The supreme court has now spoken. Prison overcrowding in California is not only a crisis; it is unconstitutional. The California prison system, the largest in the country and the most at risk legally, operated at almost 200 percent of rated holding capacity, with more than 160,000 inmates. A special three-judge federal court had found that these conditions, in which suicides, violence, and lack of health care and other social services were endemic, violate the Eighth Amendment’s “cruel and unusual punishment” clause and ordered the state to reduce its prison census by as many as 46,000 inmates, to “only” 137.5 percent of capacity.
In May 2011, the Supreme Court, in a 5-4 decision (Brown v. Plata), affirmed the lower court, upholding what dissenting justices called “perhaps the most radical injunction issued by a court in our Nation’s history,” one based on “a judicial travesty.” The razor-thin Court majority, adopting the lower court’s findings, stated that transfers of prisoners to other states or county jails, construction of new prisons, additional spending, and hiring of additional prison staff would not suffice to satisfy the constitutional standard. The Court did not mandate that California change its laws — for example, by requiring the state to decriminalize certain illegal conduct, or to reduce the length of the minimum sentences that state law now requires its judges to impose. But the Court nevertheless went quite far. It held that population reduction — a euphemism for the mass releases of inmates convicted of serious crimes — must be a significant part of the remedy. Specifically, it ruled that the prison census must be reduced by 46,000 prisoners within two years.
How did we get to this point? How did prison capacity become so egregiously overstretched that the Supreme Court would impose such an extreme remedy on the state? There is no dearth of explanations. Decades of “tough-on-crime” politics, coupled with genuine concerns about high recidivism rates and studies showing the doubtful effectiveness of most rehabilitation programs, have led states to demand longer sentences. Many laws have made the longer sentences mandatory, depriving the sentencing judges of discretion. California and some other states have enacted “three-strikes” laws, which have increased the number of very long sentences that courts must mete out, sometimes for relatively minor crimes. The decades-long war on drugs has helped to over-populate federal and state prisons and jails, as has stepped-up enforcement against immigrants guilty of immigration-related felonies.
The overcrowding problem would be much easier to solve, of course, if the significant rise in incarcerations were entirely a bad thing — if they represented only costs, not benefits. But the social benefits of putting serious criminals in prison appear to be very substantial; the only real question is how large they are. Because prisoners on average would commit a large number of crimes if they were free — estimates range as high as twelve a year — more incarcerations for longer times are partly responsible for the unprecedented decline in crime rates that the U.S. has enjoyed in the past two decades. According to James Q. Wilson, perhaps the leading student of this phenomenon, the weight of scholarly opinion is that prison sentences do deter crime and explain “much but not all of the American crime drop.”
Despite the Court’s unequivocal mandate to reduce prison overcrowding, however, California is unlikely to come even close to meeting it in the foreseeable future. Consider the constraints on what California (and other states) can do to meet the Court’s far-reaching demand:
New prison construction designed to expand holding capacity is more difficult than ever due to a combination of factors. Most states are in severe fiscal crisis, which raises their costs of borrowing. Voters resist new taxes. Fierce nimby reactions occur among people living near proposed new prison sites. California already spends $9 billion (over seven percent of its entire budget) on corrections, the fastest-growing portion of its budget after Medicaid. California has hired more staff and rented prison space in other states but this simply spreads the problem around and is approaching its limits. Double- and even triple-celling prisoners, which the state already does, encourages violence, and politically powerful guards’ unions oppose it.
Releasing prisoners is the most straightforward remedy for overcrowding, and where prisoners have accumulated sufficient “good time” credits to qualify for early parole, this is a routine policy. Not surprisingly, the Court endorsed it, along with electronic monitoring and diversion of low-risk offenders to community programs. But concerns about recidivism — and particularly about violent crimes by released prisoners — impose severe limits on premature releases of convicts. Instituting more generous good time credits to allow earlier parole of those who can safely be released and monitored is desirable, but California is already doing so, and with a felon recidivism rate of 58 percent within three years (70 percent if nonfelons are included), it is surely running out of low-risk candidates for early release. This reform, then, must inevitably reach its limit at the point where those being considered for release pose too high a recidivism risk. The precise point at which that risk becomes socially unacceptable is of course a matter of difficult judgment.
We should not be surprised that officials who must weigh the prospect of even a small number of publicized violent crimes by released prisoners will tend to be risk-averse about releasing them. Their incentives, rather, will be to resolve all doubts and exercise all discretion in favor of continued incarceration. Who can blame them? In his dissenting opinion in Brown v. Plata, Justice Samuel Alito described a court-mandated prisoner release program in Philadelphia in the 1990s:
federal courts enforced a cap on the number of inmates in the Philadelphia prison system, and thousands of inmates were set free. Although efforts were made to release only those prisoners who were least likely to commit violent crimes, that attempt was spectacularly unsuccessful. During an 18-month period, the Philadelphia police rearrested thousands of these prisoners for committing 9,732 new crimes. Those defendants were changed with 79 murders, 90 rapes, 1,113 assaults, 959 robberies, 701 burglaries, and 2,748 thefts, not to mention thousands of drug offenses.
Decriminalizing nonviolent crimes, particularly by legalizing drug possession, would reduce the prison population, and there are strong — but highly controversial — arguments in favor of doing so. But what legalization means is unclear, and its effects on prison overcrowding should not be exaggerated. According to Mark Kleiman and Jonathan Caulkins:
Contrary to widespread belief both in the United States and elsewhere, American prisons are not filled with people whose only crime was using drugs (as opposed to helping to distribute them). There are more than 1 million arrests per year in the United States for drug possession (of which a very large proportion are for possession of cannabis), but few of them result in prison time, or even jail time following a conviction. . . .The exceptional user sentenced to prison by a judge for possession alone is usually a courier carrying very large quantities, is already on probation or parole for another crime, or has a long record of non-drug crimes. Among prison inmates, those doing time for simple possession have more prior convictions for non-drug crimes than those doing time for property or violent crimes.
Shorter prison sentences are surely warranted in many cases, with little risk of weakening deterrence. As Wilson notes, “When voters demand longer maximum sentences for a crime that has recently made the headlines, they sometimes ignore the fact that many people who commit that crime do not go to prison at all and those that do may often be kept in prison long after there is any chance of their repeating the crime.” Still, there is a very strong case to be made for restoring the discretion over sentencing that judges traditionally enjoyed until it was restricted by legislatures that accused them of being too “soft on crime.”
Many states face California’s problems. The fiscal stakes in relieving prison and jail overcrowding are enormous, and these stakes, more than fear of legal liability, probably drive the efforts to reduce it. Corrections spending represents a large portion of state budgets, quadrupling in just the past 20 years and now accounting for one of every fifteen state general fund discretionary dollars. Nearly 90 percent of corrections spending goes to prisons.
High incarceration rates translate into high costs. One recent study claims that a reduction by one-half in the incarceration rate of nonviolent offenders would lower correctional expenditures by $16.9 billion per year and return the United States to about the same incarceration rate we had in 1993 (which was already high by historical standards). The large majority of these savings would accrue to financially-squeezed state and local governments, amounting to about one-fourth of their annual corrections budgets. As a group, state governments could save $7.6 billion, while local governments could save $7.2 billion. State corrections budgets, once largely untouchable, are now being slashed. According to one study, “The budgets of at least 26 state departments of corrections have been cut for fy2010, and even those whose budgets have not been cut are reducing expenditures in certain areas . . . Each year, the decisions will become more difficult.”
Even so, politicians will find it hard to sell such reforms to a deeply skeptical, crime-fearing, and tax-weary public. As a result, severe and chronic prison overcrowding will persist, leaving California and other states with a very serious social, political, and legal problem. This stalemate, moreover, must inevitably impugn the power and perceived legitimacy of the federal courts, which seem impotent in the face of such long-standing constitutional violations. It is telling that at oral argument in Brown v. Plata, the Supreme Court justices, while rebuking the state’s lawyer for California’s failure to meet the constitutional standard, was unable to suggest any ideas about what else the state should do other than spend more money on prisons. Yet, in its decision, the Court conceded, as the lower court had found, that any remedy requiring significant additional spending by the state was “chimerical” in light of its “unprecedented budgetary shortfall.”
In the wake of the Court’s decision, the state has adopted an overcrowding reduction plan whose adequacy is already in doubt. Under this so-called Public Safety Realignment Plan, which began on October 1, 2011, the state has transferred responsibility for incarcerating and supervising the parole of low-level offenders to the counties. Those convicted of crimes defined as nonserious, nonviolent, and nonsexual will, with some exceptions, serve their sentences in county jails instead of in state prisons, for terms of up to three years. This approach is imposing immense fiscal and administrative burdens on county systems, as they are now also responsible for supervising the parole of inmates convicted of nonserious, nonviolent crimes when they are released from jail or prison under the plan. The county jails must now also house any convicts whose parole is revoked. More than 10,000 convicts have been shifted to out-of-state prisons, with presumably more to come. California also plans new in-state construction projects that will add more than 50,000 new beds. New programs are being instituted for early release of nonviolent female and medically disabled prisoners.
Despite these efforts to increase capacity and reduce the inmate population, the state has almost certainly failed to meet the court’s requirement that it bring the inmate population down to 167 percent of capacity by the end of 2011. Data from December 7, 2011, showed the prisons still operating at 170.7 percent of capacity, and some facilities had much greater overcrowding, with several still exceeding 190 percent of capacity. California’s counties (some of which are under their own court orders to reduce overcrowding and improve conditions) are of course bearing the brunt of these new developments and are taking extreme measures to deal with them. Riverside County, for example, plans to charge inmates $142 per night. Fresno County is no longer incarcerating those who violate their parole conditions. Other counties are releasing inmates early — and while they insist that those so released are not dangerous, this is more wishful thinking than firm prediction.
Operating under these intense legal, political, and budgetary pressures, policymakers must search desperately for other ways to reduce prison overcrowding until the necessary but politically elusive structural and policy changes can be made. Under these difficult conditions, any policy that promises to reduce overcrowding without undue risk to public safety deserves serious consideration. In what follows, I propose such a policy. Simply stated, the federal government should deport some immigrant criminals before they enter prison, not after. This would seem to be a no-brainer.
Deportation and the law
Consider this fact: in 2009, more than 25 percent of federal prisoners and a large number of state prisoners were noncitizens and have been convicted of the kinds of crimes that render them immediately deportable once their convictions are final and affirmed on appeal. (Immigration law gives them few if any defenses.) The government wants to remove most of these easily-deportable criminals; indeed, federal law already requires expedited removal at the end of their prison sentences. But the law permits such deportation only after they have served their full sentences in our prisons. Given a level of overcrowding today that may (as in California) necessitate the release of many citizen criminals before prison officials think they are safe to parole, it would obviously make sense to deport some immigrant criminals before they consume scarce prison resources that average over $45,000 per year per inmate. Doing so would immediately reduce the need for hard-pressed officials to release prisoners whom they would not otherwise release because of the recidivism risks they pose.
Unfortunately, federal law poses a major obstacle to this remedy. Since 1917 — long before our prisons were unconstitutionally overcrowded and before immigrants convicted of serious crimes contributed significantly to this problem — the immigration statute has required that criminal immigrants complete their prison terms before they can be deported. Congress has not reconsidered this imprisonment-before-deportation policy in light of changed conditions.
Indeed, Congress complicated this problem when it enacted the Prison Litigation Reform Act (plra). Although this statute was designed to limit federal courts’ power to order the states to undertake certain far-reaching remedies — including the early release of prisoners into the community — it also provided that such remedies could be justified if the court found that no less intrusive measure would effectively remedy the constitutional violation at issue in the case. In Brown v. Plata, the Supreme Court found that the California prison system was so egregiously, persistently, and unconstitutionally overcrowded that the lower court’s prisoner release order was permissible even under the plra’s limitations on judicial remedies.
The 1917 law requiring imprisonment before deportation was amended in the 1990s to provide for its own exception. A convicted deportable criminal can be deported earlier — that is, before serving his full sentence here — if he was convicted of a nonviolent offense and if an appropriate official (the attorney general or the chief of the state prison system) requests such an earlier deportation. Note, however, that this exception is limited to nonviolent criminals, precisely those who in the aggregate consume the most prison resources, and that correction officials almost never use the exception.
Three factors probably explain the statutory limitation of the exception to nonviolent criminals and officials’ failure to invoke it: the fear of illegal re-entry by recidivists, diplomatic constraints, and political factors. (Interestingly, the United Kingdom encourages early removal of immigrant criminals despite the existence of similar conditions and arguments there.)
Illegal re-entry. Unfortunately, criminals’ countries of origin may well be reluctant to repatriate violent criminals or, if they do take them back, may fail to imprison them for very long, or at all. In that case, many will seek to reenter the United States illegally and resume their depredations against American society. Given the traditional porosity of our long borders and fecklessness of border enforcement, this is a perfectly realistic concern.
The magnitude of this risk of illegal reentry, however, depends, first, on the receiving country’s willingness to incarcerate the criminal for the full period of his U.S. sentence, monitor him when he is on parole, and so forth. This commitment in turn depends on the country’s diplomatic relationship with the United States and on the fiscal and other incentives that the United States can use to induce that country’s cooperation — factors that I discuss below. But whatever the period of incarceration in the criminal’s home country, his ability to return to the United States thereafter will of course depend on the efficacy of our border enforcement — and failing that, of our interior enforcement.
In the past, undocumented immigrants have found it relatively easy to cross the southern border, albeit at significant cost in terms of “coyote” fees, risks to life and health, opportunity costs at home, and the frequent need for multiple attempts before success is finally attained. Since 2007, however, the number of illegal entries has declined substantially due to some combination of reduced job opportunities in the U.S., an enlarged and more technologically equipped Border Patrol, and several enforcement programs including criminal prosecution and prison time for previously-deported illegal entrants. If this trend continues, the likelihood of illegal reentry will be much lower than it was traditionally.
Diplomacy. Another impediment to policy change is the need to deal diplomatically with the criminal’s country of origin. Prisoner transfer treaties (ptts) are international agreements that permit countries to remove a foreign-national prisoner to the prisoner’s country of nationality to serve out the remainder of his sentences. ptts are negotiated in the context of an international law default rule: One nation’s courts will not execute the penal laws of another nation.
Typically, a ptt requires as a condition of the transfer: 1) the prisoner’s consent, 2) the receiving country’s consent, 3) a final judgment against the prisoner, and 4) dual criminality (the crime for which the individual was convicted must be a crime in both the sending and receiving countries). In many but not all of these treaties, the transfer must also have been initiated by the prisoner. In the United States, foreign national prisoners initiate applications for transfer, which are then considered by the Department of Justice. In determining whether to approve the transfer, the doj considers the prisoner’s likelihood of social rehabilitation and the likelihood that he will return to the United States, as well as other law enforcement concerns.
Congress imposed the prisoner’s consent requirement primarily out of a concern that deportation might impose a hardship on a criminal and his family. But such solicitude is unwarranted in the common situation where he lacks family ties here. Although Congress passed legislation in 1995 urging, but not requiring, the president to renegotiate ptts to eliminate the consent requirement in ptts, it remains. In addition, the ptts require not only the consent of the sending country, receiving country, and the prisoner, but also that the prisoner initiate the request. Consent, however, is not the only impediment to quick deportation of criminals. Mexico, for example, has sometimes been unwilling or unable to verify that it was imprisoning its nationals once they were returned there; prisoners may simply be released. In addition, securing the receiving country’s consent and assuring ourselves that deportees will serve their full sentences there and will not suffer persecution or other human rights violations is sometimes difficult.
Political factors. The politics surrounding efforts to reduce prison overcrowding and reduce prison costs are fierce. Taxpayers want relief from prison costs, of course, but they also fear more crime if criminals are released prematurely — a possibility that Justice Alito’s dissent in Brown v. Plata renders vivid. This presumably explains why California officials have had so little success in early release. At the same time, localities where prisons provide scarce government jobs, prison suppliers, prison guard unions, and other interests with a strong stake in large prison populations oppose reductions. Resolving these conflicts is exceedingly difficult, as ex-Governor Arnold Schwarzenegger discovered in California and new-Governor Andrew Cuomo is discovering in New York.
A proposal: Earlier deportation
In some circumstances, sound immigration enforcement policy will favor having an immigrant criminal serve his entire sentence before being deported. For example, it might take that long for the United States to arrange for his actual deportation to the country of origin. The government might doubt the ability or willingness of any country to which he could be returned to keep him imprisoned for his full term, and might fear that if he is released there prematurely (or is not imprisoned at all), he would re-enter the United States illegally. It might want to keep him in our prisons so that he can be questioned about criminal activity by him or by others. Family ties, medical needs, or other humanitarian reasons might justify keeping him here, albeit under lock and key. There are surely other situations in which the government might conclude that deportation would be ill-advised. But just as surely, we will often want to deport criminals as soon as is practicable once their conviction and deportation order become final. Common examples are those guilty of violent crimes, those who are thought to pose a high risk of recidivism, those only recently arrived in the United States, those who are eager to be deported, and those without strong family ties here.
Changing the imprisonment-before-deportation rule. Congress should change the law in one of two ways. First, it could repeal the current imprisonment-before-deportation rule altogether. For those criminals whom the government decides to deport, it would have broader discretion about when to initiate deportation — as soon as possible, after the criminal has served his sentence, or at some point in between. Alternatively, Congress could adopt a new default rule: early deportation, perhaps even before any incarceration has occurred. Since a default rule is not mandatory, the government could still decide to imprison before deportation in any given case.
Under either approach, the law should specify the important individual and community values that the official decision maker must seek to protect and balance. (Some of these values will — or at least should — have already been brought into play in the immigration court before the final deportation order was issued.) Individual values should include: maintaining the integrity of families in the United States, especially those with children or spouses who are U.S. citizens; the length of time the immigrant has spent in the U.S.; whether he wants to be deported or not; and other factors such as acute needs for medical treatment. Community values should include: the seriousness of the offense, with violence and harm to the U.S. community weighing heavily in favor of early deportation; the risk of recidivism; the degree of prison overcrowding in his jurisdiction; the length of time remaining in his sentence, which would affect the U.S. incarceration costs to be avoided; the receiving country’s willingness and ability to ensure that he completes his sentence there; the likelihood of his reentering the U.S. illegally after that country releases him; and legitimate diplomatic considerations. Congress could specify the weight to be accorded to each of these criteria or leave that instead to official discretion. Finally, the prospect of being removed against one’s will to a prison abroad with harsher conditions and perhaps away from one’s family might require certain administrative process before early deportation can be imposed.
Renovating the PTT regime. In addition to changing the current imprisonment-before-deportation rule, Congress and the president should renovate the ptt system. They should deploy diplomatic and financial incentives to persuade the treaty partner to accept repatriation even without the prisoner’s consent, and to ensure that 1) the deported prisoner will actually serve the remainder of his U.S.-imposed term there, and 2) the conditions of imprisonment there meet the minimum health, safety, and other standards applicable to prisoners in the United States. In appropriate cases, the president should use his existing legal authority to arrange for the right of the U.S. to build, manage, and fund such prisons in the receiving countries. Needless to say, such diplomacy, which must proceed on a country-by-country basis, would be very challenging, and the incentives driving the receiving country’s negotiating position would doubtless be complex. Nevertheless, such a program could be immensely cost-effective for the U.S. and could be made attractive to the receiving country as well, depending on the value of the subsidies and other benefits that the U.S. would transfer to the receiving country along with the prisoner.
Even if such a program proved infeasible, however, my proposal to replace the traditional imprisonment-before-deportation rule with an administrative process to rationalize and support early deportation should be far superior to the status quo for those cases in which transfers pursuant to a ptt can indeed be effectuated. The criteria for deciding for or against early deportation would enable the government to consider the social risks of doing so — primarily the risks that the transferee state would not incarcerate him for the prescribed period or under humane conditions, and that he would reenter the U.S. illegally after release by that state. In those events, the U.S. might still deport him early if the criteria militated in favor of such a decision.
Overcrowded prisons constitute a very serious social problem of many dimensions: human rights threats, violence against prisoners and guards, breakdown of order and discipline, obstacles to prisoner rehabilitation and good health, fiscal integrity, and violation of constitutional or statutory rights. Despite being under immense and growing pressures to reduce their prison populations, California, other states, and the federal government have so far been unwilling or unable to reduce their prison populations significantly without creating what they believe are undue risks to public safety. A variety of policy changes have been proposed, some are being tried (especially in California, under the court order), and some of these may prove to be partly successful. Even so, the problem remains and in some states may be getting worse.
Early deportation of immigrant criminals would confer immense benefits on federal and state taxpayers (and on the remaining prisoners) at little or no marginal costs. To garner these fiscal advantages, however, the federal and state governments will have to work hard to dismantle the legal, political, bureaucratic, and diplomatic obstacles that have prevented early deportations. Although these governments have obviously not mustered the will to do so in the past, the situation may be changing dramatically, as extraordinary fiscal (and thus political and legal) pressures — not to mention the prospect of mandated releases of large numbers of potentially dangerous criminals before officials deem it safe to do so — suddenly make the previously unthinkable eminently thinkable. Even law-and-order conservatives are now supporting reforms designed to reduce corrections costs, reforms that were previously supported primarily by liberals. In hard times when difficult tradeoffs must be made, stranger things have happened.
My proposal to send back criminals who are deportable (not all of them) before they enter our prisons is certainly no panacea. The criminals’ countries of origin may not want them back and may refuse, contrary to international law, to repatriate them. And if their countries do accept them, they may not imprison them at all, or for long, so the criminals may try to reenter the U.S. illegally. This reentry threat may explain why the federal government has not updated the 1917 law or used its exception. But no policy to reduce prison overcrowding is risk-free, and the existing ones are both flawed and plainly inadequate. Diplomacy might persuade countries of origin to incarcerate their repatriated criminals, and better border enforcement (which has recently improved) can reduce reentries.
Prison officials and diplomats must try anything that the law permits and that promises to reduce overcrowding without endangering public safety. Accelerating the deportation of criminals who will almost certainly be deported eventually will not solve the overcrowding problem, but it will surely help.