Small v. Hunt

Affirmed by published opinion. Judge MICHAEL wrote the opinion, in which Chief Judge WILKINSON and Judge WILLIAMS joined. OPINION MICHAEL, Circuit Judge: Plaintiffs/appellants (“appellants”) are a class of prisoners incarcerated at 49 medium and minimum security prisons in North Carolina. Defendants/appellees (collectively, the “state”) are the Governor of North Carolina, the Secretary of the Department of Correction, and the Director of the Division of Prisons. Appellants appeal from the district court’s order of June 28, 1994, granting the state’s petition to modify a consent decree relating to conditions at the 49 prisons. 1 In addition, appellants appeal the district court’s grant of the state’s motion to alter or amend the June 28, 1994, order. Appellants raise both substantive and procedural objections to the district court’s orders. For the reasons that follow, we affirm in all respects. *792 I. This is the latest chapter in a dispute that began in 1985 when several North Carolina prisoners filed a class action against the state seeking relief from living conditions alleged to be unconstitutional at 49 medium and minimum security prisons. Altogether, the 49 prisons comprise about half of the North Carolina prison system’s adult male capacity. These prisons are dormitory style where prisoners sleep on bunks in large rooms. On December 28, 1988, appellants and the state entered into a comprehensive settlement agreement that soon terminated the lawsuit. The agreement contained specific provisions aimed at improving prison conditions. One provision required the state to provide 50 square feet of living space per inmate in each dormitory by July 1, 1994. Fifty square feet of living space per inmate is referred to as “standard operating capacity” (SOC). The agreement also contained interim provisions that allowed the state to house prisoners at 140% of SOC in medium security facilities and 125% of SOC in minimum security facilities until July 1,1994. 2 The agreement contained a number of other provisions. For instance, it required the state to replace all triple bunks with double bunks and to implement direct supervision by correctional officers in the dormitories. The provision for direct supervision required the deployment of correctional officers to patrol the sleeping areas inside each separately locked dormitory. The state also agreed to provide access to dayrooms with a minimum of 25 square feet per prisoner. Other provisions dealt with peer review of medical care, visitation policies, work and study programs, fire safety, new dormitory design, ventilation renovations, medical diets, plumbing repairs, clothing and linen policies, and locker space. The settlement agreement provided that it would become effective upon the “approval and ratification” by the North Carolina General Assembly and the approval of the district court. In approving the settlement the General Assembly in March 1989 agreed that “funds necessary to satisfy the terms and obligations of th[e] agreement will be appropriated.” 1989 N.C. Sess. Laws, eh. 8 § 1. By order of April 3, 1989, the district court approved the settlement agreement, and it assumed the force of a consent decree. The state moved on four interrelated fronts to tackle the problem of prison overcrowding and to comply with the consent decree. First, the state enacted a prison cap that limited prison population to 18,000. See N.C. Gen.Stat. § 148-4.1. Second, by 1994 the state had appropriated almost $500 million for new prison construction. Third, the state implemented and expanded programs providing alternatives to incarceration, thereby diverting thousands of offenders from prison. Finally, the state enacted the Structured Sentencing Act and later accelerated its implementation. See N.C. Gen.Stat. § 15A-1340.10 et seq. The structured sentencing legislation reduces the length of incarceration for less serious offenders and reserves the longest sentences for the most dangerous offenders. These considerable efforts by the state, however, were overwhelmed by double-digit percentage increases in prison admissions that began in 1988. When the state, through its settlement committee, 3 negotiated the settlement, the committee relied on a projection by the Division of Prisons that annual inmate admissions would increase between 3 and 5%. The committee believed, based on these projections, that the state would have the financial resources necessary to implement the settlement agreement. However, the actual rates of admission far exceeded the projec *793 tions. There were significant increases of 11.6% in 1988, 17.6% in 1989, 8.1% in 1990, 16.1% in 1991, and 8.1% in 1992. Although the increases leveled off in 1992 and 1993, total prison admissions had increased from a level of 17,000 admissions per year in 1988 to almost 31,000 per year by 1992. The state’s efforts to comply with the 1989 consent decree and to respond to its prison crisis have come at great expense to North Carolina taxpayers. By 1994 almost half a billion dollars had been spent or committed for prison construction projects. General operating expenses for the prison system more than doubled between 1988 and 1994 to $560 million per year. Many of these increased expenses are directly attributable to the unexpected rise in prison admissions. For example, the state’s settlement committee anticipated annual operating costs of approximately $68 million for newly constructed facilities, based on the projected need for about 5,000 new beds by 1996. Yet, because of the unexpected explosion in prison admissions, by 1994 the state had appropriated funds to build facilities for 18,000 new beds, about 13,000 more than initially planned. As a result, the total operating costs for completed and funded construction is estimated at approximately $169 million per year. Thus, the difference between current operating cost projections and the projections that guided the settlement committee’s decisions represents a yearly increase in operating costs of almost $101 million. These are costs that will continue each year over the lifetime of the facilities. At the time (October 13, 1993) the state petitioned to modify the consent decree, it had complied with virtually every provision except the requirement of 50 square feet of living space per inmate. 4 The 50 square feet provision was the subject of the state’s October 1993 petition. Specifically, the state sought a modification of the consent decree permitting it to (1) house inmates in the “new” dormitories at 130% of SOC, (2) house minimum and medium custody inmates in the “old” dormitories at 140% of SOC, and (3) make conforming changes to other provisions of the agreement consistent with these capacity changes. 5 The state’s petition alleged a “drastic, unanticipated increase in prison admissions rates” that “ha[d] made performance of the Agreement’s terms related to dormitory living space inequitable.” The petition said that unless the court granted modification, “the public interest w[ould] be harmed.” The district court held a four-day eviden-tiary hearing. Despite the unanticipated explosion in prison admissions, evidence offered at the hearing established that the state has been operating a secure and humane prison system, substantially above minimum constitutional requirements. The state presented testimony and affidavits from its prison administrators, superintendents, and two experts concerning the quality of living conditions in the 49 prisons at issue. The experts noted that serious assaults by inmates on inmates had greatly diminished. They saw significantly less tension between inmates and officers. They found that the addition of dayrooms had changed the dormitories for the better. One expert said that the prison units operating at the interim capacity levels provided a “better quality of life and safer place for inmates and staff to live and work than most prisons throughout the United States.” Another expert said that these units have “one of the lowest assault or violence rates of any prisons I have inspected.” In sum, considerable evidence showed that the state could operate the dormitories at capacity levels above 100% of SOC and still provide for the basic needs and safety of the prisoners without risking safety for the staff. Within a week of the hearing, the district judge visited two of the prison units without the parties or their counsel in attendance. Thereafter, by order dated June 28,1994, the district court granted the state’s petition in part and denied it in part. Specifically, the *794 court permitted the state to house inmates in the new dormitories at 125% of SOC but denied the state’s request to continue housing inmates in the old dormitories at 140% of SOC. The' district court found that the settlement committee had not foreseen the extent of prison population increases and that the state’s prison system “is still struggling to manage the vast numbers of inmates,” despite recent declines in admission rates. Small, 858 F.Supp. at 518. The court concluded that the public interest was best served by allowing the new dormitories to continue to operate at 125% of SOC and granted modification of the decree to that extent. The court refused to allow operation of the new dormitories at 130% of SOC (as requested by the state) because the court understood that the state had never before operated them at that level. The court denied any modification for the old dormitories because operation in excess of 100% of SOC required a center row of bunks, which the court believed posed risks for inmates. However, the court noted that “[t]he record is not clear whether there are other older units of different dimensions or of different bunk arrangement that would exceed 100% of SOC and still meet the Rufo [v. Inmates of Suffolk County Jail, 502 U.S. 367, 112 S.Ct. 748, 116 L.Ed.2d 867 (1992) ] standard.” Id. at 524. On July 13,1994, the state moved pursuant to Fed.R.Civ.P. 59(e) and 60(b) to alter or amend the June 28,1994, modification order. First, with respect to the new dormitories, the state argued that the court had misunderstood the evidence and thus had erroneously concluded that the state had no experience operating the new dormitories at more than 125% of SOC. The state’s evidence in fact demonstrated that the new medium security dormitories had been operated at population levels of between 130% and 140% of SOC since the consent decree had been entered. Second, with respect to the old dormitories, the state proposed a bunk configuration eliminating the middle row of bunks that had concerned the court. Thereafter, in an order dated March 27, 1995, the district court agreed that the evidence did show that the state had been operating the new dormitories at more than 125% of SOC. It also found the alternative bunk configuration acceptable. The court therefore granted the state’s motion to alter or amend the June 28, 1994, order by allowing the state (1) to house inmates in the new dormitories at 130% of SOC and (2) to house inmates in the old dormitories “at varying capacities greater than 100% of capacity, but not greater than 125% of capacity, so long as the unit does not have a center row of bunks and there is at least 32 inches of space between bunks.” The modifications to the consent decree have been appealed on both substantive and procedural grounds. II. We turn to whether the modification orders were proper on the merits. Preliminarily, we must set out the standards governing the modification of consent decrees entered in institutional reform litigation. On April 26, 1996, amendments to 18 U.S.C. § 3626, known as the Prison Litigation Reform Act (the “new Act”), were signed into law by the President. The new Act is aimed in part at making it easier for those running state prisons to seek termination or modification of federal court orders relating to prison conditions. The new Act provides that it “shall apply with respect to all prospective relief whether such relief was originally granted or approved before, on, or after the date of the enactment of this title [April 24, 1996].” Pub.L. No. 104-134, § 101, 110 Stat. _ (1996). In this ease the district court proceedings that led to the modification orders and the oral argument on appeal were all completed prior to April 24, 1996. The state has not sought to invoke the new Act, either by asking us to apply it in the first instance or by asking for a remand and reconsideration by the district court. Because the state has not mentioned the new Act and because we conclude below that the district court was correct in granting the state relief under pre-April 24, 1996, standards, we do not believe it is necessary for us to consider the provisions of the new Act. We, therefore, outline the standards without consideration of the new Act. *795 Under Fed.R.Civ.P. 60(b)(5) a court may modify a consent decree on "such terms as are just" if "it is no longer equitable that the judgment [I have prospective application." When applying Rule 60(b)(5) to consent decrees in prison reform litigation, district courts should "exercise flexibility in considering [modification] requests." Rufo v. Inmates of Suffolk County Jail, 502 U.S. 367, 383, 112 S.Ct. 748, 760, 116 L.Ed.2d 867 (1992). "`[T]he unique nature and demands of institutional reform litigation necessitate a more flexible approach to modification' than may be appropriate with respect to consent decrees between private parties." Plyler v. Evatt 924 F.2d 1321, 1324 (4th Cir.1991) (Plyler II) (quoting Plyler v. Evatt, 846 F.2d 208, 212 (4th Cir.) (Plyler I), cert. denied, 488 U.S. 897, 109 S.Ct. 241, 102 L.Ed.2d 230 (1988)). "The `uniqueness' of [such] litigation lies in the fact that it is necessarily aimed at achieving `broad public policy objectives in a complex, ongoing fact situation,' with the consequence that consent decrees settling such litigation must be viewed as embodying `not so much peremptory commands to be obeyed but] as ... future-oriented plans designed to achieve [those] objectives.'" Id. (quoting Plyler I, 846 F.2d at 212). In all events, however, a movant is not entitled to relief simply because "it is no longer convenient to live with the terms of a consent decree." Rufo, 502 U.S. at 383, 112 S.Ct. at 760. In Rufo the Supreme Court articulated a framework for examining a party's request to modify an institutional reform consent decree. A party seeking modification has the burden of first "showing [that] a significant change either in factual conditions or in law" warrants revision of the decree. Id. at 384, 112 S.Ct. at 760. 6 If the movant cites significantly changed factual conditions, it must additionally show that the changed conditions make compliance with the consent decree "more onerous," "unworkable," or "detrimental to the public interest." Id. Of course, general equitable principles require that the movant's good faith be taken into account. See Plyler II, 924 F.2d at 1324. 7 If the movant succeeds in demonstrating that a significant change in circumstance warrants modification of the decree, a court must then determine whether "the proposed modification is suitably tailored to the changed circumstance." Rufo, 502 U.S. at 391, 112 S.Ct. at 765. As a court makes this determination, "three matters should be clear." Id. First, modification of a consent decree "must not create or perpetuate a constitutional violation." Id. Second, a modification "should not strive to rewrite a consent decree so that it conforms to the constitutional floor." Id. Rather, a court should do no more than necessary to resolve the problem created by the changed circumstance. Id. Third, within the constraints just mentioned, principles of federalism require a district court to defer to local government administrators "to resolve the intricacies of implementing a decree modification." Id. at 392, 112 S.Ct. at 764. This means that while "[f]inancial constraints may not be used to justify the creation or perpetuation of constitutional violations, [] they are a legitimate concern of government defendants in institutional reform litigation and therefore are appropriately considered in tailoring a consent decree modification." Id. at 392-93, 112 S.Ct. at 764. In light of these standards we now consider (1) whether the district court correctly decided that there was a significant change in factual conditions warranting modification of the consent decree and (2) whether the modification allowed by the district court is *796 suitably tailored to the changed conditions. We review the district court’s decision for abuse of discretion, and we accept the factual findings on which the district court’s decision is based unless they are clearly erroneous. Plyler I, 846 F.2d at 212. A. Appellants do not contest the existence of a significant change in factual circumstances. However, they argue that the state failed to establish that the change in circumstances makes compliance with the consent decree “substantially more onerous,” “unworkable,” or “detrimental to the public interest.” See Rufo, 502 U.S. at 384, 112 S.Ct. at 760. We disagree. First, the district court found that compliance with the requirement of 50 square feet of living space would be substantially more onerous in light of the unanticipated explosion in inmate population. The settlement committee had projected that prison admissions would increase at a rate of 3 to 5% annually. In fact, admissions increased at an average annual rate of 12.8% between 1988 and 1993. In raw numbers, inmate admissions jumped from about 17,000 per year in 1988 to almost 31,000 per year by 1993. Thus, by 1993 the prison system was getting almost 31,000 new inmates each year and putting them in facilities with an SOC of just under 20,000 beds. Moreover, the general operating costs for the prison system had doubled since 1988. The operating costs attributable to all new beds added to the system exceeded the settlement committee’s projections by $101 million annually. Evidence such as this led the district court to find that even with the recent decline in admission rates, “the [prison] system presently is still struggling to manage the vast numbers of inmates that are admitted each year.” Small, 858 F.Supp. at 518. Appellants argue that compliance with the original decree is not substantially more onerous since the state has almost fully complied with the 50 square feet requirement despite the changed circumstances. But appellants overlook the fact that the state has expended huge sums of money, far in excess of amounts anticipated, to reach even partial compliance. We do not believe the state should be penalized with the denial of relief simply because it has been reasonably successful as a result of very diligent, good faith efforts to comply fully with the consent decree. Second, the district court determined that enforcement of the 50 square feet provision would be detrimental to the public interest. The court emphasized that the public has an important interest in having its institutions run in a fiscally responsible manner. The court also recognized that even if the state has sufficient funds to build thousands of new prison beds, “the public is entitled to have the public funds directed to the most important state projects and to reap the fullest benefit from the dollars spent.” Id. at 523. Appellants point out that the district court rejected the state’s contention that compliance with the decree would endanger public safety because early release of prisoners would increase the commission of crimes. The court was unpersuaded because the state “could not point to any solid statistics in support of that view.” Id. at 518. But the court’s rejection of this contention does not help appellants. As the district court correctly noted, the public has interests other than safety. It also has an interest in how its tax dollars are spent. Finally, in an apparent effort to show lack of good faith by the state, appellants argue that the state’s increased expenditures for prison operations simply represent policy choices by the legislature to use incarceration as a sanction. These increased expenditures, appellants say, were not really made to comply with the consent decree. We have said that “enactment of stricter criminal laws should not be considered bad faith.” Plyler I, 846 F.2d at 213. Thus, we do not see increased reliance on incarceration to punish convicts as a policy choice designed to thwart compliance with the decree. Here, the district court found that the state had acted in good faith to comply with the original consent decree, and that finding is amply supported by the evidence. The district court was correct in determining that the significant change in factual *797 circumstances warranted modification of the original consent decree. B. We now turn to whether the modification is suitably tailored. Appellants argue that even if some modification is appropriate, a permanent modification is not justified because the facts show that the problem (the unanticipated increase in prison population) is only temporary. Appellants point to evidence showing that (1) the increase in admission rates has abated, (2) certain measures, such as the Structured Sentencing Act, should result in the stabilization and reduction of the prison population, (3) North Carolina will soon divert most misdemeanants away from the state prison system, (4) the Department of Correction expects admissions to decrease substantially, (5) several thousand new prison beds are funded and under construction and will soon be available, and (6) available beds (with 50 square feet per inmate) are projected to exceed the prison population in the near future. We agree that a district court should only consider granting permanent relief if temporary relief will not resolve the problems created by the changed circumstances. “[T]he focus should be on whether the proposed modification is tailored to resolve the problems created by the change in circumstances. A court should do no more, for a consent decree is a final judgment that may be reopened only to the extent that equity requires.” Rufo, 502 U.S. at 391, 112 S.Ct. at 764. If temporary relief will resolve the problem, permanent relief is obviously not a “suitably tailored” solution. See id. We believe, however, that permanent relief is appropriate in this case because temporary relief would not resolve the problem of the enormous expense required to incarcerate the much larger population. The increased costs associated with incarcerating the unfor-eseeably large prison population are permanent, not temporary. Annual prison operating costs are $101 million in excess of what the state predicted when it entered the settlement agreement. Even if prisoner admissions stabilized, and even if the state had the facilities to keep the expanded population imprisoned at an allocation of 50 square feet per inmate, the state still did not anticipate the added financial burden of operating a prison system with twice the number of inmates expected. We therefore conclude that the district court did not abuse its discretion in granting permanent relief from the 50 square feet requirement. All in all, we conclude that the modifications (including those in the amending order of March 27, 1995) allowed by the district court are suitably tailored to the changed circumstances confronting the state. III. We next consider appellants’ procedural objections. A. The first procedural issues concern the state’s July 13, 1994, motion, pursuant to Fed.R.Civ.P. 59(e) and 60(b), to alter or amend the court’s June 28, 1994, order. We will treat the state’s motion as a Rule 59(e) motion since it was “served not later than 10 days after entry of the judgment.” See Fed. R.Civ.P. 59(e); Campbell v. Bartlett, 975 F.2d 1569, 1580 n. 15 (10th Cir.1992) (“Motions served within 10 days of judgment ‘ordinarily will fall under Rule 59(e),’ while motions served later fall under Rule 60(b).”) (quoting Van Skiver v. United States, 952 F.2d 1241, 1243 (10th Cir.1991), cert. denied, 506 U.S. 828, 113 S.Ct. 89, 121 L.Ed.2d 51 (1992)). 1. Appellants argue that Rule 59(e) provided no basis for the district court to grant the state’s motion to amend the June 1994 order insofar as it related to authorized population levels in the new dormitories. We disagree. A court may grant relief under Rule 59(e) to “correct manifest errors of ... fact upon which the judgment is based.” 11 Charles Alan Wright, Arthur R. Miller & Mary Kay Kane, Federal Practice and Procedure: Civil 2d § 2810.1, at 125 (1995). In its June 1994 order the district court said that it would not grant relief for population *798 levels beyond 125% of SOC in the new dormitories because it believed that the state had never operated them with inmate numbers in excess of 125% of SOC. Upon reviewing the state’s motion to alter or amend the June 1994 order, the district court realized that it had erroneously overlooked evidence that the state had indeed operated new dormitories at levels between 130 and 140% of SOC. Having discovered its mistake about the facts that led to its judgment on this issue, the district court was authorized to make a correction. See 11 id. The court therefore properly amended its modification order to allow the state to house inmates in the new dormitories at 130% of SOC. 2. We also reject appellants’ argument that the district court abused its discretion in considering (on the state’s motion to amend the modification order) new evidence on alternative bunk configurations in the old dormitories. Rule 59(e) relief may also be appropriate “to account for new evidence not available at trial.” Hutchinson v. Staton, 994 F.2d 1076, 1081 (4th Cir.1993). In that circumstance a party must produce a “legitimate justification for not presenting” the evidence during the earlier proceeding. RGI, Inc. v. Unified Indus., Inc., 963 F.2d 658, 662 (4th Cir.1992). We recognize that the bunk reconfiguration plan for old dormitories was not newly discovered evidence in the ordinary sense. However, it was newly requested evidence. In its June 28, 1994, order the court essentially invited the state to submit a proposed bunk configuration that would be suitably tailored under Rufo. The state had a “legitimate justification for not presenting” the reconfiguration plan earlier because until the court expressed its concern about a center row of bunks, the state had no reason to present an alternative proposal for the court’s consideration. See RGI, Inc., 963 F.2d at 662. In its management of the case, the court could have delayed a final decision on the modification until it advised the parties of its (previously unexpressed) concern about the center bunks and gave them a chance to respond. The same result was achieved by the earlier entry of the modification order, which then prompted the state’s motion to alter or amend. In this particular circumstance, we cannot say the district court abused its discretion in granting an amendment under Rule 59(e) to allow varying capacities of up to 125% of SOC in the old dormitories. 3. There is no merit to appellants’ contention that the state was required to run through the entire Rufo test anew on its motion to alter or amend the judgment. The district court had already analyzed Rufo’s first step and determined that the state was entitled to relief from the 50 square feet provision at all dormitories, both old and new. But the court denied relief with respect to the old dormitories because the proposed remedy was not suitably tailored. The motion to alter or amend was limited to the issue of whether a suitably tailored solution could be crafted (Rufo’s second step). Thus, another complete showing of entitlement to relief under Rufo’s first step would have been superfluous. 8 B. Finally, appellants argue that the district court committed prejudicial error by conducting a view of two prison facilities in the absence of counsel after the close of the evidence. The district court correctly acknowledged that the view was error, see Lillie v. United States, 953 F.2d 1188, 1191 (10th Cir.1992), so we need only determine whether the error was harmless. A district court’s use of evidentiary findings from an improper view is reviewed under the standard governing the erroneous admission of evidence. See id. at 1192. A new trial is warranted unless the other competent evi *799 dence is “sufficiently strong” to permit the conclusion that the improper evidence had no effect on the decision. Id. Applying this standard, we conclude that the view was harmless. The first evidentiary hearing on the state’s motion for modification ended on May 13, 1994. Six days later, on May 19, 1994, the district judge visited two state prisons “to get a better concept of just what the contentions were.” Counsel were given notice of the visit an hour in advance, but they were not invited along. Appellants did not raise any objection to the visit until after the initial modification order was entered on June 28, 1994. Then, on July 13, 1994, appellants made a motion asking the district court to amend its order to eliminate any reliance on the court’s visit or, in the alternative, to grant a partial new trial to permit “a properly conducted view” or the “taking of additional evidence with respect to the view.” The district court heard extensive argument on appellants’ motion. The court then carefully reviewed its June 28, 1994, order “to determine what part of its findings or relief might have been based, in any way, on its view.” The district court determined that only two categories of findings were influenced by its visit. The first category was a single finding that the two prisons were “spotlessly clean and in good order.” This finding is harmless because it was immaterial to what the court had to decide under Rufo. The second category of visit-related findings dealt with space. But the district court’s determination that the state had established changed circumstances under Rufo was based almost entirely on findings concerning increases in prison admissions, and the visit had no effect on these findings. The visit did have some relevance to the Rufo requirement of “suitably tailored” relief. However, the court’s original decision to allow the state to operate the new dormitories at 125% of SOC, despite the state’s request to operate them at 130%, was supported by sufficiently strong evidence introduced at the four-day hearing. Thus, the visit had no effect on that decision. In the original order the district court denied the state any relief for the old dormitories, so the visit caused no prejudice to appellants on that part of the order. Accordingly, the error in making the prison visit was harmless. 9 IV. The orders of the district court are affirmed. AFFIRMED. Footnotes [1] This order has been published as Small v. Hunt, 858 F.Supp. 510 (E.D.N.C.1994). [2] The district court provided the following illustration of what these percentages of SOC mean: [A] 50 inmate dorm providing 50 square feet for each inmate has 2500 square feet. Operating at 125% of capacity would mean that 62 inmates would have 40.33 square feet each. Operating at 130% of capacity would mean that 65 inmates have 38.46 square feet each. And operating at 140% of capacity would mean that 70 inmates would have 35.72 square feet each. Id. at 523 n. 9. [3] The state was represented in the negotiations by a settlement committee composed of representatives of the Governor, the Lieutenant Governor, the General Assembly, and the Department of Correction. [4] The state had also not yet completed construction of dayroom space in minimum security facilities. However, that task was completed by July 1, 1994. [5] Obviously, the new dormitories were constructed more recently. According to the state, however, "new” and "old” also relates to the size of the dormitories. Dormitories of 50 and 102 inmate capacity are “new” and all others are "old.” [6] The state does not rely on a change in law. [7] "Ordinarily [] modification should not be granted where a party relies upon events that actually were anticipated at the time it entered into a decree." Rufo, 502 U.S. at 385, 112 S.Ct. at 760. Conversely, modification is not precluded if the movant did not actually foresee the changed conditions. See id. Appellants do not appear to challenge the district court's finding in this case that the state did not foresee the explosion in North Carolina's prison population. In any event, the district court's finding is not clearly erroneous. The district court relied upon the testimony of former Director of Prisons Joseph Hamilton, who said that the settlement committee did not anticipate the dramatic rise in inmate admissions. The court found Hamilton's testimony "credible and worthy of belief," Small, 858 F.Supp. at 515, and we accept that assessment. [8] We also reject appellants' argument that the March 27, 1995, order (which amended the June 28, 1994, order) rewrites the settlement agreement to allow conditions in the new dormitories to conform to the constitutional floor in violation of Rufo. Even under the terms of the amended modification order, the state is required to operate the new dormitories in a manner that exceeds minimal constitutional requirements. [9] We pause to note that the record reveals that the district judge, who has presided over this case since 1985, has devoted much time and painstaking attention to it. We are satisfied that, the visit notwithstanding, the judge has been open and fair in his conduct of the case and that his decisions have been the product of careful deliberation.