).

As part of the 2001 settlement, the district agreed to adopt and comply with a system of monitoring and review of consent decree budgets and programs. More importantly, the settlement resulted in a new student-assignment provision laid out in Paragraph 13 of the decree to fulfill the mandate of the 1999 settlement. As agreed, race/ethnicity could no longer be used as a factor in student assignment (CD IV ¶ 13(j)).^[7] Instead, the district would assign students to schools based on: (1) where siblings have been assigned; (2) whether any specialized learning needs must be accommodated; and (3) a "diversity index" taking into account socioeconomic status, academic achievement, Englishlanguage learner status, mother's educational background, academic performance at the student's prior school, home language and geographic areas (CD IV ¶ 13(k), incorporating Attachment B).^[8] A separate plan that was consistent with Paragraph 13(j) would be developed for Lowell High School and School of the Arts. Paragraph 13(m) was amended to set forth a procedure for addressing any identifiable racial or ethnic concentration at a particular school, that the parties believed adversely affected the district's goals, although no party could propose a change to the student-assignment procedure until the 2004-05 school year.

4. RESULTS OF THE 1999/2001 SETTLEMENTS.

With the adoption of the student-assignment method which arose from the 1999 and 2001 settlements, the district has become increasingly resegregated. A dramatic achievement gap has persisted for many black and Latino students. During the 2001-02 school year, thirty schools were severely resegregated at one or more grade levels.^[9] By the 2004-05 school year, this number had risen to 43, with 27 schools entirely resegregated (Biegel Suppl. Report (April 2005) App. 1).

To put these numbers in perspective, it is useful to recall that in 1970, over fifty percent of San Francisco schools were segregated, defined as one racial/ethnic group comprising over fifty percent of the school population. By 1997-98, the most successful school year after the 1983 consent decree, only one school—or 0.8 percent of San Francisco schools—enrolled more than fifty percent of a single racial/ethnic group. Six years later, this number has climbed back up to 35 percent. Today, over one in three San Francisco public schools is resegregated, largely as a result of the student-assignment system adopted by the parties and written into the consent decree.

Academic achievement data indicate a close relationship between resegregation and the disparity in academic achievement between black and Latino students in comparison with white and Chinese-American students. The overwhelming majority of schools that have been successful in closing the achievement gap, as indicated by the large percentages of students of all ethnic backgrounds who score at the proficient or above level on state standardized tests, have maintained ethnically and racially-diverse student bodies (Biegel Report No. 22 (2004-05) App. 1 at 2). In contrast, the Academic Performance Index (API), a school-ranking system based on standardized test scores, indicates that schools with declining APIs are those that have become resegregated. For example, all of the elementary schools in Bayview-Hunter's Point that have a predominantly black enrollment, except for Carver, are now at an API of 1, the lowest possible ranking (Biegel Suppl. Report (April 2005) at 18).

Resegregated schools also have the most significant achievement gaps. For example, at Visitacion Valley Middle School, out of 113 black students, only nine scored at the proficient level on the state standardized test for English/Language Arts and only two did the same for math; out of 87 Latino students, only five scored at the proficient level for English/Language Arts and only six for Math (Biegel Suppl. Report (April 2005) App. at 9-10). In addition to standardized test scores, the consent-decree monitoring team has identified numerous other indicators of the dramatic achievement gap among students of different racial and ethnic groups, including the disparity in grade-point averages by school and by ethnicity; the disparity in attendance rates by school and by ethnicity; the number of Advanced Placement classes offered at various district high schools and the ethnic makeup of students in those classes; the number of special-education students at various schools and the ethnic makeup of those students; and the number of suspensions and expulsions by ethnicity. To take but one example, in 2004-05, African-American students comprised 13.7 percent of the student population yet received 51.69 percent of all suspensions (Biegel Report No. 22 Ohio App. 2 at 23).

It should be noted that despite this broad and disheartening story of resegregation, there has also been some slivers of progress towards the original goals of the 1983 consent decree. For example, the most recent report from Mr. Biegel highlighted thirteen schools that exemplified the most successful reforms under the consent decree (id. at 5-7).

5. RECENT HISTORY.

Both the consent-decree monitor and the school district have done an admirable job of documenting the consequences of the decree. With this information in hand, the parties have met regularly to respond to new reports and to discuss the shortcomings revealed therein. Yet, although the parties have appeared at numerous status conferences, since the modifications to the student-assignment plan made in 2001, no substantive motions have been filed to change the diversity index.

The consent-decree monitor has made clear his view that the current studentassignment plan, incorporated into the consent decree itself, has resulted in the resegregation of certain schools. He has repeatedly recommended that geographic zones be taken into account as one factor in placing students, presumably to counteract the effects of de facto residential segregation in San Francisco.^[10] Pursuant to Paragraph 13, beginning with the 2004-05 school year, the parties could meet and confer to address any concerns regarding ethnic concentration at any school and modifications to the assignment plan. Yet, while the parties have discussed the use of geography as a criterion for placement, and the district has agreed to consider this change, no party has filed a motion to compel a fix.

In response to a supplemental report submitted by the consent-decree monitor on March 12, 2004, describing at length the district's achievement gap and its relation to resegregation, the Court, on its own initiative, called for an evidentiary hearing. This was held on August 3, 2004. Dr. William Trent, Professor of Education Policy at the University of Illinois, Urbana-Champaign, testified regarding the correlation between the racial concentration of schools and students' academic performance. He testified that, regardless of race/ethnicity, students in schools that are not racially identifiable do better than students in primarily Latino or primarily African-American schools. Ms. Orla O'Keeffe, director of the district's Educational Placement Center, also testified about how the diversity index functioned to assign students to over-subscribed schools, based on the six factors described above.^[11] Yet, a completely random assignment of students might have been more successful in achieving district-wide diversity. The Court set a deadline of September 2, 2004, for any motion to modify the plan. The parties met and conferred a number of times during the month of August, but no one came forward with a motion to modify the student-assignment plan. The same discredited diversity index was left in place.

Instead, the school district—with the consent of all parties and leadership from Superintendent Arlene Ackerman (who has recently announced her resignation effective June 30, 2006)—implemented the "Dream Schools" initiative, also known as Dramatic Accelerated Educational Reform, starting in 2004. Through this educational-reform measure, the district reconstituted three of its lowest performing and most ethnically isolated schools located in Bayview/Hunter's Point during the 2004-05 school year. It also implemented a new standards-based model of instruction with additional academic support for students. The parties agreed that the Dream Schools initiative was a way to close the achievement gap. The school district targeted seven other schools for conversion. In an order dated October 18, 2004, the Court granted a joint request to amend the decree and compel the school district to take all necessary steps to implement those seven additional Dream Schools in time for the 2005-06 school year.

The local teachers union has challenged the reconstitution aspect of the Dream Schools initiative, which requires all teachers at schools designated for conversion to reapply for their jobs. This related action was removed from state court and is now pending. United Educators of San Francisco v. San Francisco Unified Sch. Dist., No. C 05-02828 WHA (N.D. Cal. removed July 12, 2005).

6. PROPOSED EXTENSION OF THE CONSENT DECREE.

As stated, all agreed that the decree would end on December 31 of this year. Now, however, counsel wish to extend the termination date another eighteen months. The proposal includes no fix to the diversity index or the student-assignment plan. To the contrary, under the compromise proposal, the procedure for making any adjustments to the student-assignment system could not even be invoked before October 15, 2006, meaning that the current student-assignment system would be locked in for yet another year. The proposal to extend the consent decree is summarized as follows:

(1) By October 1, 2005, the school district shall identify at least three additional schools (aside from the seven schools named in the order of October 18, 2004) selected from the list of underperforming schools, for conversion to Dream Schools during the 2006-07 school year. Specific plans for these Dream Schools shall be developed and implemented according to the expectations and operational guiding principles agreed to by the parties.

(2) The two experts who are evaluating the student-assignment method shall consult with the parties, other experts, the Superintendent's Community Advisory Committee On Student Assignment and the State Monitor by November 30, 2005; any recommendations for proposed changes shall be made by January 2006. Public comment must be sought by April 2006.

(3) Any student-assignment method or revisions to the current method for the 2006-07 and 2007-08 school years must comply with Paragraphs 13(j) and 13(m)(ii) of the consent decree, unless written notification of proposed changes for the 2007-08 school year and reasons for such changes (i.e., recommendations of experts) are provided by January 31, 2006. If Court approval is required, any proposed changes must be submitted to the Court by May 31, 2006.

(4) By October 15, 2005 and October 15, 2006, the school district shall make available information concerning the racial or ethnic composition of each school within the district, but no party may invoke the procedures for adjustments to the assignment method earlier than October 15, 2006.

(5) The school district shall include the following disclaimer language in any brochures or catalogs, as well as on its website:

The student assignment method currently used by the SFUSD was developed and adopted by the SFUSD; the decision to use this assignment method was not taken at the demand of any party in the desegregation consent decree lawsuits. The student assignment method was approved by the Court in its October 24, 2001 Order.^[12]

(6) The consent decree shall terminate as of June 30, 2007, without the necessity of further order of the Court. The parties agree that by that date, the SFUSD shall have taken all practicable actions to achieve unitary status, including eliminating any vestiges of past de jure segregation or ethnic discrimination to the extent practicable and complied with the consent decree substantially and in good faith for a reasonable period. The parties expressly agree not to seek any further extension of the consent decree beyond this date, by motion or otherwise.

Public comments were solicited and compiled by counsel for the Court's review. The public response to the proposed extension of the consent decree was overwhelmingly negative. Although the parties appear to tabulate the results differently in their reply briefs, this order will use the district's own numbers by way of example. Of 252 public comments, 213 generally opposed the settlement, 28 supported it, two supported it with suggested amendments and eleven failed to indicate the author's position on the issue.

At the hearing, many members of the public voiced their concerns, including parents, students, teachers and leaders of various local organizations. When pressed, many of those individuals in favor of extending the decree appeared to be arguing for the implementation of the Dream Schools initiative (or other improvements for underperforming schools), not necessarily for the decree itself or the Court's continued involvement. Some advocated an extension of the decree, but with amendments to fix or replace the diversity index. Other parents argued that the decree should be allowed to expire so the school board would be free to devise a new student-assignment plan unencumbered by the diversity-index method. Parents against the current student-assignment system also stressed the burden of having their children spending hours on buses commuting to schools across town, instead of in classrooms or at home with their families. A few parents were undecided because they were unsure whether it would be easier or harder to improve underperforming schools or change the student-assignment system after the decree expires. Some students who spoke felt that the Dream Schools initiative had benefited their schools, but that the consent decree focused too much on achieving diversity instead of improving educational programs.

ANALYSIS

This action is the oldest case on the entire docket for the Northern District of California. It has spanned more than one-tenth of our nation's existence. It has lasted nearly as long as the combined duration of all major wars against foreign powers by the United States since the Revolutionary War. The consent decree itself has lasted almost twice as long as the period of Reconstruction following the Civil War. The entire educational careers of many students were spent under its regime with their own children under it yet.

Counsel proceed as if the question is simply whether the stipulated extension is fair, reasonable and adequate under FRCP 23(e). Even if that were the issue, the extension could not be approved, for the public interest must also be considered. Waits v. Weller, 653 F.2d 1288, 1291 (9th Cir.1981). The perpetuation of resegregation allowed, if not caused, by the decree itself would be contrary to the public interest. In effect, due to the Ho settlement and its bar on revisions to the student-assignment system until the 2004-05 school year, the decree has transformed itself into court-ordered resegregation. This topsy-turvy outcome would remain untouchable until October 2006 under the proposed extension.

But FRCP 23(e) is not the proper framework for analyzing the proposal. This is not a situation where the parties are "settling" a pending class dispute. That happened two decades ago. Rather, we are at the very end of a previously-agreed transition period, with the consent decree currently scheduled to expire in a few weeks. Extending the consent decree now will undo that deadline.

Any motion to modify or extend the consent decree was supposed to have been filed by August 25, 2005. No motion was filed. Instead, counsel, perhaps accustomed to stipulations rather than formal motions and expecting the Court to acquiesce, simply stipulated to modify the consent decree and extend federal court supervision another eighteen months. Counsel in this action have earned the Court's respect and admiration. In part for that reason, the Court has made the effort to explain at length herein why their compromise proposal and joint request for extension are rejected.

The Court has repeatedly stressed that any modification or extension of the consent decree must satisfy the Dowell/Freeman test, setting forth the standard for when termination of a consent decree is appropriate. Yet, SFNAACP, the Ho plaintiffs and the school district collectively devoted less than four pages of their otherwise extensive briefs to conclusory remarks regarding the application of Dowell and Freeman. Only the State Superintendent's brief contained any analysis under this standard. Plainly put, counsel have failed to meet their burden. They have failed to show that the vestiges of past de jure segregation, assuming that there was ever any de jure segregation in the first place, have not already been eliminated. And, they have failed to cure the problem that the consent decree itself—through the student-assignment plan with its so-called diversity index—is allowing, if not causing, racial resegregation.

1. THE CONSENT DECREE INCLUDES A SPECIFIC TERMINATION DATE.

As a result of the 2001 settlement and subsequent modifications to the consent decree, Paragraph 48-1 now provides:

The consent decree shall terminate as of December 31, 2005 without the necessity of further order of the Court. To facilitate an orderly transition, the State shall continue its Consent Decree funding through the end of the SFUSD's 2005-06 school year on June 30, 2006. As the SFUSD shall have, by December 31, 2005, taken all practicable actions to achieve unitary status, including eliminating any vestiges of past de jure racial or ethnic discrimination to the extent practicable and complying with the Consent Decree substantially and in good faith for a reasonable period, the SFUSD and the State shall oppose any attempt to extend, by motion or otherwise, the Consent Decree beyond December 31, 2005.

This language is notable. First, it modified the 1999 settlement, so termination of the consent decree was no longer "subject to court approval." Significantly, all parties agreed to the termination date. Second, both the school district and the state agreed to oppose any attempt, by motion or otherwise, to extend the consent decree beyond December 31, 2005. (The San Francisco NAACP was the only party arguably permitted, and only by implication, to seek extension of the consent decree by motion or otherwise, although it too had expressly agreed to end the consent decree.)

Paragraph 48-1 specifically incorporated the Dowell/Freeman legal standard (addressed below) for ending desegregation decrees and was premised on the expectation that the school district would have met this standard by December 31, 2005. The remaining provisions of the consent decree provided a clear statement of the "practicable actions" the district needed to take to achieve compliance and eliminate any vestiges of discrimination. See, e.g., Lee v. Butler County Bd. of Educ., 183 F. Supp. 2d 1359 (M.D.Ala.2002) (terminating consent decree where parties had agreed to updated consent decree detailing remedial steps to be taken within three years and district had complied based on evidence presented to court).

The argument is now made that an eighteen-month extension is necessary for a smooth and orderly transition away from federal court supervision. This, however, was the rationale for the extension adopted in 2001. The consent decree was originally scheduled to terminate on December 31, 2002, but it was extended to December 31, 2005. Having chosen this date themselves, the additional three years (and extra six months of state funding) was presumably agreed by all parties to be sufficient for the transition. In short, we have already been in "transition" for almost three years. This argument, standing alone, is not persuasive.

2. THE DOWELL/FREEMAN TEST.

During the 1990s, the Supreme Court addressed the question of when termination of a consent decree in the school-desegregation context is appropriate. In Dowell, the school board sought termination of a class-action desegregation decree that had been entered almost twenty years prior upon a finding that the district had been operating an intentionally segregated school system. The Supreme Court remanded to the district court to determine whether the desegregation decree should be terminated, stating a two-part test.

First, the district court was to determine whether the school district had complied in good faith with the decree since it was entered. Second, the court was to focus on whether the vestiges of de jure discrimination had been eliminated "to the extent practicable." To answer this latter question, the district court was instructed to look to "every facet of school operations," including faculty, staff, transportation, extra-curricular activities, facilities and student assignment. Dowell, 498 U.S. at 249-50, 111 S. Ct. 630; see also Freeman, 503 U.S. at 492, 112 S. Ct. 1430, 118 L. Ed. 2d 108 (1992) (adding "quality of education" to the list of factors).

In Freeman, the Supreme Court recognized that a desegregation decree may be terminated in incremental steps. It held that "federal courts have the authority to relinquish supervision and control of school districts in incremental stages, before full compliance has been achieved in every area of school operations." Factors to be considered in ordering partial withdrawal of judicial control included: "whether there has been full and satisfactory compliance with the decree in those aspects of the system where supervision is to be withdrawn; whether retention of judicial control is necessary or practicable to achieve compliance with the decree in other facets of the school system; and whether the school district has demonstrated, to the public and to the parents and students of the once disfavored race, its good-faith commitment to the whole of the court's decree and to those provisions of the law and the Constitution that were the predicate for judicial intervention in the first instance." This third factor could be demonstrated by a school district's record of compliance. 503 U.S. at 491, 112 S. Ct. 1430.

There are literally hundreds of outstanding school-desegregation cases across the country. Yet, judging by the number of published opinions in these cases over the past ten years, most are languishing on court dockets. In those cases where the issue of terminating a desegregation decree has arisen, courts seem to overwhelmingly lean in favor of termination. See Wendy Parker, The Decline of Judicial Decision-Making: School Desegregation and District Court Judges, 81 N.C. L.Rev. 1623, 1639 (2003).

The Dowell/Freeman test requires a district court to examine not just process (compliance with the decree), but outcome (vestiges of de jure desegregation). This involves a fact-specific analysis, including inquiries into student-assignment procedures, faculty and staff assignment, transportation, facilities, extracurricular activities and quality of education. Here, the parties have expressly written into the consent decree that by December 31, 2005, the school district will have met this standard (CD IV ¶ 48-1).

The Supreme Court has not specifically defined what it means by "good faith compliance" or what constitutes a "vestige" of discrimination. District courts, by necessity, must have discretion in deciding whether termination is warranted. A recurring issue in these cases boils down to a question of causation: Are the current racial/ethnic disparities the result of past intentional discrimination? That is, are they "vestiges" of de jure discrimination?

Those circuit courts that have examined the issue have continued to leave the term largely undefined and failed to find any instances of "vestiges" of discrimination. See, e.g., Coalition to Save Our Children v. State Bd. of Educ., 90 F.3d 752 (3rd Cir. 1996) (within-school segregation in special-education and gifted/talented programs, decline in number of minority teachers and racially-identifiable extra-curricular activities are not vestiges of past discrimination); Wessmann v. Gittens, 160 F.3d 790 (1st Cir.1998) (achievement gap between black and Latino students versus white and Asian students, as well as lower teacher expectations for black and Latino students, are not vestiges of past discrimination); United States v. City of Yonkers, 181 F.3d 301 (2d Cir.1999) (lower teacher expectations for minority students and lack of updated curricula and teacher training are not vestiges of discrimination); Belk. v. Charlotte-Mecklenburg Bd. of Educ., 269 F.3d 305 (4th Cir.2001) (racially-imbalanced schools, less experienced teachers in minority schools and targeting of black students in student-discipline practices are not vestiges of discrimination).

These decisions are in keeping with the Supreme Court's recognition that present-day racial disparities may not be the result of past intentional discrimination. In Freeman, the Court excused segregated schools on the ground of purely private housing choices. 503 U.S. at 476, 490, 112 S. Ct. 1430. Likewise, in Missouri v. Jenkins, the Court found that racial disparities in achievement scores could be due to "numerous factors beyond the control of [the district]." 515 U.S. 70, 101, 115 S. Ct. 2038, 132 L. Ed. 2d 63 (1995).

In the present case, when the Ho plaintiffs appealed the Court's denial of their summary-judgment motion, the Ninth Circuit agreed that "vestiges" was the central issue. The opinion noted that "two issues remain for trial: Do vestiges remain of the racism that justified Paragraph 13 of the consent decree in 1983? Is Paragraph 13 necessary to remove the vestiges if they do remain?" The Ninth Circuit made clear that "the `vestiges' to be proved by it must be tied to the discriminating practices and policies that justified the consent decree fifteen years ago," and that "conclusory statements" are not sufficient evidence. Ho, 147 F.3d at 865. Given the centrality of proving "vestiges" to the outcome of trial, the district court required the parties to submit briefs arguing the appropriate definition of the term. Ultimately, Judge Orrick adopted the state's definition:

[A] vestige of segregation or discrimination . . . exists when (1) there exists a current condition in the school system that is likely to convey the message of racial inferiority that is implicit in a policy of segregation; and (2) that current condition is caused, at least in part, by the intentional racially discriminatory policies and practices by government officials that justified the Consent Decree at its inception.

Because the Ho litigation settled, this standard was never actually applied. But it seems to have been the last word on the subject in our circuit. See Ryan Tacorda, Acknowledging Those Stubborn Facts of History: The Vestiges of Segregation, 50 UCLA L.Rev. 1547, 1560 (2003).

Significantly, in its order approving the 1999 settlement, the Court acknowledged "the difficult causation issue" and noted that "[n]one of the parties to the litigation have been able thus far to demonstrate that the current problems in the SFUSD have been caused by the prior governmental discrimination that justified the adoption of the Consent Decree in 1983." 59 F. Supp. 2d 1021, 1038 (1999) (emphasis added). The 2001 settlement extended the date for termination of the consent decree from December 31, 2002, to December 31, 2005. All parties agreed that by December 31, 2005, the school district would have "taken all practicable actions to achieve unitary status, including eliminating any vestiges of past de jure racial or ethnic discrimination to the extent practicable and complying with the Consent Decree substantially and in good faith for a reasonable period" (CD IV ¶ 48-1). Proof to the contrary is not now before the Court, as discussed in detail below.

A. Good-Faith Compliance.

The first prong of the Dowell test examines whether or not the district has complied in good faith with the desegregation decree since it was entered. District courts have a great deal of discretion in determining "good faith compliance." In 1997, when this Court denied the Ho plaintiffs' summary-judgment motion, it cited the following areas of racial disparity where the district had failed to comply with the consent decree: special-education classes, student discipline, hiring and the achievement gap on standardized tests. Ho, 965 F.Supp. at 1326-27. Despite this evidence, there are also numerous examples of the district's good-faith efforts to comply with the consent decree in more recent years. For example, to address the achievement gap and growing resegregation in the school district, the district began converting some of its lowest-performing schools into "Dream Schools." To address the disproportionate representation of black students in special-education classes, the district has taken steps to improve its referral and assessment system, including requiring each school to have a positive behavior system in place (see Joint Response of All Parties to Issues Raised by Consent Decree Monitor's Apr. 11, 2005 Suppl. Report).

Perhaps the best indication of the district's good-faith compliance with the consent decree is that, since the 2001 settlement, no party has opposed any action taken by the school district. No one has filed a motion to enforce the consent decree or to hold the school district in contempt for failing to comply. Plaintiffs now belatedly assert that the school district has not yet taken all practical steps to achieve unitary status. Counsel argue that had such motions been brought, they would have been granted. This is not necessarily true. Unsubstantiated predictions of how the Court might have ruled on never-filed motions are insufficient to demonstrate that the district has not complied in good faith with the consent decree. This order finds that the district has plainly tried to comply in good faith.

B. Vestiges of Past De Jure Discrimination.

The second prong of the Dowell test requires a court to examine whether the vestiges of discrimination have been eliminated "to the extent practicable" in all areas of school operations, including staff, transportation, extra-curricular activities, facilities, student assignment and quality of education. Undoubtedly, the San Francisco Unified School District is resegregated today. This segregation can be measured both across schools and within individual school programs. Segregation alone, however, does not provide a federal court with the legal authority to continue enforcing a desegregation decree. Unless the current segregation is a "vestige" of past discrimination, a desegregation decree cannot be extended.

The key question is whether the current segregation is the result of any past intentional racial discrimination by government officials. See Ho v. San Francisco Unified Sch. Dist., 147 F.3d 854, 865 (9th Cir.1998); 59 F. Supp. 2d 1021, 1038 (N.D.Cal.1999). At no time in the entire twenty-seven-year history of this litigation has any party proven that the district was engaging or had ever engaged in intentional racial segregation.

Paragraph 8 of the original 1983 consent decree stated:

The parties stipulate and agree that, if proof were presented in formal proceedings, the Court would be justified in making factual findings and legal conclusions sufficient to require the system wide remedies that are set forth in this Consent Decree.

Stipulation aside, however, the Court's order noted that "no constitutional violations on [defendants'] part have been proved" and "[p]laintiffs' proof of discrimination, primarily against black students in the Western Addition, circa 1954 is weak in several respects." San Francisco NAACP, 576 F.Supp. at 45-48. Throughout the Ho litigation, the "vestiges" issue resurfaced. Each time, this Court and the Ninth Circuit have pointed out the weak or nonexistent evidence linking current segregation to past intentional segregation. See Ho, 147 F.3d at 865; San Francisco NAACP, 59 F.Supp.2d at 1038.

In sum, nothing in the lengthy record of the related SFNAACP and Ho cases indicates that any current segregation is causally linked to intentional segregation that existed in 1983. Nonetheless, it is worth examining some of the specific facets of school operations cited in Dowell for evidence of "vestiges" of past discrimination.

(1) Student Assignment Across Schools.

Assuming, arguendo, that the district did assign students to schools in an intentionally, racially discriminatory manner in the years preceding the original consent decree, it seems clear that the current segregation in San Francisco schools is not a "vestige" of this past discrimination. As a result of the 1983 consent decree and the ethnic-distribution requirement it imposed on the district, San Francisco successfully desegregated its schools. By 1997-98, only one district school enrolled more than fifty percent of a single race/ethnicity (see Siegel Report 16 (1998-99)).

Significantly, it was the Ho settlement in 1999, (and subsequent modifications to the consent decree), that prohibited the district from continuing to use race as a factor in student assignment. In the six years since then, the district has quickly become resegregated (see Biegel Suppl. Report (April 2005) App. 1). Thus, the current segregation across schools is not a vestige of past discrimination, but rather a result of the evolution of the consent decree itself. If anything, the revised student-assignment plan imposed by the consent decree (with its so-called diversity index) has been the main culprit.

(2) Student Assignment Within Schools: Special Education.

Whether the current segregation of students within schools—particularly in special-education classes—is a vestige of past intentional discrimination presents a more difficult question. Discrimination in special education has been extensively litigated in San Francisco. In Larry P. v. Riles, black students challenged the process, especially the use of I.Q. tests, used to place students in special classes for the educable mentally retarded. The district court held that the use of I.Q. tests which have a disparate impact on black students violated the Rehabilitation Act, the Education for All Handicapped Children Act, Title VI and the Fourteenth Amendment. 495 F. Supp. 926 (N.D.Cal.1979). The Ninth Circuit, however, reversed the finding of intentional discrimination. 793 F.2d 969 (9th Cir.1984). Larry P. supports a contention that the current segregation of special-education classes is related to the long history of placement policies with a discriminatory impact on black students. Yet, it fails to prove that the district or state engaged in past intentional discrimination in this area.

(3) Faculty Hiring.

It is in the area of faculty hiring that evidence of "vestiges" of past discrimination has some conceivable traction. In its 1983 order approving the consent decree, the Court noted that in the area of faculty and administrative hiring and placement, "plaintiffs were factually strong and likely to prevail at trial." 576 F. Supp. 34, 38-39. The most recent report of the consent-decree monitor reiterated that "the District has never come close to meeting the goals of Paragraphs 34-35," namely that staff at each school would reflect the student population of the entire school district (Biegel Report 22 Ohio App. 2 at 17). The following chart demonstrates the current disparities (see id. at 17-18):

% Certified                                    % Student
Staff         % Teachers    % Administrators   Population
Black             6.4           5.2          15.4               13.7
Latino            9.5           9.4          12.1               21.6
Filipino          2.8           2.8           2.5                6.2
Asian-American   20.9          20.7          21.7               42.8
White            50.9          52.0          42.9                9.3

Because the disproportionate underrepresentation of minority teachers has been an ongoing pattern since 1983, plaintiffs conceivably could have succeeded, had they attempted to prove that current disparities among faculty are vestiges of past discrimination in this area. But that link has not been proven on this record and the deadline for doing so has passed. In sum, this order finds that the available evidence suggests that any vestiges of de jure discrimination have been eliminated to the extent practicable.

C. Importance of Local Control Over Schools.

In laying out the standard by which federal courts could extricate themselves from their role supervising desegregation decrees, the Supreme Court has emphasized the value of local control of schools. In Dowell, Justice Rehnquist, writing for the majority, stated that injunctions in school-desegregation cases "are not intended to operate in perpetuity." He continued:

Dissolving a desegregation decree after the local authorities have operated in compliance with it for a reasonable period of time properly recognizes that "necessary concern for the important values of local control of public school systems dictates that a federal court's regulatory control of such systems not extend beyond the time required to remedy the effects of past intentional discrimination." 498 U.S. at 247-48, 111 S. Ct. 630 (citation omitted).

The Supreme Court echoed this sentiment in Freeman, noting that accountability to the public remains even after court supervision ends.

Returning schools to the control of local authorities at the earliest practicable date is essential to restore their true accountability in our governmental system. When the school district and all state entities participating with it in operating the schools make decisions in the absence of judicial supervision, they can be held accountable to the citizenry, to the political process, and to the courts in the ordinary course. 503 U.S. at 490, 112 S. Ct. 1430.

The Supreme Court's continued emphasis on the value of local control over schools militates in favor of terminating the decree in this case. After twenty-two years of judicial supervision and substantial compliance by the school district, it is entirely appropriate to return full control of the district to the school board and superintendent. These professionals are in the best position to make educational policy decisions. The normal political process should be sufficient to hold the school district accountable. San Francisco, it can be safely added, is a beacon for human rights. It would be one of the last places in America where racist school policies would be expected.

3. THE RUFO STANDARD.

In the alternative, Rufo v. Inmates of Suffolk County Jail, 502 U.S. 367, 112 S. Ct. 748, 116 L. Ed. 2d 867 (1992), decided one year after Dowell, established a two-part test for modifying consent decrees in institutional-reform cases. The party seeking modification "bears the burden of establishing that a significant change in circumstances warrants revision of the decree." If this burden is met, the district court must consider "whether the proposed modification is suitably tailored to the changed circumstance." "Changed circumstances" include a significant change in factual circumstances or in law. If "changed factual conditions make compliance with the decree substantially more onerous" or "a decree proves to be unworkable because of unforeseen obstacles," modification may be appropriate. However, "modification should not be granted where a party relies upon events that actually were anticipated at the time it entered into a decree." The Supreme Court noted that its decision in Rufo, providing a more flexible standard for relieving a party from a final decree than had previously been applied, was consistent with its decision in Dowell, which provided a flexible standard for terminating a desegregation decree. Id. at 383-85, 112 S. Ct. 748.

It is somewhat unclear whether the Rufo standard or the Dowell/Freeman standard should be applied to a motion to extend (i.e., modify the termination date of) a consent decree. The Ninth Circuit has not explicitly addressed this issue. The Fourth Circuit, regarding a motion to terminate a consent decree involving state welfare programs, has stated that "[t]he standards employed in Dowell and Rufo are but variations on a single theme." Alexander v. Britt, 89 F.3d 194, 197 (4th Cir.1996). The Sixth Circuit, on the other hand, has more explicitly stated that Rufo is the appropriate standard when deciding whether to modify a consent decree entered in a school-desegregation case. Lorain NAACP v. Lorain Bd. of Educ., 979 F.2d 1141, 1149 (6th Cir.1992). Likewise, the Eleventh Circuit, in a class-action employment-discrimination case, ordered the district court to apply the Rufo standard to modification motions and the Dowell standard to termination motions. United States v. City of Miami, 2 F.3d 1497, 1503-09 (11th Cir.1993).

Judge Orrick previously suggested that Rufo might be the applicable standard in his 1997 decision denying the Ho plaintiffs' motion for summary judgment. 965 F. Supp. at 1326 ("As plaintiffs by their motion seek to dissolve or terminate the entire Consent Decree rather than to modify it, the Court does not address at this time the propriety of modification under Rufo v. Inmates of Suffolk County Jail."). Likewise, in its decision approving the 2001 settlement, Judge Orrick discussed the strength of the district's argument to extend the consent decree under the Rufo standard. He referred to those "changed circumstances" that bolstered the district's case, including the resignation of the superintendent, the district's financial mismanagement troubles, unforeseen litigation and the change of lead counsel. San Francisco NAACP v. San Francisco Unified Sch. Dist., 2001 WL 1922333, *6.

It is true that after the proposal was tendered, Superintendent Ackerman announced her resignation effective June 30, 2006. And, the teachers union has recently filed suit, challenging the school district's hiring practices for the new Dream Schools. But these circumstances will not make compliance, at least for the few remaining months of the consent decree, "substantially more onerous" or "unworkable." Indeed, Superintendent Ackerman's resignation is not effective until June 30, 2006, after the consent decree is currently scheduled to expire. The Rufo standard for modifying a consent decree has not been met.

4. THE STUDENT-ASSIGNMENT SYSTEM AND DREAM SCHOOLS.

The parties have failed to meet their burden of demonstrating that an extension of the consent decree is legally justified. More than that, the Court is convinced that an extension will do more harm than good. The current student-assignment system is not working, placing an unnecessary burden on families in San Francisco. Oddly, the proposed modification to the decree would bar any adjustment to the student-assignment method before October 15, 2006. In essence, the school district would be forced to continue using the diversity index for the 2006-07 school year and possibly beyond.

The current student-assignment system based on the diversity index has not only failed to accomplish its desired purpose, but has in fact allowed, if not caused, resegregation of the school district. At the hearing, counsel for SFNAACP claimed that the diversity index itself was not responsible for the resegregation of underperforming schools because it only affected student assignment at over-subscribed (i.e., more desirable), not under-subscribed (i.e., less desirable), schools. This argument fails to explain why the original consent decree was so successful in producing racially integrated schools, but the current system is failing. It would be wrong to compel the continued use of a flawed student-assignment system, one that at best is a vain exercise and at worst actually promotes racial segregation. It is no answer to say that more studies will be done in the future. This Court has pleaded with the parties to fix the diversity index. They have not. This order declines to perpetuate a flawed, court-ordered student-assignment system.

* * * * * *

The Dream Schools initiative may be an important step towards closing the achievement gap. Improving academically underperforming schools may even indirectly help achieve diversity, by making those schools more attractive to children of all racial/ethnic groups. The school district is prepared to go forward with its plans for Dream Schools with or without an extension of the consent decree.

This order stresses that the Dream Schools initiative was an idea that began with the school district itself. It was subsequently incorporated into the decree in mid-2004 via a stipulated amendment that made the initiative mandatory under the decree. The Court routinely approved the stipulation without a hearing. In hindsight, perhaps this was unwise. At least one effect of writing the initiative into the decree was to strengthen the hand of the district against teachers, parents and others not joined in this litigation. When the teachers union (and others) later complained over the details of implementation, the district answered that the Dream Schools were compulsory under the decree. Litigation in state court over the displacement of teachers at newly-converted Dream Schools, for example, was removed to federal court by the school district. Similarly, when a family sued in state court over their son's assignment to a particular school, the district removed the case to this Court on the grounds that the lawsuit challenged the consent decree. Johnson v. San Francisco Unified Sch. Dist., No. C 05-00245 WHA (N.D. Cal. removed Jan. 18, 2005).

Once the initiative was built into the decree, the Court felt obliged to monitor its implementation. Regular hearings were held. The school district was required to meet various self-imposed deadlines, each of which was jointly requested by stipulation. The school district, however, fell further and further behind its own schedule. It frantically tried to ready the Dream Schools in time for the annual parent-student selection process. The result: When the parents were called upon to make their decisions, many teachers and many courses were not yet in place at the designated Dream Schools. Family decisions had to be based on vague, skeletal themes. Moreover, the teachers union claims that the teacher-selection process was too compressed. Wildly erratic hiring led to unfair replacement of many experienced teachers, it claims. Perhaps if the Dream Schools had not been incorporated as part of a mandatory decree, the district might have done a better job by focusing its energies on converting one school at a time.

The Court now finds itself being blamed for the district's rushed implementation of the initiative. Despite the multiple hearings and intermediate deadlines established to effect the Dream Schools, it is clear that no federal judge is as well-equipped as school professionals to make the educational policy decisions involved. Now, counsel wish to do it again and build yet more mandatory Dream Schools into the decree (while locking in the same discredited student-assignment system for another year).

The decree has drifted too far. The modus operandi for this case has evolved to this: Counsel broker a deal for their clients. Then they write it into the decree. Then they use the supremacy of the decree to override any opposition from parents, teachers and other interests not represented in this suit. The checks and balances of the traditional governmental and political processes are thus short-circuited. Counsel and their clients have become a kind of supreme council able to fast-track educational initiatives of their own choosing and to bypass concerns of those unrepresented herein. This would be acceptable were it remedying the vestiges of past de jure segregation. But that is not so. Instead, the rationale for the initiatives has drifted to doing "what is best" for the children of San Francisco. Excellent as it sounds, this rhetoric has no anchor in the original source of the district court's power to intrude upon local government. What is best for the children of San Francisco should be left to the professionals in the district, subject to the voices of all in the community, whether or not they have a seat at counsel table.

In short, since the settlement of the Ho litigation, the consent decree has proven to be ineffective, if not counterproductive, in achieving diversity in San Francisco public schools. De jure segregation has never been proven, much less vestiges thereof. As the decree has come to be used, the Court must pretend to supervise decisions better left in the hands of education professionals subject to the rough and tumble of local politics and government. "Returning schools to the control of local authorities at the earliest practicable date is essential to restore their true accountability in our governmental system." Freeman, 503 U.S. at 490, 112 S. Ct. 1430. The time has come to do so for San Francisco.

CONCLUSION

For the foregoing reasons, the extension of the consent decree is DENIED. As previously stipulated, the consent decree shall expire on December 31, 2005.

IT IS SO ORDERED.

NOTES

[1] In truth, the history of this litigation arguably dates back even earlier, to another lawsuit filed. in 1969. Johnson v. San Francisco Unified Sch. Dist., 339 F. Supp. 1315 (N.D.Cal. 1971), vacated and remanded, 500 F.2d 349 (9th Cir.1974). The Ninth Circuit reversed the finding of de jure segregation, as intervening law had clarified the legal standard, and remanded to determine whether defendants possessed the requisite intent to segregate. A second action, filed while the appeal was pending, was voluntarily dismissed in 1976. O'Neil v. San Francisco Unified Sch. Dist., No. C-72-0808 RFP (N.D. Cal. filed May 5, 1972). Because the Johnson plaintiffs made no effort to prove de jure segregation on remand, on June 22, 1978, the court sua sponte dismissed that complaint without prejudice, for mootness and failure to prosecute.

[2] No summary of this litigation would be complete without recognizing the judge who presided over this case for much of its history. From 1978 until his retirement in 2002, the late Judge William Orrick oversaw both the original SFNAACP litigation and the implementation of the consent decree. Under his vigilant watch, the case resulted in a successful settlement and consent decree, and the San Francisco schools became integrated. Years later, he also presided over the related action involving the Ho plaintiffs. In his last year on the bench, Judge Orrick acquiesced in the 2001 settlement in these actions (and amendments thereto), which resulted in the current student-assignment plan.

[3] "Reconstitution" is a drastic method of educational reform whereby an existing school is recreated from scratch. While the building and the students remain the same, the entire faculty and administration are replaced with a new staff. Reconstitution in San Francisco also involved (i) the adoption of eleven "Philosophical Tenets" to establish expectations for learning and behavior; (ii) determination of specific student outcomes for each grade; (iii) advancement of available instructional technology; (iv) increases in adult-student ratios; (v) increases in staff development to implement these components; (vi) selection by staff of effective, unique instructional tools; and (vii) encouragement of parental involvement. See Kelly C. Rozmus, Education Reform and Education Quality: Is Reconstitution the Answer?, 1998 B.Y.U. Educ. & L.J. 103, 113-14 (1998).

[4] The longitudinal reports tracked the progress of the same students over time for each targeted school.

[5] These demographic trends continue to the present day. The Chinese-American student population has increased from 19.5 percent in 1983 to 32.1 percent in 2005 and the Latino student population has increased from 17.2 percent in 1983 to 21.6 percent in 2005. In contrast, the black student population has declined from 23.1 percent to 13.7 percent during this time, and the white population has declined from 16.9 percent to 9.3 percent. Numbers for other racial and ethnic groups have remained relatively stable since 1983 (Biegel Report No. 22 (2004-05) App. 2 at 2).

[6] The consent-decree advisory committee was appointed by the Court in 1993 to "provid[e] a neutral forum for discussion of suggestions of the State's monitor and advis[e] the Court from an educational policy perspective on the progress towards achieving the Consent Decree goals." See San Francisco NAACP v. San Francisco Unified Sch. Dist., 1993 WL 299365 at *2 (N.D.Cal.1993).

[7] The Ninth Circuit has recently held that the use of a race-based "tie-breaker" in a studentassignment system used by Seattle public schools does not violate the Equal Protection Clause. Parents Involved in Cmty. Schs. v. Seattle Sch. Dist. No. 1, 426 F.3d 1162 (9th Cir.2005) (en banc). In theory, a similar student-assignment system could be implemented in San Francisco. If the consent decree were extended, however, the terms of the Ho settlement would prevent the school district from using race or ethnicity as a factor at all, even as a tie-breaker. Also, the proposed extension would not allow any changes to the student-assignment system before the 2007-08 school year.

[8] Attachment B stated that "the District will be proposing by August 15, 2001, in accordance with the terms of the Stipulation, diversity index criteria related to `Geographic Areas'" (Amended Stipulation and Order Re Modification and Termination of Consent Decree, Attachment B, n. 1, July 12, 2001). In a Joint Stipulation dated December 4, 2002, the school district agreed to consider using geographic zones and state its conclusions by August 15, 2003. The NAACP disputed the district's compliance with this promise, but no contempt or compliance motion was ever filed by plaintiffs or anyone else. To this date, the diversity index still does not take into account "Geographic Areas." See SFUSD Enrollment Guide for 2005-2006 Enrollment Period at 16.

[9] "Severely resegregated" was defined as sixty percent of the student population of a single race/ethnicity.

[10] While this admittedly would not be aimed at remedying de jure segregation, "in a consent decree, defendants may agree, within limits, to do more than a judicially imposed injunction could have required." Alexander v. Britt, 89 F.3d 194, 200 (4th Cir.1996).

[11] Ms. O'Keeffe testified that a profile for each student is created by assigning a "1" or "0" for each of six binary factors: socioeconomic status, academic achievement status, mother's educational background, language status, academic performance index and home language. The diversity score for each factor is calculated as the sum of the squared proportions of "1s" and "0s" subtracted from one; the optimal score, representing perfect diversity, would be "0.5." The composite diversity score is calculated by averaging the diversity scores for each of the six factors.

[12] This language has been modified to read: ". . . the decision to use this assignment method was not taken at the demand of any other party in the desegregation consent decree lawsuits (Suppl.Br. 4, fn. 1)."