In April Congress presented New York and municipalities across the country with an unexpected and much-needed legislative gift: the Prison Litigation Reform Act. Placing strict limits on what federal courts can do to improve conditions in state and local prisons, the act is a watershed in the struggle over whether judges or elected mayors and legislators should have the power to make municipal policy. Wasting no time invoking the new law, the Giuliani administration won a federal court ruling in July that vacated the exhaustively detailed decree under which the court had supervised management of New York’s prisons since 1978. If the appellate court affirms the ruling, the mayor and the City Council will again be able to run the prisons, from cleaning cells to feeding prisoners, without first having to ask permission from a judge.

In New York and other cities, court decrees still control vast areas of municipal policy, from special education to the placement of foster children to housing for the homeless. Allocating huge chunks of municipal budgets on policies that elected officials haven’t chosen and that often have proved counterproductive, these judicial decrees undercut democracy. The Prison Litigation Reform Act provides an especially instructive model for what Congress and the state legislatures should do to restore authority to local officials.

The act revolutionizes the administration of state and municipal prisons nationwide. It is a revolution longed for not just by New York but by many other localities. Twenty-four prison agencies across the country chafe under population caps imposed by the federal courts, and many more are subject to court orders regulating prison conditions in general. In Michigan, for example, as Senator Spencer Abraham recounted during floor debate over the act, a federal court controlled the warmth of prison food, the brightness of prison lights, the comfort of prison air and water, the availability of electrical outlets in prison cells, and the professional credentials of prison barbers. Michigan had consented to the decree in 1982 and was still wrestling with it 14 years later.

The act makes two fundamental changes in the authority of federal courts over prisons. First, it puts a stop to open-ended decrees by imposing a term limit. Now, unless plaintiffs can show ongoing violations of law, no prison decree can last more than two years if a local government objects. Second, the act places boundaries on the content of decrees, forbidding them to impose obligations simply because judges, plaintiffs, and officials think they are a good idea. Federal courts now must gear decrees to fixing particular constitutional violations. With breathtaking simplicity, Congress restored the prerogatives of local self-government while at the same time confirming the courts’ duty to uphold constitutional and statutory rights. Indeed, the act does nothing more than confine courts to this duty, and its consequences will be so profound only because courts have routinely gone so far beyond their legitimate role.

For New York, the new law blew down a vast legal edifice built brick by litigated brick, beginning when the Legal Aid Society’s Prisoners’ Rights Project filed several suits against the Beame administration over city management of its prisons. In 1978 the newly elected Koch administration settled the case by accepting a 51-page federal court decree. Over the years the parties to the decree amended it more than 90 times, extending it to every imaginable aspect of prison management. No detail was too picayune: the city was obliged, for instance, to clean prison showers with the disinfectant Boraxo, to provide inmates with the Sports Channel via satellite, and to furnish the court-appointed prison monitor with a car no more than one grade below that provided to the commissioner of corrections. The various orders, amendments, and work plans bulked thicker than two Manhattan telephone directories.

As the Giuliani administration quickly learned, the decree interfered with even the most basic innovations in managing the prisons, turning the city into a perpetual, often disappointed, supplicant before the federal judge and, inevitably, the Legal Aid Society, the attorney of record for those incarcerated in city jails. The administration wanted to scuttle Mayor Dinkins’s plan to construct a single large facility in Rockland County to prepare food for all the city’s inmates, having figured out that the city could save $250 million by upgrading existing facilities. The decree stood in the way. The administration wished to ban jewelry that identified gang members, believing that it precipitated fights. The decree stood in the way. After bloody incidents during prisoner trips to the law library at Rikers Island, the administration wanted to keep prisoners already adjudged as violent in their cells to do their legal research. Again, the decree stood in the way.

When the city asked the court to modify the decree to allow these changes, the Legal Aid Society strenuously objected. Its lawyers pressed the court to keep in place the amended 1978 decree and challenged the city to prove the facts supporting its plans. Legal Aid reacted the same way to a motion that the city filed in December 1995 claiming that it had complied with the decree in ten of the 30 categories of obligations set out in the original decree and asking the court to delete them outright, along with the related amendments. The city argued that the decree’s particular requirements in these areas—for example, that prisoners get free telephone calls and newspapers, that the prisons use only licensed barbers, and that windows be washed four times a year—were no longer necessary to ensure the city’s compliance with the Constitution.

Trying to make these simple policy changes required reams of documents, lengthy legal briefs, oral arguments before the court, and affidavits and depositions from prison officials, prisoners, and experts. After months of litigation and negotiation, Legal Aid finally agreed to eliminate the provisions of the decree that mandated the Rockland County food facility and that prevented the city from banning gang jewelry. As for access to the law libraries, Legal Aid agreed that violence was a problem but argued that the city should have to make a second evaluation of each prisoner’s potential for violence. Judge Harold Baer Jr., who took over the city prison cases after the first judge retired in 1995, directed the city, as a test, to continue law library visits for five weeks, during which time 11 slashings and other violent episodes took place. Judge Baer promptly granted the city’s motion to provide law books to the designated violent prisoners in their cells (albeit with an excruciatingly specific order on how they were to be delivered).

To understand how courts have managed to stray so far beyond their province of enforcing rights, suppose that a judge finds that a prison violated the Constitution because the food provided fell below government nutritional and sanitary standards or didn’t conform to a prisoner’s religious dietary requirements. Preventing future violations requires only that the food be made to meet the minimal standards implicit in the Constitution. How to do so is up to the prison officials. But it is the court’s job to make sure that they do obey the Constitution, and for that reason, court orders usually do not stop at a simple direction that the food be nutritious. The judge may conclude that the order should go into specifics such as the content of the diet, the method of food storage and preparation, the construction of the kitchen, and the schedules for contracting with suppliers and hiring cooks. The rationale for such specificity is that milestones make it easy to detect violations, that without a precise road map government would fail to reach its constitutional objective, and that, as lawbreakers, government officials cannot be trusted.

Once the parties begin to negotiate, the discussion inevitably begins to stray further from fixing the constitutional violations. Suppose that the prison officials’ plan for permanently remedying the constitutional violation requires the reconstruction of the prison kitchen, a two-year project. To secure plaintiffs’ acquiescence in the delay, the local prison authority might offer an interim food plan that calls for cooked food to be trucked in at high cost from a particular vendor, or one that meets a certain standard of quality, or one to be chosen later with the approval of the plaintiffs’ attorneys. Plaintiffs may have complaints that they would like taken care of as part of the bargain—for instance, that they should have more access to the law library or that unindicted and unconvicted prisoners shouldn’t have to suffer the indignity of having to wear prison garb. Result: court orders stray far from the original claims and become burdened with obligations that, even if reasonable or acceptable at the time, become oppressive to the administrations that inherit them.

This process creates opportunities for plaintiffs and judges to empower themselves at the expense of representative government. Even a judge who resists the temptation to impose his own policy preferences in the decree still gains personal oversight and control of highly visible governmental programs. The greater the specificity, the greater the control. Plaintiffs and defendants also find much common ground in specificity. Plaintiffs’ attorneys go from being outsiders to powerful insiders with a direct, personal voice in the actual management and control of governmental programs, free from even the normal restraint of real clients who pay the bills and must be consulted. Government officials like specificity, too, because they gain a trump card against budgetary or other constraints on what they want to do. A rigid milestone written into a court order becomes an untouchable funding mandate when the overall city budget must be cut. Judge Morris Lasker, who signed the original court order in the New York City prison cases, clearly perceived this reality. Noting that correction officials seemed to want more funding for the prisons than the mayor desired, Judge Lasker mused, “Sometimes I really wonder in this case whether the defendants themselves shouldn’t be represented by separate counsel, about whether there is a conflict of interest between the commissioner of correction, for example, and the mayor of the city.”

The upshot is the kind of decree often seen in cases against state or local government—a document setting out broad and wide-ranging duties, written with lawyer-like specificity and marked by multiple deadlines, all having the effect of locking in budgetary commitments and restricting choices by future officials.

Under court rules laid down by the U.S. Supreme Court, local governments can’t easily escape from such court orders. To modify one, the government must demonstrate an unforeseen change of circumstances that warrants a revision. If the parties were concerned about crowding in jails as they negotiated the court order, a subsequently enlarged prison population due to the advent of crack or toughened sentencing rules or better police strategy is arguably not unforeseen. This requirement gives a powerful lever to the plaintiffs to prevent the modification of court orders. They can argue three different theories at once: that no change of circumstance has occurred, that any change that did occur was foreseen, and that the proposed modification goes further than necessary to correct for the unforeseen circumstance.

To terminate a court order altogether, the local government must also meet a heavy burden; it must demonstrate compliance with the decree for an extended period, even when the underlying constitutional violations have ended. While there are examples of court orders ending based on compliance, most orders simply go on and on. One reason is that governments typically fear opening the question of compliance, for which they bear the burden of proof. Plaintiffs will challenge terminating the decree (and their role in enforcing it) by demanding evidence of compliance. Even the best-run governmental institutions dealing with human services, prisons, and the like cannot show a spotless record, and most do not come close to fulfilling every one of the multitude of exacting requirements imposed on them. Asking a court to terminate a decree opens governments to the risk of failure, entailing possible new obligations and public exposure of the compromises that are inevitably made in public institutions. As a result, governments generally allow a court order to lie unchallenged even when a credible claim could be made for terminating it in part or even in its entirety.

The Prison Litigation Reform Act frees one vital area of policy—prison governance—from these constricting rules. Whether entered into with a government’s consent or over its objections, a prison decree may be no broader than needed to remedy a proven violation of rights. And a decree now lasts not forever but for a strictly limited period of two years. For a decree to last longer than that, plaintiffs must show that the original problem still exists; if they can’t, a government defendant can ask the court to terminate the decree. In short, the act effectively places the burden of proof where it belongs, on the plaintiffs.

In New York, the Legal Aid Society tried to keep the act from sweeping away the city’s prison consent decree, claiming that the act itself was unconstitutional because Congress lacked the power to void a duly executed contract—which, they claimed, is what a consent decree really is. The conclusion from their argument was that the city had to hold up its end of the consent decree just as it had to make good on the bonds it gives to those who lend it money. Judge Baer quickly squelched their specious argument. Whereas bondholders give up money in exchange for bonds, plaintiffs give up nothing in exchange for a consent decree. Their right against unconstitutional prison conditions does not entitle them to any particular set of policies. If government acts within the Constitution and lesser laws, no group of citizens has the right to veto its decisions.

Though the problem of decrees against state and local governments might be fixed in any number of ways (as we suggested in “Government by Decree,” City Journal, Summer 1994), the limits on duration and content that form the heart of the Prison Litigation Reform Act are perhaps the most promising road to reform. Congress should broaden the act, and legislatures across the country should enact similar laws for state courts, which also issue their share of overreaching decrees.

In the first place, such a change would make courts recognize the primacy of democratic decision making and the bounds of their own authority, and would be more consistent with deep-seated norms of judicial conduct. Courts ordinarily refuse to take on political questions, even when invited to do so by the legislature. If a state legislature enacted a law requiring judges to decide how much money localities must spend for homeless shelters, judges would immediately invoke the doctrine of separation of powers and refuse to comply. This same principle should lead courts to curtail the damage that their orders do to the democratic process. Just because political officials have consented to a decree does not mean that a judge has done justice to democracy. Elected officials must have the power, within constitutional limits, to change policy in light of experience and the changing wishes of voters. This is such an important principle that the Constitution doesn’t even give one legislative session the power to bind the next, either by law or budgetary appropriation.

Second, the making of government policy in a democracy should be open to all those affected by it. In giving vast powers to plaintiffs’ attorneys, decrees exclude a range of other concerned parties. By contrast with the rough-and-tumble of the normal political process, in the courtroom only official litigants get to speak or offer evidence. Thus, absent from the New York City prison case were the city’s Board of Correction, which sets standards for the prisons; the City Council, which appropriates funds for the prisons and has oversight responsibility for them; the guards and the inmates victimized by their violent fellow prisoners, who are most directly affected by prison management; and, of course, the taxpayers, who ultimately pay the bills.

Even some plaintiffs find themselves without an independent voice under a decree. Often too poor and unorganized to hire an attorney, they must depend on their original representatives in the case, usually activists with their own idea of who the poor are and what they want. When the Legal Aid Society, invoking its authority as attorney for all public housing tenants under the 1971 Escalera consent decree, recently tried to stop the New York City Housing Authority from speeding up the eviction of drug dealers from public housing, many tenants were outraged. They felt that Legal Aid was far more sympathetic to the criminals in their midst than to their own claims as law-abiding citizens. Two of the largest tenants’ organizations, claiming to represent the overwhelming majority of residents, hired their own attorneys to file amicus briefs in support of the Housing Authority.

Finally, policies need to change to meet changing circumstances. Decrees that last for decades make it difficult for governments to solve new problems and to experiment with different means for solving old ones. To give just one example, Kay Hymowitz recently documented how the 17-year-old decree in the Jose P. case has forced the city to maintain an enormously expensive special-education program, despite its dubious value to the children it serves (“Special Ed: Kids Go In, But They Don’t Come Out,” City Journal, Summer 1996). Placing term limits on all decrees would keep such mistaken policies from being carved in stone.

Even before Congress and the state legislatures get around to curbing the baneful effects of judicial decrees, the courts can do a great deal on their own. There’s no reason why judges today can’t confine decrees to remedies for proven violations, provide for them to be reopened automatically on a fixed date, and terminate them when plaintiffs can no longer show ongoing problems. Such an approach would be far more consistent with our Constitution’s strong emphasis on the separation of powers.

Courts should try to force policy choices back into the hands of elected officials, where they belong. Legislators may not always think this a favor, of course. Decrees very often free them from responsibility for the most intractable problems, placing the burden of compliance on mayors or governors. But legislators are the ones with the political and budgetary power to forge real solutions, and they are far more likely than courts to appreciate both the limits of delivering on good intentions and the need to distribute limited resources reasonably among a wide range of worthy undertakings. And unlike courts, they are directly answerable to the people.

By exercising such restraint, judges would show a much deeper understanding of their constitutional duty. Under the U.S. Constitution and its cousins in the states, government is directly accountable to the people. When a court goes beyond its responsibility to protect rights and needlessly constrains political officials in matters of policy, it undermines that accountability. Whatever other rights a court protects, it must always consider the fundamental right of the voting public to elect officials who will have some room to deal with complex issues like crime and welfare, education and housing. That is what democratic self-government means.

In upholding the Prison Litigation Reform Act, Judge Baer questioned its wisdom, claiming that the court’s 18-year dominion over New York City’s prisons improved them. Whether that’s true is beside the point. Judges should run prisons and other government programs only to the extent absolutely necessary to protect constitutional rights. When they go further, they cease to act like judges.


City Journal is a publication of the Manhattan Institute for Policy Research (MI), a leading free-market think tank. Are you interested in supporting the magazine? As a 501(c)(3) nonprofit, donations in support of MI and City Journal are fully tax-deductible as provided by law (EIN #13-2912529).

Further Reading

Up Next