Printed in 115 Harv. L. Rev. 1838, 1868 (2002) as part III of Developments in the Law: The Law of Prisons
Be sure and check the paper copy or Lexis/Westlaw before citing or quoting!
Private prisons are on the rise. Privately operated juvenile facilities — mostly community-based group homes or halfway houses — and federal adult halfway houses have been common in the United States since the 1960s. In 1979, private firms began contracting with the Immigration and Naturalization Service to detain illegal immigrants pending hearings or deportation. Private, large-scale investment in the construction and management of conventional prisons and jails dates from the mid-1980s. Prison privatization has been driven not only by the growing support among lawmakers and the public for private provision of traditional government services, but also by exploding prison populations resulting from stricter drug and immigration laws and changes in sentencing procedures.
By the end of 2000, there were 87,369 state and federal prisoners
in private detention
facilities in the
Comparative studies on the cost and quality of private and public prisons give reason to be cautiously pleased with private prison performance. The empirical evidence is consistent with economic theory, which predicts that with privatization, costs will fall and quality (however defined) may rise. The idealist could ascribe the satisfactory performance of private prisons to the power of market incentives; the cynic could point out that given public prisons’ bleak history and patchy present, private prisons perform satisfactorily compared to a rather low baseline. Each would be right.
Public prisons are not the most accountable of government systems; in fact, under certain circumstances, private prisons may be more accountable. In the qualified immunity context, recent Supreme Court decisions such as Richardson v. McKnight and Correctional Services Corp. v. Malesko have held private prisons to at least as high a standard of constitutional protection as public prisons. Judges’ and juries’ greater skepticism of private agencies than of government may also make private prisons more accountable; moreover, government oversight of private prisons may be less deferential than government oversight of its own operations. In addition, private prisons have substantially greater market accountability because they are concerned with winning new contracts and renewing old ones, and with avoiding both adverse publicity and drops in stock price. The continued promise of private prisons requires three concurrent innovations. First, evaluators must develop a rich set of performance measures, and prison data must be gathered and publicized. Second, the government must implement performance-based contracts that tie compensation to actual results. Finally, the government should maximize the efficiency gains from privatization and minimize opportunities for capture by institutionalizing competition between public correctional departments and private prison firms and making contract monitoring independent of both the public and the private sectors.
Critics have argued that statutes authorizing private prisons unconstitutionally delegate core government functions to private parties. This contention has never been tested in court, but such arguments seem dubious given the uneven history of the nondelegation doctrine and the Supreme Court’s recent decision in American Trucking Ass’ns v. Whitman. True, private prison officials “determine when infractions occur, impose punishments and . . . make recommendations to parole boards,” but as long as they implement well-defined correctional policy with sufficient oversight, this delegation seems unobjectionable on federal constitutional grounds.
One modern-day objection to private prisons stems from opposition to corrections and criminal policy generally: if the problem is the incarceration of too many people, making prisons cheaper or more efficient is a false solution and may exacerbate the problem. Another objection is the “expressivist” critique “that to turn over responsibility for administering prisons and jails to private, for-profit companies at some level compromises the legitimacy of the state’s exercise of its authority to punish.” This Part takes a frankly consequentialist view of private prisons and this does not address these critiques. Private prisons might be inherently problematic under some moral theories and acceptable under others (both consequentialist and deontological), but space constraints preclude engaging this debate.
What about the specter of corruption? Industry lobbies government, and regulatory agencies can be captured by the entities they regulate; the private prison industry is no different. Not only may private prison companies lobby for preferential treatment, they may also, as entities that directly profit from incarceration, influence substantive criminal legislation by supporting tough-on-crime candidates, scaring the public about crime, and advocating tougher sentencing. The story is plausible, but it does not explain current levels of prison privatization or modern-day demand for more and cheaper prisons because the forces leading to the explosive growth of the prison population substantially predate the modern growth of the private prison industry.
Moreover, though private prison companies do lobby state and federal governments, so do prison guard unions, which also benefit from increased incarceration rates and prison construction. Prison guard unions generally
contribute vastly more money to politicians than do private prison companies. The
Quite apart from whether political influence peddling distorts criminal policy, does such peddling weaken the case for privatization? Not necessarily, particularly when one considers different kinds of influence peddling: corruption and patronage. If a politician is corrupt and uses his power to extract money from the contractor, then privatization is likely to be inferior to public provision. Conversely, if a politician is involved in patronage and uses his power to pursue other political objectives, like serving politically powerful interest groups such as public employees’ unions, private provision is preferable.
1. Obstacles to Effective Assessment. — Effectively evaluating prisons requires specifying goals and objectives, developing measures and indicators, and collecting comparison group data. Unfortunately, the political process does not value rigorous evaluation highly. Some states require only cost evaluation; only a handful require comparisons of both cost and quality, often to comply with statutory targets. Some neglect evaluation altogether. What monitoring data exist are often inadequate for outcome evaluation.
As if that were not enough, “cost” and “quality” are not clearly defined. Public agencies and private firms measure costs differently. Public prison budgets usually exclude various central administrative and support expenses, such as medical, legal, and personnel administration services, which other state agencies typically handle. Private budgets include these costs but do not include the government’s costs of preparing and monitoring contracts. As for quality, definitions differ across studies, and quality is difficult to compare in any case. If one facility has fewer assaults but more escapes than another, is it better or worse?
Few studies are rigorous. Even reasonably good studies leave much to be desired. No study seriously controls for many important factors that influence misconduct rates, such as staff-to-inmate ratios, custody technology, correctional policies, or age and race of inmates.
Furthermore, comparative studies do not adequately address serious overcrowding problems. Overcrowding, which may increase inmate violence and the incidence of infectious and stress-related diseases, thereby contributing to unconstitutional conditions, is a serious problem in prisons and jails. Less expensive prisons allow for more capacity because the same prison budget can build more prisons, and increased capacity can relieve overcrowding. (Of course, cheaper prisons do not guarantee greater capacity, and greater capacity does not guarantee decreased crowding; still, it is reasonable to expect that cheaper prisons will not exacerbate crowding.) Thus, to the extent that private prisons decrease costs, privatization can improve conditions in both public and private prisons. Comparative studies between private and public prisons at a specific moment in time cannot register this across-the-board quality increase.
Finally, most studies do not analyze both cost and quality and thus are of limited value in assessing private prisons. Studies that do not look at both elements simultaneously cannot begin to analyze the costs and benefits of private prisons.
2. Evidence from the Studies. — Studies that look at cost or quality alone do, however, provide some information. The most rigorous studies find clearly positive cost savings. On the quality side, comparisons are trickier, as there is no single metric representing quality. But none of the more rigorous studies finds quality at private prisons lower than quality at public prisons on average, and most find private prisons outscoring public prisons on most quality indicators. Most of these quality studies do not examine cost, but as private prisons are not expected to be more expensive, this result belies statements in the prison literature that assume that cost reductions must come at the expense of quality. Indeed, the few methodologically sound studies that evaluate both cost and quality suggest that cost is no higher in private facilities and quality is at least roughly equivalent.
Finally, a 2000 study of
3. That’s Fine in Practice, but How Does It Work in Theory? — Private prisons fare rather well in quality comparisons, but why? Contracts are necessarily incomplete: because the government and the private provider can only describe a general service and cannot specify beforehand in full detail exactly how the contractor should provide that service, the contractor has wide latitude in running the prison. This latitude permits the contractor to cut corners, reducing costs by reducing quality. It is no surprise that early economic models of privatization predicted that private ownership would reduce both cost and quality.
But more recently, economists have observed that cutting corners is not the only way to make money. It is easy to assume that an aversion to out-of-pocket costs will deter firms from implementing a “quality innovation” — but this assumption ignores opportunities for contract renegotiation. Because private prison companies can suggest such innovations to the government and renegotiate their contracts (or, in the real world, include extra services in a higher bid), they can capture some of the gains from quality innovation. They therefore have greater incentive to innovate in this way than their public counterparts, who cannot capture such gains.
Thus, while economic theory predicts that costs will decline under private management, it does not necessarily predict the same of quality. Whether quality increases or decreases overall depends on whether cost-cutting or quality innovation has a greater effect.
This simple model does not take into account other factors that may strengthen the case for privatization. For instance, the government may observe a provider’s harmful cost cutting and hold it against the provider in future rounds of bidding or decline to renew the contract. If even a few private providers compete against each other, the government can seek an alternative provider once it observes harmful cost cutting, rather than having to retake control of the prison. In addition, because of inefficiencies in current prison practices, there may be opportunities for cost cutting that does not reduce quality.
4. How Does the Private Sector Save Money? — The private sector saves money in several ways. First, private companies save money at the design and construction stage. They can typically design and build prisons in half the time required for governments to do so, because they can avoid the layers of red tape that play a role in safeguarding against government corruption but are arguably unnecessary when the government purchases a service from the private sector (using a procurement process that itself has safeguards against government arbitrariness). Private firms are also usually free of purchasing restrictions and subcontracting quotas. Contracting out prison design and construction reduces costs by 15 to 25%.
Second, private companies save money at the operation stage. The main savings come from reducing labor costs, both through lower wages and through more efficient use of labor. A private firm that had a role in designing a facility would be likely to use innovative design techniques that could minimize the number of guards required to monitor inmates. Moreover, because they are not bound by civil service rules in managing their personnel, private prisons use roughly one-third the administrative personnel of government prisons and use incentives to reduce sick time and consequent overtime expenditures. Private firms also save money by “maintaining tight control of inmates and keeping them well-fed and occupied with work, education, or recreation.” Finally, private firms are free from many bureaucratic purchasing rules and can often buy supplies at lower cost than the government.
Abuses happen in any system. But how do different systems react to abuse when it occurs? While there is no systematic information about the reaction to prisoner abuse in public and private prisons, case studies may provide a flavor of the accountability mechanisms at work.
1. Private Prisons’ Legal Accountability. — The traditional hostility of juries toward corporate defendants, private prison guards’ inability to claim qualified immunity in § 1983 civil rights suits, and courts’ unwillingness to defer to the judgment of corporations all increase private prisons’ legal accountability relative to public prisons.
(a) Jury Hostility. — Empirical studies have found that juries are more likely to award large verdicts against corporations than against governments. Jury
hostility also affects settlements, which are made in the shadow of expected recovery amounts at trial. Consider, for example, the case of the
(b) Qualified Immunity: The New Bifurcated Regime. — A danger inherent in privatization is that public responsibilities will be performed by private individuals without effective oversight. Recent cases, however, have sought to avoid this pitfall and, in some ways, hold private prisons to an even higher standard than public prisons. Under Harlow v. Fitzgerald, government officials (including prison guards) performing discretionary functions are shielded from liability for civil damages if their conduct does not violate “clearly established statutory or constitutional rights of which a reasonable person would have known.” In Richardson v. McKnight, however, the Supreme Court held that private prison guards do not enjoy this qualified immunity. Thus, McKnight bifurcates the treatment of public and private prisons in a way that makes private prisons more, not less, accountable.
Does this dichotomy make sense? If governments are cost-minimizers, it does not. In the presence of liability, cost-conscious governments would hire timid guards, who require less indemnification, for public prisons, and would choose timid contractors, who submit a lower bid, for private prisons; thus, the argument for qualified immunity would not depend on whether the prison is public or private. But governments often do not minimize costs. They often face soft budget constraints, may award contracts to friends or political allies, and may monitor prisons more or less depending on their political ideology. Because government officials’ motivations are complex, it is difficult to predict how the effects of qualified immunity will vary as between public and private prison operators. The efficiency arguments for only extending qualified immunity to public guards are thus somewhat indeterminate.
Considerations of compensation and accountability, however, cut against immunity for both public and private prisons. On the compensation side, immunity presumably makes a difference in some cases, and where it does, it may prevent victims from being compensated. On the accountability side, immunity allows prisons to externalize the cost of constitutional violations onto prisoners — not a particularly well-represented segment of society.
Nor does qualified immunity seem necessary to reduce the drain on the court system. Most § 1983 prison lawsuits are considered frivolous, and almost all fail before getting to the merits. Qualified immunity would probably not change inmates’ filing behavior, and even if it did, it is not clear that this would be desirable, because our system deliberately tolerates a certain amount of meritless litigation to avoid missing the occasional worthwhile claim. A regime of liability, with high procedural hurdles (such as requirements of exhaustion of administrative remedies) or penalties for frivolous litigation (such as the frequent-filer penalties in the Prison Litigation Reform Act), may better serve the goals of compensation and accountability for both government and private actors.
Thus, given the bifurcated regime established in McKnight, private prisons are, if anything, more accountable for their constitutional violations than are public prisons. The presence of this additional judicial check should in turn increase private prison quality.
(c) Other Legal Accountability Mechanisms. — Not only may privatization make juries more skeptical of prisons and guards more susceptible to suit, but it may also make judges less deferential. Courts have traditionally been deferential to the government in prison suits because of the common judicial unwillingness to second-guess the political branches. Though this unwillingness is understandable, deference is the enemy of robust accountability. Yet courts have often been hostile to private, for-profit delegations; while the nondelegation doctrine is almost never applied, judicial hostility manifests itself through more vigorous due process scrutiny. As administrative power moves to private prison corporations, courts may abandon their deference to prison officials, because conflicts of interest will appear more obvious and contractual terms will necessarily be incomplete.
In addition to litigation, government monitoring adds a further layer of review. While public prisons have government monitors, monitors of private prisons (even if captured to some degree) are likely to be more independent than monitors of government-run prisons.
2. Private Prisons’ Market Accountability. — Market mechanisms, such as governments’ ability to rescind or decline to renew private firms’ contracts, and more generally, the potential for bad publicity to cause a drop in firms’ stock prices, further increase private prison companies’ accountability.
(a) Contract Rescission. — A private prison may have its contract rescinded. This possibility is not always as easy as it sounds — the private prison industry is a somewhat concentrated oligopoly, though that may change with increased privatization. But as long as more than one firm is operating and the government continues to run part of the prison system, someone will be available to take over a dysfunctional prison, making the government’s threat to rescind a contract somewhat credible. At any rate, even such imperfect discipline is more difficult to impose on public prisons — the government cannot take over its own prison except by firing civil servants, and it cannot have a private firm take it over except by opening a new bidding process, which is more difficult than finding someone to take over an existing contract.
In 1995, for example, an investigative reporter and a surprise inspection
revealed appallingly crowded, unsanitary, and abusive conditions
at the Bowie County Correctional Facility, a private “warehouse turned prison facility”
At the Seal Beach City Jail, operated by the private Correctional Systems, Inc. since 1994, two private prison guards were indicted in August 2001 for allegedly “arranging and concealing an attack on a drunken inmate who was singing boisterously in the jail’s detoxification cell.” Within ten weeks of the attack, both guards had been fired; one guard was charged with federal civil rights violations and the other with being an accessory after the fact for trying to conceal the attack. The first guard was sentenced to over four years in prison. This incident prompted the city to review its contract with CSI and to consider other options.
(b) Adverse Stock Price Effects. — Private corporations are sensitive to drops in their stock prices. Contract rescission, as well as the possibility of lawsuits with high damage awards, affects profitability, and perceptions of a company’s profitability are reflected in the price of its stock. Thus, a private corporation is punished financially for bad news, and possibly for mismanagement that may impose costs in the future. For example, the INS detention center in Elizabeth, New Jersey, run by Esmor Corrections, erupted in a massive riot in 1994 because the company had continuously cut corners on food and facility upkeep. Esmor’s stock price dropped from $20 a share to $7 after news of the riot became known.
While fear of stock price drops may make private prisons conceal their problems — CCA has been accused of covering up escapes from the Youngstown prison described above — such secrecy has its limits, because escapes, riots, and prisoner litigation are hard to cover up. Following disclosure of the problems at the Youngstown prison, CCA’s stock hit a one-year low. A more plausible story is that private companies are more concerned with keeping their stock prices high over the long term by insisting on sound management and that guards and wardens can be encouraged to act responsibly through stock ownership in the company.
(c) Greater Market Flexibility. — Not only are private prisons under greater market pressure to keep conditions from getting out of hand, they are also more able to change because they are free from some of the constraints of government management. In July 2001, for instance, four hundred prisoners in a CCA prison in Kentucky started a riot in which “inmates set mattresses on fire and toss[ed] TVs and toilets through the windows. Two weeks later, CCA fired the warden and his top assistant, citing ‘policy violations.’” Government employment protection, in this case, would have been an obstacle to disciplining the offending prison officials.
3. Are Public Prisons Any More Accountable Than Private Prisons? — (a) The Dynamics of Organizational Stasis. — Susan Sturm describes the “dynamics of organizational stasis,” which “lock in the current conditions in prisons” (both public and private) and “disable regular prison participants from achieving institutional self-correction”:
Even in the face of riots, violence, or public exposure of brutal conditions, legislatures frequently provide minimal support for change. The familiar study commission or task force conducts a public hearing or investigation, culminating in a report with recommendations that are infrequently enacted into law. Even when the legislature does respond to abuse with legislation, there is little accountability for its enforcement and the administration may comfortably ignore it without legislative sanction.
Much of Sturm’s description of prisons applies equally to public and private prisons — inmates are generally powerless to change prison conditions, and both guards and administrators often resist reform. Other elements, however, apply primarily to public prisons: unions concentrate on preserving their members’ pay, benefits, and seniority, and the budgetary process, in which this year’s allocation is presumed to be the baseline for next year’s, supports the status quo. It is well known that public bureaucracies have different incentives and attitudes toward change than do private companies. Companies concerned about winning a bid, retaining their contracts, maintaining their stock price, or just being marginally more protected against prisoner lawsuits seem more likely to overcome their institutional stasis.
(b) Corcoran State Prison. — Contrast the above factors, which promote accountability in private prisons, with the sad tale of Corcoran State Prison, a California public prison, in which guards killed seven inmates and wounded forty-three others between 1989 and 1995. Rival gang members fought in human cockfights overseen by prison guards; officers abused and beat inmates; problem inmates were disciplined by being locked in a cell and then beaten or raped by an inmate enforcer dubbed the “Booty Bandit.”
Conditions at Corcoran are not typical of public prisons. What is striking about Corcoran, however, is how little was done to alter these conditions even once they were discovered. The Corcoran episode serves as a reminder of the weakness of public prisons’ accountability. When replacing one’s service provider is difficult — and especially when the provider is identified with the purchaser, capturing the policy advice process — moral outrage at abuses can only get one so far.
In 1994, the FBI began its probe of Corcoran, and in 1996, prompted by media attention, the state launched a pair of investigations, one by the Department of Corrections and another by the Attorney General’s office. These investigations were stymied by political pressure from the governor’s office and the prison guard union, and the state probes yielded not a single criminal charge. Instead, state investigators “spent considerable manpower trying to dig up dirt” on the whistleblowers who had reported these abuses to the FBI. One whistleblower, Richard Caruso, was the only officer disciplined as a result of the state investigations (though he sued and received a large settlement from the state, the largest amount ever given to a whistleblowing officer in California). Another whistleblower changed jobs within the prison system because of alleged retaliation by prison officials (and recovered nothing). The FBI inquiry ultimately resulted in charges against eight officers, but all were acquitted in June 2000.
While private prisons today provide acceptable quality at a lower cost than do public prisons, they will only continue to do so as long as their buyers — federal and state governments — care. Bad things still happen at private prisons. Private prison companies have not always been forthcoming with information about the types of inmates they are holding, and reports of maximum-security prisoners being housed in portions of private prisons designed for the general population are not uncommon. Effective judicial, legislative, and administrative oversight continues to be necessary. Perhaps the scope of prison litigation should be expanded to keep both public and private prisons honest, but there is no substitute for performance contracts that encourage quality improvements, effective monitoring, and information gathering and disclosure.
1. Performance Measures. — Susan Sturm notes that prison officials often know little about their own operations. Prison administrators gather some information to prove the prison’s compliance with court orders, but it would be preferable for them to do so without being forced. One way to increase private prisons’ disclosure, at least in the case of federal institutions, is by subjecting them to the same Freedom of Information Act disclosure requirements as public prisons — though this, too, is a piecemeal solution.
Perhaps the environmental information regime can provide a model for the disclosure of prison information. The Toxics Release Inventory (TRI), which requires that industrial facilities report the release and transfer of specific chemicals, has been credited with dramatic reductions in pollution. In part because of the existence of these reporting requirements, much formerly fragmented and incomparable data has become standardized and can be aggregated to allow comparisons of different firms or states, or to track performance over time.
Of course, TRI mandates only that information be gathered and made public, not that anything be done with that information. But the mere availability of information, whether environmental information about industrial facilities or correctional information about prisons (floor space, number of violent incidents, recidivism), has real effects. First, “you manage what you measure” — administrators have a natural desire to improve what they have data about. Second, information gathering encourages peer monitoring within an industry, and industry self-regulation can be a valuable supplement to other forms of monitoring or control. Third, information is valuable to regulators, would-be regulators, industry actors seeking to stave off regulation, community monitors and “informal regulators,” capital markets, labor markets, and customers (in the prison case, state and federal government entities).
Information as regulation has its critics. Regulation proponents charge that information does not guarantee any action or results and that it is most useful in combination with other regulation. From the other side, critics charge that because information is highly contextual, any mandated information will be too simplistic and perhaps misleading, especially when presented at a high level of aggregation. But while particular information initiatives may be flawed, and while mere disclosure may not be sufficient, collecting and publicizing a rich (and meaningful) set of performance indicators will increase the incentives for improvement that are already inherent in the competitive contracting process.
2. Improving Performance Through Contract. — At best, contracts “represent potentially useful accountability instruments [and] vehicles for achieving public law values, such as fairness, openness, and accountability.” Private prison contracts should have specific terms, graduated penalties, and strong oversight by a “contract manager” working for the public agency; they should require private prisons to “observe minimal administrative procedures such as notice and hearing requirements,” and perhaps explicitly give inmates or surrounding communities third-party beneficiary rights, which would allow oversight through contract litigation when government oversight fails. States could also require, as contractual terms, compliance with American Correctional Association and National Commission of Correctional Health Care standards.
Moreover, states could mandate that private prisons provide the same training that is required of public prison guards, though requiring certain inputs is presumptively less effective than looking to outcomes where these are measurable. Contracts could require that private firms carry civil rights liability and other insurance; that they disclose conflicts of interest; that they allow access to records and entry to the facility by inspectors; or that they be independently monitored or audited by certified professionals. Finally, contracts could tailor termination rights and provide for easy amendment of contractual terms. In short, contract designers can be highly creative — and thorough — in writing accountability into contractual terms.
Most basically, corrections departments should move toward performance-based contracts. Ideally, performance-based contracts should “clearly spell out the desired end result” but leave the choice of method to the contractor, who should have “as much freedom as possible in figuring out how to best meet government’s performance objective.” These contracts also structure contractor payments to encourage the desired results, rewarding the contractor for improvements and penalizing it for poor performance or rising costs.
This approach seems feasible for corrections. The American Correctional Association is revising its accreditation standards to include performance measures, and the Office of Juvenile Justice and Delinquency Prevention is developing performance-based standards for juvenile correctional facilities. Performance measures for prisons could include process measures such as the number of educational or vocational programs, or outcome measures such as the Logan quality of confinement index, the number of assaults, or the recidivism rate. Governments could even require that contractors pay for elements that are often externalized, such as the cost of escapes. Because no single statistic adequately captures “quality,” and because focusing on any single measure could have perverse effects, performance-based contracts should tie compensation to a large and rich set of variables.
3. Awarding and Monitoring Contracts To Maximize the Gains from Privatization and Minimize Capture. — Accountability can also be improved by redesigning correctional agencies. Under the basic model of accountability, the public correctional agency sets contract terms, awards contracts, and monitors compliance. Capture is frequent, cross-fertilization rare.
But there are other models of accountability. In the United Kingdom, “the role of the chief inspector of prisons brings some external scrutiny to the prison system generally, including the private sector . . . . This aspect of the accountability system is vigorously independent, with no danger of capture to date.” However, the chief inspector’s reports are not binding, and the government has resisted attempts to give the office more teeth. In Florida, private prisons have their own regulatory authority that operates independently of that of public prisons; the private prison monitor is in turn independent of the private prison authority. This system minimizes capture, but it also minimizes cross-fertilization. Moreover, “[t]here is also a danger that the monitors may develop a loyalty to the Privatization Commission which in turn might inhibit their willingness to make public criticisms — a variant of the capture principle.”
Perhaps the most promising model is one suggested by British penologist Richard Harding, in which both public and private prisons are subject to, and monitored by, an independent body that takes over new prison projects from the outset and allows both public and private systems to bid on them. After the contract term expires, both public and private systems may bid on the prison again, so that facilities can move among private firms and between the private and public sectors, promoting both accountability and cross-fertilization. This model separates the roles of government as purchaser and provider, making it more difficult for service departments to capture the policy advice process and use their power to recommend themselves as service providers.
This Part has not engaged the “softer” arguments about privatization, but has rather taken correctional policy as given and prison privatization as ethically neutral in itself. Does political influence peddling weaken or strengthen the case for privatization? It depends whether corruption by corporations is worse than patronage of public prison guard unions — a question that calls for further research.
Turning to the cost and quality comparisons, what imperfect empirical evidence there is suggests that private prisons cost less than public prisons and that their quality is no worse; it is perhaps unsurprising that prison privatization behaves in this respect much like privatization of other state and local services. Moreover, there are many reasons to believe that private prisons are more accountable than public prisons — both because of heightened legal and market accountability for private forms and because accountability in the public sector is so limited. There are also numerous untapped opportunities for improving private prisons even further with richer information gathering and dissemination, performance contracts, and better monitoring. In short, despite all of their possible faults, private prisons are a promising avenue for the future development of the prison system.
 See Douglas McDonald, Elizabeth Fournier, Malcolm Russell-Einhorn & Stephen Crawford, Abt Assocs. Inc., Private Prisons in the United States: An Assessment of Current Practice 4–5 (1998) [hereinafter Abt Report].
 Abt Report, supra note 1, at 7–10; Judith Greene, Bailing Out Private Jails, Am. Prospect, Sept. 10, 2001, at 23, 26 (describing the sudden interest of the Federal Bureau of Prisons in privatization, and attributing the Bureau’s need for tens of thousands of new beds to harsher drug sentencing laws enacted in 1986 and to the 1996 Immigration Reform Act).
 Bureau of Justice Statistics, U.S. Dep’t of Justice, Bulletin: Prisoners in 1999, at 7 tbl.10 (2000) (reporting that there were 71,208 inmates in private facilities (5.2% of all state and federal prisoners), of which 3828 were federal (2.8% of all federal prisoners) and 67,380 were state (5.5% of all state prisoners)).
 For a table of state and federal statutory authority to outsource prison operation, see Charles W. Thomas & Sherril Gautreaux, The Present Status of State and Federal Privatization Law, at http://web.crim.ufl.edu/pcp/html/statelaw.html (last visited Mar. 18, 2002).
 Richard G. Kiekbusch, Jail Privatization: The Next Frontier, in Privatization in Criminal Justice: Past, Present, and Future 133, 135 (David Shichor & Michael J. Gilbert eds., 2001) [hereinafter Privatization in Criminal Justice].
 122 S. Ct. 515 (2001). See id. at 515 (holding that companies operating private correctional facilities, like agencies operating analogous public facilities, are not subject to civil rights suits under Bivens v. Six Unknown Named Agents of Federal Bureau of Narcotics, 403 U.S. 388 (1971)).
 See, e.g., Joseph E. Field, Note, Making Prisons Private: An Improper Delegation of a Governmental Power, 15 Hofstra L. Rev. 649 (1987). But see Ira P. Robbins, The Impact of the Delegation Doctrine on Prison Privatization, 35 UCLA L. Rev. 911, 915 (1988) (arguing that the case against private prisons on delegation grounds is “extremely close”).
 Greene, supra note 5, at 26 (“[L]ike the state legislators before them, members of Congress were madly building new prisons . . . , searching for cheap new private-prison beds, and refusing to consider changes in the draconian sentencing laws that were causing most of the increase in prisoners.”); Ahmed A. White, Rule of Law and the Limits of Sovereignty: The Private Prison in Jurisprudential Perspective, 38 Am. Crim. L. Rev. 111, 145 (2001) (arguing that all imprisonment is “intensely problematic and in many ways inherently irrational,” “dysfunctional,” and “socially malignant,” and that government should “wage its own wars against its citizens . . . in an obvious and maximally costly way”).
 Sharon Dolovich, The Ethics of Private Prisons 73–74 (Nov. 1999) (unpublished manuscript, on file with the Harvard Law School Library); see also Dan Markel, Are Shaming Punishments Beautifully Retributive? Retributivism and the Implications for the Alternative Sanctions Debate, 54 Vand. L. Rev. 2157, 2233–34 (2001).
 The consequentialist treatment does not abandon the moral high ground — consequentialism itself implies certain normative commitments. Even for a nonconsequentialist, private prisons may not be especially problematic. See, e.g., Markel, supra note 26, at 2234–40 (arguing that private prisons are consistent with the “confrontational conception of retribution”).
 See Richard W. Harding, Private Prisons and Public Accountability 42–47 (1997) (discussing capture or the risk of capture in private prisons in Australia, Florida, and the U.K.); id. at 159 (arguing that under the “basic model” of accountability, in which the public sector correctional agency is at the center of privatization decisions, “[c]apture occurs frequently”); David Shichor, Punishment for Profit: Private Prisons/Public Concerns 240–41 (1995) (noting the involvement of Tennessee state officials with the Corrections Corporation of America (CCA) while CCA was bidding on the management of Tennessee’s prison system in 1985); Greene, supra note 5, at 27 (describing links between the federal government and private prison companies).
 This very claim is made about prison labor contractors in the nineteenth
century. Id. at 37 (“[T]hree
governors [of Tennessee] during the 1870s and 1880s expressed a negative opinion
of the convict lease system but realized
that the abolition
of this practice would create a severe financial burden on the state, and therefore they continued, albeit reluctantly, this policy.”) (footnote
omitted); Beverly A. Smith & Frank T. Morn, The History
of Privatization in Criminal Justice, in Privatization in Criminal
note 11, at 3, 17 (“Prosecutions, sentences, and paroles
were all manipulated to ensure a supply of a disproportionately high number of black inmates,
in what some have seen as replication of or an economic replacement for slavery without
a capital investment in workers.”). The claim is also made
about modern-day prisons. See, e.g.,
Greene, supra note 5, at 23 (describing how the warden of CCA’s Tulsa Jail in
 See Mark Arax & Mark Gladstone, State Thwarted Brutality Probe at Corcoran Prison, Investigators Say, L.A. Times, July 5, 1998, at A1 (“[State investigators] had watched the [prison guard] union under president Novey ride the prison construction wave, growing from a kind of social club into one of the more powerful forces in the state, with a rank-and-file 27,000 strong.”).
 See Adrian T. Moore, Private Prisons: Quality Corrections at a Lower Cost 33–34 (Reason Public Policy Inst., Policy Study No. 240, 1998) (comparing correctional officers’ $1.5 million donations to Pete Wilson alone during his 1990 and 1994 California gubernatorial bids with private prison companies’ $150,000 total political contributions nationwide in 1995–1996). Straight by-the-numbers comparisons are not enlightening because private prisons remain smaller players than public prisons. The interesting point is merely that whoever provides prison services will seek to influence the political process.
 Dan Morain, Davis To Close State’s Privately Run Prisons, L.A. Times, Mar. 15, 2002, at A1, available at 2002 WL 2461282 (noting that the Correctional Peace Officers’ Association spent $2.3 million to help elect Governor Gray Davis); see also Arax & Gladstone, supra note 32 (listing donations of $5.2 million to state candidates from 1987 to 1998, including $667,000 to gubernatorial candidate Pete Wilson, $159,000 to Attorney General candidate Dan Lungren, and an additional $760,000 in 1990 to defeat Wilson’s gubernatorial opponent, Dianne Feinstein).
 See, e.g., Ohio Rev. Code Ann. § 9.06(A)(4) (West 2001 Supp.) (requiring that the contractor “convincingly demonstrate” that it can operate the facility and provide required services with at least a 5% savings over the projected cost to the public entity).
 See, e.g., Ariz. Rev. Stat. § 41-1609.01(G) (2001) (requiring the contractor to offer cost savings); id. § 41-1609.01(H) (requiring the contractor to offer services of “at least equal” quality); id. § 41-1609.01(K) (requiring a biennial comparative public-private performance comparison study); id. § 41-1609.01(L) (requiring a five-year cost comparison); Tenn. Code Ann. § 41-24-104(c)(2)(A) (1997) (requiring quality to be at least equal to that provided at state prisons); id. § 41-24-104(c)(2)(B) (requiring cost to be at least 5% less than the state’s cost); id. § 41-24-1-105(a) (mandating the establishment of objective performance and cost criteria); id. § 41-24-1-105(c) (requiring evaluation of performance, with the contract to be renewed only if the contractor provides “essentially the same quality” as the state at 5% lower cost, or superior services at “essentially the same cost” as the state, where “superior” is defined as 5% better and “essentially the same” is defined as within 5%). The Florida statute requires evaluation of both cost and quality, though only cost savings and not quality improvements are required by the statute. Fla. Stat. Ann. § 957.07 (West 2002) (requiring 7% cost savings, where a private entity’s corporate income and sales tax payments count as an offset to costs, and where the cost of services provided to the public entity at no direct cost by other government agencies is allocated to the public entity); id. § 957.11 (requiring evaluation of “costs and benefits” of contracts).
 Durham, supra note 36, at 67; see Robert B. Levinson, Okeechobee: An Evaluation of Privatization in Corrections, 65 Prison J. 75, 76, 88 (1985) (citing the vague or nonexistent goals specified in the 1982 contract to run the Florida School for Boys in Okeechobee, a secure facility for adjudicated delinquents; the contract focused on administrative inputs rather than outcomes or even outputs).
 For a summary of the cost and quality studies on private prisons and a brief discussion of which studies are more methodologically sound, see 2 Geoffrey F. Segal & Adrian T. Moore, Weighing the Watchmen: Evaluating the Costs and Benefits of Outsourcing Correctional Services 2, 3 & tbls.2A–2C, 9–10 & tbls.4A–4B (Reason Public Policy Inst., Policy Study No. 290, 2002). For a highly critical evaluation of all prison privatization cost and quality studies (even some of the ones the Reason report considered sound), see also Abt Report, supra note 1, at app. 2.
 Scott D. Camp & Gerald G. Gaes, Private Adult Prisons: What Do We Really Know and Why Don’t We Know More?, in Privatization in Criminal Justice, supra note 11, at 283, 285 (attributing the insufficiency of current empirical studies to a failure to control for structural factors and only weak attempts to control for age and race); see also Abt Report, supra note 1, app. 2, at 18 (“[T]hese background data are a sine qua non of valid institution performance comparisons. What may appear to be differences in institution performance may be nothing more than differences in the background characteristics of inmates. . . .”).
 See U.S. Gen. Accounting Office, Private and Public Prisons: Studies Comparing Operational Costs and/or Quality of Service 13 (1996) (“[I]t is important that any study focus on both operational costs and quality of service.”). Interestingly, in the GAO study itself (which is really just a literature review), only the Tennessee studies compare both cost and quality, id. at 23–25, and these are not particularly rigorous, see Segal & Moore, supra note 41, at 3 tbl.2B, 10 tbl.4B. Surprisingly, given the solid methodological discussion, the GAO did not examine the Louisiana State University study, the Arizona DOC studies, or the Florida OPPAGA study, which analyzed both cost and quality. Nonetheless, later writers have called the GAO report a “thorough review,” Greene, supra note 5, at 25, “one of the more comprehensive reviews,” Bureau of Justice Assistance, U.S. Dep’t of Justice, Emerging Issues on Privatized Prisons 26 (2001), and the “most comprehensive study to date,” Joel Dyer, The Perpetual Prisoner Machine: How America Profits from Crime 229 (2000).
 Id. at 3 tbl.2A. Most of the less rigorous studies also report cost savings, and none reports that private prisons are more expensive. Id. at 3 tbls.2B, 2C. The causes of private prison cost savings are explored below in Section B.4.
 See, e.g., Greene, supra note 5, at 25 (“[Quality] problems seem to be endemic to the enterprise [of private prisons] — a result, in great part, of the private companies’ mission to hold down costs.”).
 William G. Archambeault & Donald R. Deis, Jr., Executive Summary, Cost Effectiveness Comparisons of Private Versus Public Prisons in Louisiana: A Comprehensive Analysis of Allen, Avoyelles, and Winn Correctional Centers, Phase I (rev. ed. Dec. 10, 1996).
 Id. at 2–4. A 1997 study of Arizona’s correctional system, which compared a private prison to fifteen public prisons, found 14% cost savings in the private prison relative to the public prison average and 17% cost savings after adjusting for the private prison’s property tax payments. Charles W. Thomas, Comparing the Cost and Performance of Public and Private Prisons in Arizona: An Overview of the Study and Its Conclusions (1997), available at http://web.crim.ufl.edu/pcp/research/Ariz.html (last visited Mar. 18, 2002). But see Abt Report, supra note 1, at 45 (criticizing this study’s methodology).
 See Jean-Jacques Laffont & Jean Tirole, A Theory of Incentives in Procurement and Regulation 231 (1993) (noting that “[a] high concern for quality [requires] low-powered incentive schemes” because providers who care too much about cost will reduce quality).
 The Hart-Shleifer-Vishny model assumes that all cost cutting decreases quality — in other words, that the public sector already operates at the frontier of technological efficiency. Id. at 1133. But no one knows the most efficient way to run a prison; this information must be discovered. Markets can encourage the discovery of efficient methods; government bureaucracies often do not. See generally Friedrich A. Hayek, The Use of Knowledge in Society, in Individualism and Economic Order 77, 77–78, 84, 86 (reprint 1980) (1948).
 See Dyer, supra note 47, at 214 (noting that “[m]ost guards at state and federal facilities earn union-scale wages and receive both retirement benefits and health insurance,” while “their counterparts at private prisons . . . often earn as little as $7.00 an hour, with little or no benefits”); Greene, supra note 5, at 25 (citing “wages and benefits substantially lower than those in government-run prisons”).
 Id. at 18. Moore provides an interesting example of private sector cost savings: Virginia prisons once kept thirty days of food on hand in expensive warehouses. This practice possibly dated back to when prisons were remote and supplied by mule train, and was never questioned until a private company did something different. Id. at 19.
 There is no statistical evidence on jury attitudes toward corporations in the prison context, but more general jury studies suggest that juries tend to be turned off by a defendant’s wealth and by arguments that a defendant’s behavior was profit maximizing. See Daniel Kahneman, David A. Schkade & Cass R. Sunstein, Shared Outrage, Erratic Awards, in Punitive Damages: How Juries Decide 31, 40 (2002) (showing higher punitive damage awards for corporations with greater wealth); cf. W. Kip Viscusi, Corporate Risk Analysis: A Reckless Act?, 52 Stan. L. Rev. 547, 550 (2000) (discussing juror bias against companies merely for having undertaken a risk analysis); see also Greene, supra note 5, at 23 (reporting that in December 2000, a South Carolina jury awarded $3 million in punitive damages against a CCA juvenile prison for abusing a youth using a level of force “malicious to the conscience of mankind”).
 Alphonse Gerhardstein, Private Prison Litigation: The “Youngstown” Case and Theories of Liability, 36 Crim. L. Bull. 183, 198 (2000). CCA also paid class counsel of the city of Youngstown $803,000 in litigation expenses, and agreed to pay annual fees to Youngstown for ongoing monitoring. Id.
 See 42 U.S.C. § 1997e(d)(3) (capping attorney fees at 150% of statutory hourly rates); id. at § 1997(e) (barring recovery for emotional injury in the absence of physical injury). The limitation on recovery for emotional injury may not apply in many cases, but the attorney fee cap may be more significant in cases in which inmates seek damages. Even if the settlement by its terms included attorney fees, the total amount of the recovery would probably be lower than it otherwise would be because expected recoveries affect settlement amounts.
 In Correctional Services Corp. v. Malesko, 122 S. Ct. 515, 517 (2001), the Court held that private companies operating federal halfway houses, like their public counterparts, were not subject to civil rights suits under Bivens v. Six Unknown Named Agents of Federal Bureau of Narcotics, 403 U.S. 388, 395–97 (1971), though the offending officials could personally be sued under Bivens, which allows awards of monetary damages against federal agents for a violation of federal constitutional rights. Malesko, 122 S. Ct. at 517. Malesko held private operators to the same standard as public agencies.
 See id. at 126. Different prisons may react to liability in different ways; one might imagine reactions that either increase or decrease costs. For instance, a prison may save money by instructing guards to discipline inmates less. Or a prison may spend more money if guards, who would rather avoid litigation even if they are fully indemnified, always call for more backup before disciplining an inmate, to avoid the possibility of a violent confrontation that could lead to a civil rights suit. But under competitive bidding with a cost-conscious government, the cost-cutting reactions will tend to win out over the cost-increasing reactions. Because some forms of timidity decrease costs, while overzealousness presumptively increases costs, a liability regime fosters timidity.
 28 U.S.C. § 1915(g) (1994 & Supp. II 1996) (prohibiting a prisoner from suing in forma pauperis if, on three previous occasions while he was imprisoned, an action of his was dismissed as frivolous, malicious, or failing to state a claim, unless he is “under imminent danger of serious physical injury”).
 Dyer, supra note 47, at 200–01; Lipsher, supra note 101. When the prison crowding crisis that prompted Colorado to send inmates to Texas was over, Colorado returned its inmates to in-state prisons, some of which were private. Id.
 Id. at 204; see also id. at 211 (dubbing the tendency to cover up bad news the “Esmor effect”); id. at 221 (calling honest disclosure “financial suicide” for a private prison firm and interpreting private firms’ fiduciary duty to their shareholders as “an obligation to prevent bad press”).
 See id. at 214 (noting that some private firms offer guards and wardens “stock in the prison company in an effort to reduce turnover,” but arguing that such stock ownership exacerbates the Esmor effect described above, supra note 110).
 Id. (describing also how union officials told guards not to cooperate with the earlier county prosecution of two lieutenants who “delivered a Taser jolt to an inmate’s testicles in 1989 and engaged in a cover-up”).
 Dyer, supra note 47, at 206 (describing a case in which Texas authorities learned that a CCA facility was housing sex offenders from out of state only when a few of them escaped); id. at 207 (describing a case in which, without the knowledge of local authorities, a privately run Denver drug rehabilitation program accepted felons from out of state, one of whom, after being expelled from the program, allegedly raped and murdered a local resident); see also Gerhardstein, supra note 83, at 184–85 (reporting that the private prison in Youngstown did not inform local authorities of the transfer of maximum-security inmates to a medium-security prison).
 See Dyer, supra note 47, at 206 (noting that the CCA Houston facility was not designed for inmates with a “violent criminal history”); id. at 207 (reporting that at a Texarkana private prison, “murderers and other violent offenders from Colorado were . . . housed in the general population, despite the facility not being rated for maximum-security inmates”); Gerhardstein, supra note 83, at 184–85 (noting that at Youngstown, a medium security prison in Ohio, hundreds of inmates “were transferred directly . . . from the Maximum Security Facility at Lorton”).
 See Dyer, supra note 47, at 216; Nicole B. Cásarez, Furthering the Accountability Principle in Privatized Federal Corrections: The Need for Access to Private Prison Records, 28 U. Mich. J.L. Reform 249, 296–99 (1995).
 See Performance Standards Home: Performance-Based Standards for Juvenile Correction and Detention Facilities, at http://www.performance-standards.org (last visited Mar. 18, 2002); see also 1 Geoffrey F. Segal & Adrian T. Moore, Weighing the Watchmen: Evaluating the Costs and Benefits of Outsourcing Correctional Services 15–16 (Reason Public Policy Inst., Policy Study No. 289, 2002).
 See Harding, supra note 28, at 113–15 (describing Charles Logan’s index of prison quality, which includes measures of “security, safety, order, care, activity, justice, conditions and management”).
 See, e.g., Adrian T. Moore, Geoffrey F. Segal & John McCormally, Infra-structure Outsourcing: Leveraging Concrete, Steel, and Asphalt with Public-Private Partnerships (Reason Public Policy Inst., Policy Study No. 272, 2000) (discussing water and wastewater, solid waste, and highway and street maintenance, in addition to jails and prisons). For a wider-ranging argument in favor of privatizing more of the criminal justice system, see Bruce L. Benson, To Serve and Protect: Privatization and Community in Criminal Justice (1998).