This case pits a community's effort to send 22 children to the school that will best serve their educational needs1 against the district court's interpretation of a 27-year-old statewide school desegregation decree. At issue is whether the injunctive decree prohibits a neighborhood subdivision from changing school districts, even though the move would only alter the racial composition of the district from which the subdivision is detached by approximately 2.7 percent.

Because we conclude that the proposed boundary change does not violate the United States Constitution or the district court's desegregation order, properly construed, we REVERSE.

Since 1971, the Texas public education system has been governed according to a federal court order2 designed to ensure that "no child will be effectively denied equal opportunity to educational opportunities on account of race, color or national origin." The modified order provides, inter alia, that the state of Texas, the Texas Education Agency ("TEA"), its officers, agents, and employees

shall not permit, make arrangements for, approve, acquiesce in, or give support of any kind to changes in school district boundary lines--whether by detachment, annexation, or consolidation of districts in whole or in part--which are designed to, or do in fact, create, maintain, reinforce, renew, or encourage a dual system based on race, color, or national origin.

According to the modified order, the board of trustees of any school district desiring to annex or consolidate with a nearby district must notify the Texas Commissioner of Education ("Commissioner") of its intentions. The Commissioner is required to investigate "the effects of such a projected change of boundaries on the desegregation status of all the school districts concerned." The Commissioner must then report the results of his investigation to the appropriate county and local officials, stating whether the proposed change is in violation of law.

In January 1991, the Forest Springs subdivision, which lies on the boundary of the Goodrich and Livingston Independent School districts, sought detachment from Goodrich and annexation to Livingston. If approved at that time, the detachment would have removed approximately 380 acres and twenty-two students (twenty of whom are white) from Goodrich. After notice and public hearing, the Livingston I.S.D. school board unanimously approved the proposed annexation. Goodrich, however, disapproved the detachment request, and the Forest Springs subdivision appealed to the Commissioner.

Following a hearing on the merits, the Commissioner found that the proposed detachment and annexation neither violated modified order 5281 nor would impose significant adverse educational, economic, or social effects on the students in the affected territories.3 See Forest Springs Subdivision v. Goodrich Indep. Sch. Dist., No. 240-R6-391 (Comm'r Educ. Sept. 1992). The Commissioner made pertinent findings of fact and conclusions of law:

....

12. [Forest Springs's] detachment from Goodrich I.S.D. will not result in racial imbalance; however, it does change the majority and minority percentage of [Goodrich I.S.D.'s] school population, based on average daily attendance, by more than one percent.

....

14. There is no evidence that the detachment of Petitioner Forest Springs Subdivision from Respondent Goodrich I.S.D. is designed to or does in fact, create, maintain, reinforce, renew, or encourage a dual school system based on race, color, or national origin.

15. There is no evidence that the annexation of Petitioner Forest Springs Subdivision from Respondent Livingston I.S.D. is designed to or does in fact, create, maintain, reinforce, renew, or encourage a dual school system based on race, color, or national origin.

....

....

2. The change of boundaries resulting from detachment of the affected territory from Respondent Goodrich I.S.D. and the annexation of Forest Springs Subdivision, Petitioner, to Livingston I.S.D. does not violate [modified order] 5281.

3. Detachment of the affected territory from Respondent and annexation of that territory to Livingston I.S.D. will have no significant adverse educational, economic, or social effect on the school districts or students involved.

Id. The Commissioner ordered that Forest Springs be annexed to Livingston. See id.

Goodrich sought judicial review of the Commissioner's decision. A state district court reversed the order and remanded the cause to the Commissioner to enter a new order denying the detachment and annexation. But the Austin Court of Appeals reversed the district court. See Texas Educ. Agency v. Goodrich Indep. Sch. Dist., 898 S.W.2d 954, 956 (Tex.App.-Austin 1995, writ denied). The appeals court reasoned that due deference must be accorded the Commissioner's judgment. In Texas, an agency's determination must be affirmed where it is supported by substantial evidence. Applying that test, the appeals court concluded that "[t]he Commissioner's finding that the proposed boundary change would not create, maintain, reinforce, renew, or encourage a dual school system based on race, color or national origin is supported by substantial evidence" and rendered judgment affirming the Commissioner's decision. Id. at 959.

After the Texas Supreme Court denied its application for writ of error, Goodrich filed a Motion to Intervene in Civil Action 5281, complaining now to the federal court that the proposed boundary change violated the court's modified order and requesting an injunction. Livingston I.S.D., its Board of Trustees, and individual residents of the Forest Springs subdivision who had signed the original petition for detachment and annexation were joined as Defendants-Intervenors.

The district court held a hearing on August 22, 1996, in which evidence was introduced--including the entire administrative record of the proceedings before the TEA--and testimony of lay and expert witnesses was taken. Participation by the State of Texas and the United States, the named parties to the original action, was minimal. Although Goodrich's motion for injunctive relief was denied, the district court proceeded to issue a declaratory judgment, holding:

After careful consideration of the evidence presented to this court in this proceeding, it is determined that the proposed detachment [of] Forest Springs subdivision from Goodrich and annexation to Livingston Independent School District would reinforce, renew and encourage a dual school system based on race and color, in violation of this court's order of July 17, 1971.

The residents of Forest Springs filed this timely appeal.4 They argue that the district court erred in (1) finding that under principles of issue preclusion the state court proceedings should not bar Goodrich's subsequent intervention in federal court; (2) finding that the proposed detachment and annexation would reinforce, renew or encourage a dual school system; and (3) refusing to apply present standards governing school desegregation law, rather than those set out in the modified order. Neither the state of Texas nor the TEA has appealed, but both the United States and Goodrich have filed briefs as appellees.

Before we consider Forest Springs's issues, the Appellees contend, we must decide its citizens' constitutional standing as intervenors to proceed with this appeal. See Diamond v. Charles, 476 U.S. 54, 68, 106 S. Ct. 1697, 1706, 90 L. Ed. 2d 48 (1986). The Residents must establish (1) that they have suffered an "injury in fact," (2) which is fairly traceable to the challenged action, and (3) which will be redressed by a favorable decision of this court. See American Forest & Paper Ass'n v. EPA, 137 F.3d 291, 296 (5th Cir.1998).

Because the Residents' claims so easily fulfill "the irreducible constitutional minimum of standing," Lujan v. Defenders of Wildlife, 504 U.S. 555, 560, 112 S. Ct. 2130, 2136, 119 L. Ed. 2d 351 (1992), it is a shame that the government resorts to specious arguments to deny standing. Following a favorable ruling from the state appellate courts, the Forest Springs subdivision won the right to be detached from Goodrich and become part of Livingston I.S.D. For the Residents of Forest Springs this victory entitled parents to send their children to a public school system many felt had "better educational and extracurricular opportunities." Texas Educ. Agency, 898 S.W.2d at 959. The decision also allowed these citizens to support Livingston I.S.D. with their tax dollars and to participate as members of the local community. When Goodrich intervened in federal court and secured a favorable declaratory judgment overturning the annexation, all that was lost.

The Appellees contend, however, that because neither the state nor the TEA has appealed, the Residents alone lack standing. Appellees cite Diamond in support of their position, but their reliance is misplaced. Diamond was a pediatrician engaged in private practice in Illinois. After the federal court of appeals declared an Illinois abortion law unconstitutional and the State's attorney general elected not to defend the law any further, Diamond appealed to the Supreme Court. The Court held that Diamond was unable to satisfy the constitutional requirements of standing because private parties have no judicially cognizable interest in the prosecution or nonprosecution of state laws. The Court reasoned that a private citizen cannot compel a state to apply its laws with a particular degree of vigor, nor must the state enact a statute in accord with his wishes. See Diamond, 476 U.S. at 65, 106 S.Ct. at 1705. As the Court concluded, "[b]ecause the State alone is entitled to create a legal code, only the State has the kind of 'direct stake' ... in defending the standards embodied in that code." Id. at 65, 106 S. Ct. at 1705.

Unlike the appellant in Diamond, the residents of Forest Springs do not pursue this appeal in order to compel the state to enforce its laws, and they are not trying to defend the constitutionality of a legislative judgment in the absence of a state representative. Instead, the Residents are defending a private right they gained in state court to annex their subdivision to and send their children to school in Livingston I.S.D. These Residents are not merely "concerned bystanders" invoking the exercise of judicial power "simply as a 'vehicle for the vindication of value interests,' " id. at 62, 106 S.Ct. at 1703 (quoting United States v. SCRAP, 412 U.S. 669, 687, 93 S. Ct. 2405, 2416, 37 L. Ed. 2d 254 (1973)). They are concerned parents with a direct stake in the quality of education they are able to secure for their children and, hence, in the outcome of this appeal.

Moreover, a favorable decision from this court would directly redress the harm alleged in this case. The Commissioner's order required that Forest Springs be annexed to Livingston I.S.D. An order of the Commissioner remains in effect until it is modified or set aside by the courts. See Temple Indep. Sch. Dist. v. Proctor, 97 S.W.2d 1047, 1052 (Tex.Civ.App.-Austin 1936, writ ref'd). A favorable ruling from this court would remove the last obstacle preventing Forest Springs's detachment from Goodrich and annexation to Livingston. Although a new date for the annexation would be required, state law provides that an appealed decision of the Commissioner becomes effective on a date set by the district court in Travis County. See Tex. Educ.Code Ann. § 13.005(c) (West 1996). Neither the state of Texas nor the Commissioner is required to take any action to effectuate the Commissioner's earlier order. As a result, the redressability prong as well as the other standards in the Constitution's Article III standing requirement is satisfied.

The Residents argue first that the district court erred in refusing to apply preclusive effect to the final state court judgment, which affirmed the Commissioner's finding that the proposed detachment and annexation would not violate Modified Order 5281.

Common-law doctrines of issue and claim preclusion ordinarily shield final judicial determinations of administrative bodies, whether under the aegis of federal or state government. See Astoria Fed. Sav. & Loan Ass'n v. Solimino, 501 U.S. 104, 107-08, 111 S. Ct. 2166, 2169, 115 L. Ed. 2d 96 (1991). Federal courts must give a State agency's decision "the same preclusive effect to which it would be entitled in the State's courts." University of Tennessee v. Elliott, 478 U.S. 788, 799, 106 S. Ct. 3220, 3226, 92 L. Ed. 2d 635 (1986).

Under Texas law, a party seeking to invoke issue preclusion must establish: (1) the facts sought to be litigated in the second action were fully and fairly litigated in the prior action; (2) those facts were essential to the judgment in the first action; and (3) the parties were cast as adversaries in the first action. See Bonniwell v. Beech Aircraft Corp., 663 S.W.2d 816, 818 (Tex.1984). Judgments and decisions of state administrative agencies are entitled to preclusive effect where the administrative agency has acted in adjudicatory, judicial or quasi-judicial capacity. See Railroad Comm'n v. Phillips, 364 S.W.2d 408, 411 (Tex.Civ.App.-Austin 1963, no writ); see also Muckelroy v. Richardson Indep. Sch. Dist., 884 S.W.2d 825, 830 (Tex.App.-Dallas 1994, writ denied).

In this case, the first two requirements for issue preclusion clearly apply. The parties' dispute was litigated to the hilt in state administrative and judicial proceedings, including the issue whether the proposed boundary change violated "[modified order] 5281 or result[ed] in significant adverse educational, economic or social effects, or whether the boundary changes [were] in the educational interests of the students." Forest Springs Subdivision v. Goodrich Indep. Sch. Dist., No. 240-R6-391 (Comm'r Educ. Sept. 1992).

The appellees urge, and the district court erroneously held, that issue preclusion is inapplicable because the state courts did not rule definitively on whether the proposed boundary change complied with modified order 5281. Technically, as appellees observe, the only issue before the state courts was whether the Commissioner's findings were supported by substantial evidence.5 But this is all that issue preclusion requires in the administrative context. A state court reviewing the decision of an administrative agency need not proceed as though it were writing on tabula rasa. See Kremer v. Chemical Constr. Corp., 456 U.S. 461, 481 n. 21, 102 S. Ct. 1883, 1897 n. 21, 72 L. Ed. 2d 262 (1982).

The district court also rejected issue preclusion arising from the state proceedings with a puzzling statement that

if the parties had sought to relitigate in state court the issues that had been decided by the modified order, this court could have been moved to enjoin the state court proceedings in order to "protect or effectuate" its judgment. 22 U.S.C. § 2283.

This statement may be correct, but it is quite beside the point. The parties never sought to relitigate issues in state court that had been decided by the modified order. On the contrary, the administrative proceeding before the Commissioner followed the express directive of the modified order, which required the Commissioner to investigate and issue findings concerning how any proposed school boundary change affects the schools' desegregation status and whether it comports with "the law," presumably including modified order 5281. If there is a procedural anomaly in this approach, it was created by the modified order and not by the Commissioner or the parties.

But while the state administrative proceedings would generate issue preclusion among their participants, Forest Springs and Goodrich, this cannot occur as to the United States, which was not a party to the state proceedings. Normally, a judgment or decree cannot bind strangers to the litigation. See Martin v. Wilks, 490 U.S. 755, 762, 109 S. Ct. 2180, 2184, 104 L. Ed. 2d 835 (1989). Although there are some exceptions to this rule,6 the United States is rarely "barred from independent litigation by the failure of a private plaintiff." United States v. East Baton Rouge Parish Sch. Bd., 594 F.2d 56, 58 (5th Cir.1979); but see Tyus v. Schoemehl, 93 F.3d 449, 456 (8th Cir.1996). Barring a showing, not made here, that the government maintained a "laboring oar" in the state court litigation, Montana v. United States, 440 U.S. 147, 155, 99 S. Ct. 970, 974, 59 L. Ed. 2d 210 (1979) (quoting Drummond v. United States, 324 U.S. 316, 318, 65 S. Ct. 659, 660, 89 L. Ed. 969 (1945)), preclusion doctrine is inapplicable to the United States.7

The district court rejected every assertion of the Forest Springs residents supporting their proposed annexation to Livingston I.S.D. The court found as matters of fact that if the annexation were approved, it (1) would limit Goodrich's financial ability to run a school district, (2) would create a perception that race is a relevant factor in establishing district boundaries, and (3) would likely trigger further annexations and increased school segregation. The court interpreted its modified order to incorporate a 1% "guideline" for evaluating the resegregative impact of boundary changes on the racial makeup of affected school districts. Finally, it refused to read the 27-year-old modified order in light of more recent Supreme Court decisions concerning school desegregation decrees. Portentous significance must flow, in the district court's view, from the detachment of a few hundred acres and 22 students from Goodrich.

Unlike the district court, we believe that much less is involved in this minuscule boundary change. First, it does not contravene even the court's own modified order, properly construed. Second, the district court's factual findings are either clearly erroneous or too attenuated to demonstrate that the proposed boundary change will result in the creation of a segregated Goodrich school district. Third, the 27-year-old modified order could and should be interpreted with an eye toward significant intervening Supreme Court precedent. Each of these points requires elaboration.8

Modified order 5281 forbids boundary line changes

which are designed to, or do in fact, create, maintain, reinforce, renew, or encourage a dual school system based on race, color, or national origin.

The court found that the Forest Springs boundary change would reinforce, renew and encourage a racially dual system. Critical to, though not dispositive of, the court's analysis was its adaptation of a 1% guideline from a separate portion of the decree dealing only with student transfers. The other provision of the order forbids defendants from approving student transfers

where the effect of such transfers will change the majority or minority percentage of the school population ... by more than one percent (1%) in either the home or the receiving district....

Amendments to modified order 5281, August 9, 1973, at 2. The court considered this percentage "a useful rule of thumb" for evaluating the proposed Forest Springs and future boundary changes. According to the court's arithmetic, the Goodrich school-age population would be 2.7% less white if the boundary change occurred.9 Inferentially, the court found that because a 2.7% shift of children is greater than the 1% "rule of thumb," the boundary change presumptively violated modified order 5281. This reasoning is seriously flawed.

The court offers no basis, legal or factual, for its decision to import the 1% transfer standard into the boundary change portion of the decree. The goal of affording notice to parties affected by the decree is wholly ignored by this ad hoc rewriting. It matters not, and the court did not rely on the fact, that apparently the Commissioner has also informally followed a 1% rule on boundary changes. The decree simply has no such limit, and common sense suggests why, even if a 1% rule was justifiable for student transfers, the decree distinguished between student transfers and boundary changes. Student transfers are much easier for school districts to implement and would have afforded a convenient subterfuge for parties bent on undermining desegregation efforts. Boundary changes, on the other hand, are permanent, and they irrevocably affect district population, tax base, size and allocation of resources. Viewing boundary changes solely from the perspective of a tiny transient percentage change in school racial makeup is too rough a measure of their significance. See Texas Educ. Agency, 898 S.W.2d at 957-58.

That the court characterized its new 1% standard as a rule of thumb, not a rigid benchmark, does not ameliorate the problem, for the decree effectively guides day-to-day school administrative decisions. Cautious school officials, not to mention unhappy homeowners, will opt not to buy a lawsuit by proposing boundary changes that challenge this low threshold. One look at the protracted seven-year battle waged by Goodrich against the Forest Springs residents proves the litigation potential inherent in a vague and anachronistically construed remedial order.

The 1% rule of thumb is also a static statistic with little analytical power for measuring segregative population changes. The district court recognized as much by citing its own earlier order, which stated that the "one percent guideline ... cannot stand as a benchmark by which to rigidly approve or deny transfers." See order re Intervention of Mary Hightower, C.A. 5281 (Sept. 8, 1980) (permitting student transfers that would increase the white population of the transferor school by 7%).10 In the Hightower order, the district court bowed to governing Fifth Circuit law, which had already come to reject purely quantitative analysis in the desegregation context. See Davis v. Board of School Comm'rs of Mobile County, 393 F.2d 690, 693 (5th Cir.1968). The district court recognized both in Hightower and in this case that the proper inquiry is whether the proposed boundary change would undermine the effectiveness of its plan to remedy specifically identified constitutional violations.11 See, e.g., Lee v. Eufaula City Bd. of Educ., 573 F.2d 229, 232 (5th Cir.1978) ("In measuring the cumulative effect of a student transfer program on desegregation, the Court must do so from a qualitative viewpoint, without blind deference to an objective mathematical formula."); see also United States v. Lowndes County Bd. of Educ., 878 F.2d 1301, 1307-08 (11th Cir.1989).

Finally, the 1% guideline has absurdly little to do with the facts of this case. Goodrich experienced an annual student "mobility rate" of about 35%, which means that the population turnover was dramatic. The population fluctuated significantly both up and down, and an apparent decline in the white student population over several years is attributable to a doubling of Hispanic students. Amid these seismic shocks, a 1% guideline, or even an actual 2.7% student population change, is barely a tremor.12

Not only is the 1% rule of thumb too small to be probative of anything, it is, standing alone, legally insufficient to evaluate compliance with the court's modified order. Moreover, the 1% "guideline" is not and may not be treated as if it were a part of the boundary change portion of the modified order. The court erred by relying on this "rule of thumb" for any purpose concerning the proposed boundary change.

In addition to the "violation" of the newly-minted 1% guideline, the district court found that the boundary change would violate modified order 5281 because it would "reinforce, renew, and encourage a dual school system based on race and color, in violation of this court's order of July 13, 1971." The district court relied on findings that the proposed annexation would impair Goodrich's ability to run an integrated school district; would likely trigger further annexations and increased school desegregation; and would reinforce the perception that school district boundaries were drawn on account of race.

The district court first found that the money the district would lose in the wake of Forest Springs's detachment would adversely affect the quality of the teaching, the availability of resources (e.g., computers), and the type of programs Goodrich would be able to offer.

The court's sub silentio overruling of the Commissioner's finding of no adverse financial impact on Goodrich is troubling. The Commissioner specifically evaluated the pertinent state law issue whether Forest Springs's detachment would reduce Goodrich's tax base by a ratio at least twice as large as the ratio by which it would reduce student population. See Texas Educ.Code Ann. § 19.022(d) (West 1991). It would not, and the parties so stipulated, yet the federal court embarked on its own fact-finding without tether to the standard that Texas law provides.

The district court's result is also doubtful standing alone. According to unchallenged testimony elicited during trial, state and local ad valorem taxes each contribute approximately one-half of the school district's funding. Using the methodology the district court followed,13 the detachment of Forest Springs would result in a loss of approximately 22 students. Goodrich spends approximately $5,300 per student. Accordingly, the loss of the Forest Springs students could reduce Goodrich's expenditures by $116,600. If one-half of this amount were supplied by the state, the remaining $58,300 must have been financed through local ad valorem taxes. However, Forest Springs's ad valorem contribution to Goodrich I.S.D.'s coffers was only about $52,000, almost $6,300 less than the district spent to educate Forest Springs's children. From this evidence, the district court should have concluded that Goodrich I.S.D. would actually benefit from the proposed detachment. In other words, when the boundary change is analyzed with proper regard for cost savings, not merely revenues lost, the resources available to Goodrich to spend on individual pupils would rise following the detachment of Forest Springs. The evidence does not support the district court's finding that the quality of education within Goodrich I.S.D., as measured by money spent, would be adversely affected merely by the loss of Forest Springs's revenue.14

Similarly problematic is the district court's finding that approval of the boundary change would "trigger further annexations and increased school segregation." Relying on the opinion of an expert witness, the district court reasoned that the departure of Forest Springs would push the white population one step closer to minority status within Goodrich I.S.D. According to a phenomenon known as "tipping," as whites approach minority status "white flight" ensues. Thus, Goodrich's loss of Forest Springs would accelerate the departure of other white families and would " 'certainly' make other subdivisions more likely to seek similar detachments and annexations." Neither the law nor the evidence in this case supports such a finding.

Although the district court's authority over the Texas educational system under the venerable desegregation decree is quite broad, the Supreme Court has reminded us that " 'there are limits' beyond which a court may not go in seeking to dismantle a dual school system." Pasadena City Bd. of Educ. v. Spangler, 427 U.S. 424, 434, 96 S. Ct. 2697, 2704, 49 L. Ed. 2d 599 (1976) (quoting Swann, 402 U.S. at 28, 91 S.Ct. at 1282). The modified order reaches the reinforcement, renewal or encouragement of a dual school system, and it binds public officials to prevent racial segregation. The key to the order's scope is state action. Absent any showing that school authorities "have in some manner caused unconstitutional segregation," the district court lacks any power to enjoin a change in school boundaries. Spangler, 427 U.S. at 434, 96 S.Ct. at 2704; see also Swann, 402 U.S. at 28, 91 S.Ct. at 1282.

It hardly needs to be added that there is no evidence that the demographic change forecast to occur in Goodrich I.S.D. in the wake of Forest Springs's detachment is in any way attributable to "segregative actions chargeable to the [State]." Spangler, 427 U.S. at 435, 96 S.Ct. at 2704. Neither the small number of students involved in the instant case, nor the reduction in the white school-age population from 57.8 to 55.1%, taken alone or together, would directly produce a dual school system, and the court did not so find. The court did not find that the boundary change was itself a vestige of or flowed from a vestige of a dual system.15 The court also did not find that the vestiges of a formerly dual school system would be maintained by the proposed boundary change. The record is uncontroverted that both Goodrich and Livingston voluntarily desegregated many years ago, and neither district was ever subject to a specific local desegregation decree. Finally, the boundary change will not adversely affect the district's overall financial status. State action, in short, has not violated the modified order, nor does the proposed boundary change have a segregative effect.

Residential mobility is a virtue of a free and dynamic society. See Swann, 402 U.S. at 31-32, 91 S.Ct. at 1283-84. No doubt residential choices turn on economic and social considerations or even, at times, on private discrimination, but as long as they remain attributable to individual decisions, born of free choice, they are devoid of constitutional implications. See Freeman v. Pitts, 503 U.S. 467, 495, 112 S. Ct. 1430, 1448, 118 L. Ed. 2d 108 (1992). Although changing residential patterns inevitably affect the racial composition of schools, it is beyond the authority of the federal courts to counteract demographic changes in school districts that are the product of private choice and not state-sanctioned discrimination. See id. at 495, 112 S.Ct. at 1448. The district court could not predicate a violation of the modified order on the forecast that individual residential decisions, unrelated to the effect of past segregation, would change the district's racial makeup in the future. Again, it must be emphasized that the modified order does not reach private conduct but only state action that results in a dual system.

Moreover, even if it were relevant, the evidence furnishes weak support for the finding that the loss of Forest Springs would necessarily result in an exodus of white students from Goodrich. Goodrich's expert, Dr. Richard Murray, testified that studies suggest that white parents will "tolerate" an integrated school district as long as the percentage of minorities in the local population remains within a "comfort zone" of approximately 20%. As the minority population

rises, so do the anxieties of white parents. When the minority population reaches 40 to 50% of the local population, the studies imply that whites tend to withdraw from school districts in high numbers and seek refuge in more homogeneous communities.

Applying this theory of "tipping" to a rural district like Goodrich is difficult for two reasons. First, Dr. Murray admitted on cross examination that he was unaware of any academic study that has found an occurrence of tipping in rural communities. Although he stated that the phenomenon has occurred in rural Mississippi, he admitted that this observation was purely his own, not part of a formal analysis, and, hence, not subject to rigorous academic scrutiny.16

Second, in urban settings where tipping has been examined, studies suggest that it may be limited to areas where other neighborhoods and school districts are readily accessible. In urban neighborhoods, families can relocate short distances--within a large metropolitan area or to a nearby suburb--and select among public and private schools. A rural area like that where Goodrich is located affords fewer choices. Private schools are less common, and alternative neighborhoods, job opportunities and public school districts tend to be farther away.17 Thus, neither the theoretical likelihood of tipping nor its practicability in and around Goodrich is strongly supported.18

Finally, the district court found that if the proposed annexation were granted, the "public perception of [Goodrich I.S.D.] as the Black school district would be strengthened." But perceptions alone cannot form the basis for federal court intervention into the administration of a public school system. Even assuming the testimony relied upon by the district court accurately reflects community sentiment, this evidence reveals almost nothing about whether the demographic changes Goodrich I.S.D. might undergo in the wake of the proposed boundary change would be the product of private choices or a state-sponsored racial classification. The court's finding also fails to square with the persistent high turnover of the Goodrich school population, its significant fluctuations in size, and its growing Hispanic contingent--all of which facts are relevant to local perceptions, and all of which are far more significant than a one-time 2.7% population shift. District courts cannot substitute subjective judgments of this character for a finding of an actual segregative effect.

Because the district court's findings of segregative effects are too speculative, or are not supported by the record, or are rooted in private conduct rather than state action, they are clearly erroneous and legally insufficient. The district court erred in concluding that the proposed boundary change would violate either federal law or the modified order.

The district court declined to consider whether its interpretation of modified order 5281 is in accord with post-1971 Supreme Court decisions. It rejected this contention of Forest Springs based on the absence of any Fed.R.Civ.P. 60(b) motion to modify the decree. Several errors are imbedded in this part of the court's decision.

First, except for the superimposed 1% rule--which connotes impermissible racial balancing "for its own sake" rather than a remedy for past state action, Freeman, 503 U.S. at 494, 112 S.Ct. at 1447--the pertinent portion of modified order 5281, which deals with boundary changes, is not facially inconsistent with evolving school desegregation jurisprudence. The gist of post-1971 cases has been to confirm federal courts' broad remedial jurisdiction over those facets of school operations which represent or flow from an earlier de jure discriminatory system, while acknowledging that federal remedial jurisdiction goes only so far as the correction of the constitutional infirmity. See, e.g., Missouri v. Jenkins, 515 U.S. 70, 97-98, 115 S. Ct. 2038, 2053-54, 132 L. Ed. 2d 63 (1995); Freeman, 503 U.S. at 496-97, 112 S.Ct. at 1448-49; Spangler, 427 U.S. at 434-37, 96 S.Ct. at 2704-05; Milliken v. Bradley, 433 U.S. 267, 282, 97 S. Ct. 2749, 2758, 53 L. Ed. 2d 745 (1977); Milliken v. Bradley, 418 U.S. 717, 750-53, 94 S. Ct. 3112, 3130-31, 41 L. Ed. 2d 1069 (1974); Swann, 402 U.S. at 16, 91 S.Ct. at 1276. The modified order, although written broadly, as was necessary at the outset of the court's enforcement efforts, easily lends itself to the reading mandated by the Supreme Court. The order does not expressly mandate racial balancing in boundary changes. Cf. Spangler, 427 U.S. at 434, 96 S.Ct. at 2703. Nor does it expressly detach the consideration of boundary changes from prior unconstitutional state action. The district court has, in the Hightower ruling, acceded to intervening Fifth Circuit caselaw that rejected trivial statistical population differences as a stand-alone basis for federal intervention. See Lee v. Eufaula City Bd. of Educ., 573 F.2d at 232. Put otherwise, since there is no reason why the order must be interpreted to extend the district court's remedial jurisdiction beyond limits articulated by the Supreme Court, prudence and deference to the High Court strongly counsel enforcement of the order consistent with rather than in the teeth of its pronouncements.

Second, in light of the compatibility of modified order 5281 with later-articulated law, the district court's insistence on a Rule 60(b) motion as a prerequisite for re-evaluating its modified order is a red herring. The district court thought that Forest Springs must bear the burden of proving that the entire decree--covering the whole state of Texas--must be vacated or modified in order to justify the Residents' desire to transfer 22 students to Livingston I.S.D. The court's error lies in its perception that the Residents may only claim relief if the order is modified; because the modified order is limited to a remedy of the effects of past state-imposed segregation, the Residents' burden consisted of demonstrating only that the proposed annexation would not confound the remedy.

Third, to the extent the district court believed that it lacked the discretion to modify its order sua sponte, it erred. This court has clearly held that a district court has the authority to modify or terminate a consent decree19 sua sponte when it becomes manifest that there has been a change in the facts or the underlying law that gave rise to the decree. See Williams v. Edwards, 87 F.3d 126, 132 (5th Cir.1996) ("[T]he district court has the discretion to modify a decree when the court is made aware that the factual circumstances or the law underlying that decree has changed--regardless of the parties' silence or inertia."); Alberti v. Klevenhagen, 46 F.3d 1347, 1365-66 (5th Cir.1995). Moreover, in the context of this old case and the significant clarifications of school desegregation law that have occurred in the quarter-century since the modified order was issued, the court should have been wary of enforcing its order if the court perceived a significant gap between the order and ensuing legal developments. This error, of course, is irrelevant, where, as here, no such gap appears when the modified order is properly interpreted.

The Residents of Forest Springs have doggedly pursued their detachment from Goodrich for seven years. The victory they deserve is now probably Pyrrhic, as their children have grown and their subdivision's legal bills have mounted. It is most unfortunate that a noble decree to end school segregation has degenerated into a petty war of attrition against 22 students and 360 acres of tax base. The declaratory judgment of the district court is REVERSED.