This class action was commenced under 42 U.S.C. Sec. 1983 in the United States District Court for the District of Connecticut in December 1989. Plaintiffs are neglected or abandoned Connecticut children. Defendants are Connecticut's Governor and its Commissioner of the Department of Children and Families ("DCF"). Plaintiffs claimed that the structure and administration of Connecticut's child-welfare system violated their federal constitutional and statutory rights in many respects.

From the beginning, the parties and the district court have made extensive and admirable efforts to resolve the parties' disputes through agreement and cooperation. With the parties' consent, and even before defendant had answered the complaint, Judge Nevas appointed a three-member mediation panel consisting of a representative of the plaintiffs, a representative of the defendants, and Senior District Judge Robert C. Zampano. After over five months of concentrated effort that included negotiations between the parties, public hearings, and extensive deliberations, the panel proposed to the parties and to the court a comprehensive, multifaceted consent decree. The 120-page decree lays out a detailed plan for improving the management, operation, procedures, staffing, and funding of Connecticut's DCF. In January 1991, after a public hearing and with all parties consenting, Judge Nevas approved the proposed decree. The parties and the court then set about implementing the decree, which was designed to restructure a large part of the DCF.

Implementation required development of a manual for each section of the decree. Over a period of 19 months, under the aegis of the mediation panel, the parties developed and stipulated to a total of 12 manuals. Adopted as court orders, the manuals were incorporated as part of the consent decree on September 1, 1992. These manuals are guidelines and handbooks that set forth directives and details concerning the procedures, timetables, additional staffing requirements, funding requirements, and other matters necessary to implement and monitor the mandates in the consent decree. Significantly, the manuals also establish staff qualifications and caseload size for DCF's investigative and treatment personnel.

Because of the difficult issues involved, as well as the importance to the plaintiff class of enforcing the decree, the decree called for extensive monitoring. The original consent decree, of course, provided that the district court "shall have continuing jurisdiction of this action to ensure compliance with this consent decree." Initially, implementation of and compliance with the provisions of the consent decree were monitored by a monitoring panel, made up of the members of the mediation panel.

After the manuals were issued, however, Judge Nevas on December 1, 1992, signed, with the parties' consent, a detailed monitoring order that transferred both monitoring and the dispute-resolution process from the monitoring panel into the hands of a single, court-appointed, neutral monitor. One provision of the order states, "In monitoring the Decree and Manuals, the Monitor shall focus on patterns of compliance and non-compliance."

Since consenting to the decree and manuals, defendants have continued their cooperative effort to meet the problems of Connecticut's abandoned and abused children and, indeed, many worthwhile changes have been made in the system. One area--the need for staff increases--has proved to be particularly sensitive, and it forms the basis for this appeal, the first one in this institutionalreform action, which has been under administration now for four-and-a-half years. Recognizing that increased staffing is essential to DCF's operations, defendants prepared a plan for hiring more social workers and included its cost in the biennial budget that was submitted to the state legislature in March 1993. That body, however, cut over $8.7 million from the amount that DCF had requested.

Plaintiffs, fearing that this steep cut might undermine the decree's most important provisions, invoked the dispute resolution process provided in p IV B of the monitoring order. They alleged that the effect of the budget cuts would be to place DCF in non-compliance in four critical areas: (a) social work and support staffing; (b) hiring of nurse practitioners; (c) payments to foster parents; and (d) program enhancement.

Section IV of the monitoring order provides, in part:

B. If the Plaintiffs assert that DCYS [DCF] is, or is likely to be in non-compliance with any provision of the Consent Decree, Manuals, or Agreements, they must immediately notify DCYS [DCF] and the Court Monitor in writing. After such notice is given, the parties and Court Monitor shall meet within five (5) working days and attempt to resolve the issue. If the parties and the Court Monitor are not able to reach a timely agreement, then the issue shall be presented to the Trial Judge.

C. If any issue is presented to the Court, the Court Monitor shall certify to the Court in writing the issues to be decided, along with the Court Monitor's recommendations. At the Court hearing, the Court Monitor or any member of his staff may be called as a witness by a party or the Trial Judge.

In such a dispute, therefore, the monitor's function is to investigate the dispute, determine what the issues are, and make recommendations to the district court for their resolution. In this instance, the court monitor held hearings, at which defendants presented oral testimony, and both sides introduced numerous exhibits into evidence. Defendants admitted that the budget cuts would force them into a position of noncompliance, but, in an attempt to avoid imposition of further orders, they presented to the monitor a "revised hiring plan", claiming that the plan would place them in "substantial compliance" with the staffing and foster-parent payment requirements of the decree and manuals. When the monitor issued his "Certification of Issues and Recommendations", he rejected the adequacy of defendants' plan and found that it did "not achieve full compliance or a level of staffing that could be described as 'substantial compliance.' "

All the evidence developed at the hearing before the monitor was presented to Judge Nevas in June 1993. After considering that evidence and after hearing arguments from counsel, Judge Nevas adopted the monitor's findings of fact and, with some modifications, adopted his recommendations as a court order.

Two aspects of the district court's order are relevant to this appeal. The first is a phased timetable for hiring additional personnel. The manuals provided that various caseload standards must be in effect no later than June 30, 1994. To achieve those standards new people had to be hired. The order established a schedule for the hirings. It required that by July 1, 1993, DCF hire 145 social workers, 29 social work supervisors, 29 senior clerks, 8 program supervisors, 2 program directors, and 29 case aides, with a second wave, equal in numbers, to be hired by January 1, 1994. The new timetable and number of social workers to be hired had been taken from DCF's own budget request previously submitted to the state legislature.

The second aspect of Judge Nevas's order relevant to this appeal is a requirement that defendants reimburse foster parents at 100% of the United States Department of Agriculture (USDA) foster-parent rates by July 1, 1994. Section XV(C)(9) of the consent decree requires DCF, under the direction and with the approval of the monitoring panel, to "develop and implement schedules and conditions for reimbursement to foster parents at one hundred percent of USDA foster parent rates". Those rates are set periodically by the Department of Agriculture and represent an amount needed to reimburse foster parents for expenses incurred for housing, food, clothing, and the expenses associated with providing foster care. The effect of Judge Nevas's order was to set a specific date for achieving the 100% level required by the decree.

While defendants concede that the legislative cuts imposed on the DCF budget rendered them unable to achieve timely full compliance with the consent decree and manuals, they principally argue on appeal that: (1) Judge Nevas erred because he modified their legal obligations merely on the basis of predictions of prospective noncompliance, without actually finding any constitutional violation; (2) their proposed "substantial" compliance was sufficient to avoid judicial intervention by the court; (3) Judge Nevas could not adopt the facts found by the court monitor; and (4) Judge Nevas's order violated due process, because he did not provide defendants with a "fair hearing".

We have carefully considered all of defendants' contentions on this appeal in light of the history and importance of this significant litigation, and we find no basis for reversing Judge Nevas's order. Because of the public significance of this matter, and because it is likely that the parties will continue for some time to operate under this consent decree as implemented and modified by the manuals and the monitoring order, we write at some length so that the parties, the monitor, and the district court can be properly guided as they continue their cooperative efforts to improve Connecticut's child-welfare system.

1. "Modification" of the Decree.

Defendants argue that under general standards governing modification of consent decrees, the district judge erred because his order increased defendants' obligations beyond what they had consented to in the decree.

Consent decrees have a dual nature: they blend a court order with the parties' agreement. Consent decrees, of course, are injunctions, and their modification is reviewed under an abuse-of-discretion standard. Lorain NAACP v. Lorain Board of Education, 979 F.2d 1141 (6th Cir.1992). Modification is a remedy not to be lightly awarded, "especially where the design is not to relieve a party of obligations but to impose new responsibilities." Walker v. HUD, 912 F.2d 819, 826 (5th Cir.1990). "A contrary rule would discourage compromise for fear of adverse judicial modification." Id.

Consent decrees are also contractual, and while they "should be construed basically as contracts, without references to the [claims the plaintiffs] originally sought to enforce but never proved applicable through litigation," United States v. ITT Continental Baking Co., 420 U.S. 223, 236-37, 95 S.Ct. 926, 934-35, 43 L.Ed.2d 148 (1975), in implementing the purposes of a decree, a court is not rigidly confined only to the terms contained within the four corners of the parties' agreement.

In institutional-reform litigation a more flexible standard may be applied when modifying a consent decree. Rufo v. Inmates of Suffolk County Jail, 502 U.S. 367, 112 S.Ct. 748, 116 L.Ed.2d 867 (1992). In Rufo, the court noted that the "upsurge" in reform litigation involving governmental institutions "has made the ability of a district court to modify a decree in response to changed circumstances all the more important." Id. at ----, 112 S.Ct. at 758. The court added that "the public interest is a particularly significant reason for applying a flexible modification standard in institutional reform litigation." Id. at ----, 112 S.Ct. at 759. Under this standard, a party may obtain modification of a consent decree by establishing that there has been a significant change in circumstances, factual or legal, and that the proposed modification is suitably tailored to deal with the changed circumstances. Id. at ----, 112 S.Ct. at 760.

In this case, there has been no modification of the decree. The court monitor recommended, and the court ordered, that the original commitment to achieve certain caseload levels by a given date should be assured by staggering the new appointments over the remaining time so that by the time the agreed date was reached, the necessary new hires could be placed and trained, ready to handle the cases. Rather than modifying the decree by increasing defendants' obligations, as defendants claim, the court order in this case simply ensured compliance with the time frame originally established by the manuals.

Significantly, just this kind of amplification and specification had been anticipated by the parties and the court. In the first place, explicit language in the decree gave the court authority to issue orders to ensure compliance. "This court shall have continuing jurisdiction of this action to ensure compliance with this Consent Decree." Decree, p II(16). See also Berger v. Heckler, 771 F.2d 1556, 1568 (2d Cir.1985) ("Consent decrees are subject to continuing supervision and enforcement by the court").

Moreover, the monitoring order, agreed to by defendants, established a dispute-resolution process anticipating that the monitor, and eventually the court, would intervene, not just in cases of actual noncompliance, but also when plaintiffs showed that DCF was "likely" to be in noncompliance. Working under that procedure, the monitor investigated here and found that the defendants simply could not be in full compliance with the required hiring and training of caseworkers unless the court imposed the deadlines that he recommended. We therefore reject defendants' claim that the district court's order violated the legal principle that a party may not be held, under the guise of modification, to more onerous terms than those agreed to in a consent decree. This order was not a modification; it was designed to ensure full compliance with the original decree, and implementing the decree in this fashion was not an abuse of the district court's discretion.

Even if the June 25th order were to be viewed as a modification of the decree, we would still conclude, by applying the flexible Rufo standard, that the district court did not abuse its discretion. In the first place, we reject defendants' suggestion that the Rufo standard for flexible modification can benefit only defendants seeking relief from burdensome requirements. Rufo 's flexibility is designed to permit details in complicated decrees, much as this one, to be adapted to changing conditions so that the public interest can be preserved. As the Court noted, "[m]odification is * * * appropriate * * * when enforcement of the decree without modification would be detrimental to the public interest." Rufo, 502 U.S. at ----, 112 S.Ct. at 760. The failure of Connecticut's General Assembly to grant the funds requested by DCF constitutes a significant enough factual change to justify the changes ordered by the district court. Those modifications were necessary to ensure timely implementation of the decree and provide for the plaintiff class the protections and services originally agreed to by the parties and ordered by the court.

2. Substantial Compliance.

Defendants also argue that the district court erred in particularizing the hiring schedule, because the defendants were at least in substantial compliance with the decree. However, the decree does not contemplate anything less than 100% compliance with the caseload standard, and as already noted, the district court's order merely required specific numbers of new hires and a graduated hiring timetable, found to be necessary in order to ensure full and timely compliance. Defendants' argument of substantial compliance might become relevant if ever there should be a hearing for contempt based on claimed noncompliance by defendants. U.S. v. Commonwealth of Massachusetts, 890 F.2d 507, 509 (1st Cir.1989) (establishing as the test in contempt proceedings, "whether the objectives of the decree have been substantially achieved"). "Substantial compliance" has no bearing, however, on the district court's power to interpret the decree and, when necessary, to make its terms more precise and realistic.

3. Monitor's Factfinding.

DCF claims that the monitor was not authorized to make findings of fact, but that even if he were, the district court erred because it did not make its own independent findings, but instead adopted the monitor's report without addressing DCF's objections or permitting DCF to present evidence about its efforts to comply with the decree.

The powers of the court monitor are set forth in the district court's monitoring order of December 1, 1992. It states that the court monitor "shall be responsible to the Trial Judge, but shall work actively with the parties to ensure timely and effective compliance of the provisions of the Consent Decree". To carry out these duties and responsibilities, "the Monitor is empowered to take any and all action to effect timely compliance * * *." Specifically, he is empowered to monitor implementation and compliance, to convene a meeting of the parties, establish a reporting structure that "enables the monitor to effectively assess the progress of the implementation of the Juan F. Consent Decree," obtain information from DCF, issue compliance reports, attempt to resolve disputes, and "review requests by either party for modification * * *, and if necessary, to make a recommendation to the Trial Judge regarding the request for modification." If the monitor cannot resolve a dispute between the parties, then he must "certify to the Court the issues to be decided" and must also submit his own recommendations for their resolution.

Obviously, the monitor could not do many of these things without making an investigation and determining what are the essential facts that guide his analysis and recommendations. In this case, to determine what it was that plaintiffs were complaining about, the monitor held a lengthy evidentiary hearing, made findings of fact, and then certified the issues and his recommendations to the district court--all in order to effect timely compliance with the provisions of the decree.

As to what kind of judicial officer the "court monitor" is, it seems apparent that he is a special master, albeit by another name. Like a special master, who is subject to the limitations of the order of reference, the court monitor has the power to hold meetings, request information, issue reports, hear evidence, and issue findings of fact. See F.R.C.P. 53(c). A special master's findings are entitled to deference. "In an action to be tried without a jury the court shall accept the master's findings of fact unless clearly erroneous." F.R.C.P. 53(e)(2). Because of the similarity in the roles of the court monitor and a special master, we think it proper to apply the same "clearly erroneous" standard to the monitor's findings of fact. Here, the monitor's findings of fact were not clearly erroneous. Defendants had conceded that they could not fully comply with the provisions in the decree, and the monitor's findings, analysis, and recommendations to the district court, all based on ample evidence, did no more than ensure compliance with the decree, by making more specific the dates when stages in the compliance plan should be achieved.

While defendants object to the district court's having deferred to the "findings" that underlay the monitor's "recommendations", it seems apparent that the parties considered the hearing that the monitor conducted over a three-day period to be, in effect, a trial. The parties made opening statements; they placed documents in evidence; they entered into formal stipulations of fact; a court reporter kept a verbatim record of the proceedings; defendants called four witnesses to testify, placed them under oath, and subjected them to direct examination, followed by cross and redirect examination; and both sides presented legal arguments. The issues were thus made clear, and ultimately the court monitor filed his report and recommendations with the district judge. The parties' actions demonstrate their understanding that a trial was being held and that at the conclusion of the trial the court monitor was going to report to the district judge his analysis of the issues and recommendations, including the facts on which they were based.

4. Due Process.

Defendants contend that "[d]ue process required the court to hold a hearing before ordering the defendants to perform certain obligations not previously agreed to." (emphasis added). Of course, as we have already pointed out, the court's order did not require defendants to perform any new obligations; it merely made more precise and realistic the required performance of obligations that defendants had already undertaken.

With respect to defendants' claim that a fuller hearing was required before the district court, defendants specifically complain about not having been able to present new evidence to the district judge. The district judge, however, asked for and received a proffer of the proposed new evidence, and he properly determined that it was irrelevant. Commissioner Senatore's proposed testimony about implementation of defendants' "revised hiring plan" did not relate to what was needed for compliance; instead, it sought to show that defendants had already achieved "substantial compliance", which as we demonstrated above, would be relevant only with respect to a possible contempt sanction, not in issue on this appeal.

We note in passing that Commissioner Senatore and three other defense witnesses had testified at the hearing before the monitor. To require the district judge to hear again the testimony of these and perhaps other witnesses would circumvent and undercut the work of the court monitor who, under the consent decree as modified by the monitoring order, is the centerpiece of the alternative-dispute-resolution process which so far has functioned so well in working out this complex dispute.

The entire transcript of the hearings before the monitor, the exhibits that had been admitted into evidence, the stipulations of fact, and the monitor's recommendations were all before the trial judge for his review and consideration. We conclude that the defendants were not denied due process by these procedures, which properly and effectively carried out the intent of the consent decree and the monitoring order.

Resolution of this complex case by consent decree would not have been possible without the admirable cooperation of the parties and the careful, diligent work of the court monitor. Their joint efforts have effectively addressed over one-hundred issues that plaintiffs have advanced in their broad-scale challenge on behalf of Connecticut's foster care and adoptive children. Because of the parties' cooperation, wisdom, and good faith displayed so far in dealing with these important problems, lengthy and expensive formal judicial proceedings have been avoided, and relief to the plaintiff class has been expedited. We fully expect that the same attitudes will continue to prevail as the parties, the court monitor, and the district court continue to wrestle with the problems that still remain to be resolved under the consent decree.

We affirm the order appealed from, but of course recognize that the district court and the parties may find it necessary to modify the originally ordered compliance dates in order to compensate for the time necessarily lost in processing this appeal.