 |
 |
 |
 |
 |
The Pros of Privately-Housed Cons:
New Evidence on the Cost Savings of Private Prisons
Matthew Mitchell
Foreword
Harry Messenheimer, Ph.D.
President, Rio Grande FoundationDuring the three years since its founding, the Rio Grande Foundation has been emphasizing the bad policies that have held New Mexico back compared to other states in the nation and region. Indeed, our constant harping on New Mexico's high tax rates and their economic consequences may have contributed to a major shift in thinking about tax policy in the state. Now we have good news: There is a major personal income tax rate reduction in the offing. There is still more good news. New Mexico has been on the cutting edge of innovation in privatizing its prisons. In this study research economist Matthew D. Mitchell conducts an interstate econometric test of the relative efficiency of private-run versus government-run prisons. The econometric test • demonstrates to state policy makers nationwide strong empirical evidence of potential savings from privatization and • provides them with a good estimate of the magnitude of those savings.
The econometric model consists of a cross-section comparison of 46 states' corrections outlays per prisoner as a function of the portion of their prisoners that are housed privately. To account for differences among states the econometric model includes interstate differences in labor market conditions (does the state enjoy right to work?), pay for prison guards (using entry level pay for state police as an indicator) and the amount of crowding (utilization of bed space). The results suggest that privatization saved New Mexico roughly $51 million in 2001. This equates to $9,660 per prisoner. We really can do something right!
States facing a budget crunch would do well to follow New Mexico's lead. Privatization can release money for tax reductions or other needs.While not our primary aim, another result of the study deserves emphasis: The magnitude of estimated saving per prisoner in 2009 for states enjoying right to work legislation was $9,365 per prisoner. That means the average state could save roughly 28 percent of its corrections budget by legislating right to work. One wonders how much could be saved elsewhere by right to work. Are you listening, New Mexico?
I. Abstract
Three-fifths of all U.S. states contract with private corporations to house a portion of their state prisoners. A host of studies have analyzed the cost of incarceration in many of these prisons. This study takes a broader approach. It compares state per-prisoner department of corrections budgets across 46 states. Accounting for a number of cost factors, significant perprisoner savings were found in states that house a portion of their prisoners privately. All other factors being equal, states such as New Mexico with forty-five percent of their prisoners in private custody spent about $9,660 less per-prisoner in 2001 than non-privatized states. Given New Mexico's prison population of 5,300 this is an annual savings of $51 million. Forty-five percent privatization is expected to reduce the typical department of corrections budget by about one-third.
The paper begins with a short history of the privatization movement and a discussion of the motivation to privatize.
II. Background
Part A. Brief History of Privatization In the late 1970s and early 1980s, public backlash against soft-on-crime policies delivered a generation of tougher judges to the bench. Levitt noted in 1996 that "the incarceration rate in the United States has more than tripled in the last two decades." Federal and state criminal statutes - particularly those dealing with drugs - were also strengthened and law enforcement budgets redoubled. Between 1982 and 1999, the federal government increased its police expenditure by 485 percent ($35 to $40 billion dollars a year go to the War on Drugs alone). Over the same period, states increased their police expenditures by 239 percent. Both trends out-paced inflation and overall growth in government spending by a wide margin. The inevitable result was an explosion in the prison population. Between 1980 and 1999, the U.S. prison population grew fifteen times faster than the population at large. By 1986, "all but seven states were operating their prisons in excess of 95 percent capacity." The overcrowded prisons begat quality lapses. In 1983, "Only about one-fifth of all state and federal prisons were accredited by the Commission on Accreditation for Corrections." More seriously, courts began to intervene, asserting that states' old and crowded facilities violated the Constitution.
When in the early 1980s Tennessee's entire correctional system was found unconstitutional, the state considered contracting with a private firm. The firm, the Corrections Corporation of America (CCA), had been incorporated in 1983 to contract with the Federal Immigration and Naturalization Service to detain illegal immigrants pending hearings or deportation. Tennessee refused CCA's offer. But not long after, prison privatization began in earnest. In 1985, Florida's Bay County contracted with CCA to operate its jail.
The next year, CCA contracted with Santa Fe County, New Mexico to run its jail. By 1987, there were about 3,000 people held in private prisons nationwide. This represented little over one half of one percent of the entire prison population. By 2001, the private prison population had soared to over 91,000 inmates. Despite such rapid growth, only about seven percent of all prisoners were in private custody in 2001.
Part B. Motivations to Privatize
Capacity management and speed of delivery continue to drive privatization. According to one survey, 21 percent of state agencies who privatized say they turned to private firms because of their reputation for speedy delivery. This is because private firms can construct jails and prisons in about half the time it takes government to do so. But in addition to capacity management and speed of delivery, states also turn to private firms in order to improve quality and lower cost.
There are a number of reasons to believe that private prisons offer a better, safer product. One is accreditation. The American Correctional Association is an independent, non-profit professional corrections organization. They accredit public and private prisons. Forty-four percent of all private prisons are currently accredited. Just ten percent of public prisons are accredited. Court orders offer another perspective on quality. "In 2001, of the 50 state correctional departments, entire departments were under a court order to relieve unsatisfactory conditions." Not a single private prison has ever been placed under a court order for unsatisfactory conditions. Finally, there is evidence from a number of independent studies. Sixteen of eighteen studies surveyed by the Reason Public Policy Institute found private prisons to perform as well or better than public prisons.
In addition to capacity and quality, many states choose to privatize in order to save money. State prisons are often encumbered by expensive court rulings and statutory regulations. Quality lapses in New Mexico's state prison system, for example, have led to a costly court order. After a bloody 36 hour riot in Santa Fe's state prison in 1980, a federal judge ordered that the state could not house more than one prisoner per cell. As noted, private prisons have altogether avoided such court trouble. In New Mexico, the private prisons with which the state contracts are still permitted to house prisoners two to a cell.
State constitutions have also restricted prison finance. If states are at their legal bonding capacity, they cannot borrow further to build new prisons. But, if they contract with private prisons, they can get around the constraint. Private prisons are free to spend their contracted money on debt service.
But beyond any legal or statutory constraint, there is a more fundamental reason private prisons are cost effective. The power of economic competition virtually forces them to be so. And there is good empirical evidence that private prisons save money. Segal and Moore identify twenty eight private prison cost studies. They report that "virtually all of them found private prisons to provide significantly lower cost - on average between 5 and 15 percent." These studies suggest a new approach: How does prison privatization affect costs state-to-state? Answering that question is the main task that follows.
III. New Approach: Interstate Comparison of How Privatization Affects Cost
Part A. Previous Studies
Most work on prison privatization has been conducted at the micro level - usually comparing specific prisons to one another. (Many of the studies were commissioned by states whose privatization laws require cost analysis.) The studies, therefore, often look at prisons in one (or occasionally a handful) of state(s). Arizona, for instance, commissioned a study that compared one private prison to the state's fifteen government run facilities. Likewise, Texas conducts a biannual review of the average cost of private and public Texas prisons. Such focused studies may have little widespread applicability. They fail to isolate and quantify an important source of cost savings: the effect of competition induced by privatization. By focusing on the prisons themselves, rather than the broader public and private mix of correctional services, these studies may miss the forest for the trees. Because competition and its threat forces cost discipline, one would expect all prisons - private and public - in a competitive market to be more efficient than prisons in an all-public, monopolistic market. Empirical evidence, in fact, supports this expectation. In Florida, researchers found statewide prison costs were reduced by privatization. Likewise, researchers in Arizona found that after privatization, public prisons lowered their costs, narrowing the gap between public and private prison spending.
Studies often contain the caveat that any public/private cost comparison is an "apples to oranges" exercise. Because private and public prisons operate within such radically different environments, critics say, one ought to be skeptical of the apparent cost savings of private prisons. This has led many to conclude that the only sound comparison is one between prisons similar in all respects except for ownership. This was the tack taken by studies conducted in California, Washington and Tennessee, for example. But as Segal and Moore argue: Achieving identical facilities, populations, and conditions is perhaps impossible. More important, creating identical situations puts the cart of cost comparison before the horse of motivations for privatization in the first place. If a government and private prison are identical in every detail that could affect cost, what is the point of privatizing, since many of the driving factors for privatization (quality, innovation, etc.) are necessarily lost? Many of the dangers of this kind of micro-analysis may be avoided by taking a "bird's eye view" of the issue. Rather than focusing on individual prisons, this study focuses on states. The data are collected from 46 states (the most complete data set obtainable). This allows us to isolate cost savings induced by privatization rather than the specific cost savings of a particular prison. Because data are from a wide cross-section of states, the findings have much broader implications.
Part B. What is the Appropriate Measure of Cost?
An important difference in the way publicly and privately-owned prisons finance capital assets (e.g., building, land, equipment) makes cross-state cost comparisons difficult. Private firms attempt to spread the cost of an asset throughout the asset's useful life. Public managers make no such attempt. Instead, the public accounting practice is to count the capital expenditure only in the year it was made. Since the fee charged by a private firm incorporates capital financing, the ideal public measure of cost would also include some annualized portion of capital investment. Such a measure is not available. We must therefore rely on a measure of spending which excludes public capital expenses (which can vary wildly depending on whether it is a building year), but not the capital expenses of privately-owned prisons. This may bias the data to make private prisons appear more expensive than public prisons. Still, it is the best measure possible.
Part C. Why do Per-Prisoner Costs Vary From State to State?
Our primary interest is in how privatization affects per-prisoner costs. But what factors affect a state's capital-exclusive per-prisoner department of corrections budget? Input costs such as labor conditions vary wildly from state to state. So too does the extent to which a prison can efficiently use its capacity. To isolate the impact of privatization, we must also account for these other factors.
In selecting factors that might affect cost, it is important to mind Segal and Moore's warning that all too often researchers put the cart of cost comparison before the horse of privatization. Assuming that privatization affects prison management policy, it is important not to "control" for any prison policy differences that could be due to privatization. The study, therefore, does not include factors over which prison managers (public or private) have influence. The reader will find no mention of cost-affecting factors such as the type of food or laundry service the state's prisons establish or the salary paid prison employees. Instead, the factors represent a priori conditions that each state's corrections department must deal with.
Wages paid no doubt affect cross-state differences in prison spending (recall that our focus is on non-capital spending, so labor is likely the single most important input). Though prison 26 Only 59 percent of privately managed prisons are also privately owned (and therefore privately financed). See McDonald, p. 35 for further discussion.
Managers may have some discretion over what they choose to pay, they must compete with other employers for comparably skilled employees. Managers are restrained, therefore, by the state's prevailing market wage.
Another important cost factor is organized labor. Where the power of organized labor is strong, labor markets tend to be less competitive. An uncompetitive labor market, in turn, can drive up a number of prison costs. A union of prison guards, for example, might be able to obtain more liberal leave than would be determined absent the union monopoly. Unions may also push up the cost of an employee benefit package. In addition to salary and benefits, unions affect the size of a workforce. They may erect barriers to entry in order to restrict the supply (and therefore raise the price) of labor. Or they may make it difficult for an employer to shrink his workforce when market conditions change. This can lead to over-staffing (featherbedding). Other expenses are also affected by union power. For example, organized labor may raise the price of contracted services such as food, laundry or maintenance. The power of organized labor and the labor conditions created by union-encouraged state laws vary from state to state. Finally, prisons in different states face different constraints on their ability to use prison space. If a prison can place more prisoners in a given space, it can save money on important variable costs such as guards, surveillance equipment and even electricity. But prison managers face different state laws and court rulings that govern the legally permissible crowding level. Furthermore, states differ in their fiscal outlook and ability to match growth in prison populations with new prisons. These differences lead to disparities in prison density. Of course, higher prison density can also save money on construction costs as well. Recall, however, that to the extent possible, this study excludes construction costs by focusing on non-capital expenditures. Because crowded conditions tend to lead to more probationers and parolees, we can help control for differences in what departments of corrections spend on these punishments by controlling for differences in prison density.
In sum, to better understand the effect of privatization on department of corrections budgets, it is important to control for other factors that affect spending. This study identifies the market wage for prison employees, union power within a state and restraints on space utilization as the most important of these other factors. The next section discusses how we included these differences.
Part D. Model and Predictions
To estimate the relationship that exists between per-prisoner spending and cost factors (including privatization), the study relies on an economic model that summarizes how perprisoner costs are explained by the factors discussed in Part C. The factors explaining perprisoner costs are known as "explanatory variables." Data on each of these variables were gathered and the regression technique of ordinary least squares (OLS) was employed to estimate how they explain per-prisoner cost.
For each state's cost measurement, the model uses the statewide, per-prisoner cost of incarceration in 2001. This is based on each state's total department of corrections expenditure (excluding public prison capital costs), and each state's prison population.30 As noted in Section C, any prison - public or private - must offer a salary that can compete with alternative jobs. Though state data on prison guard salary is unavailable, we do have data for state police officer pay, and that should be a good gauge of the market wage for guards.31 We should find a positive relationship between state police salary and per-prisoner cost. 29 To these factors, it might seem appropriate to add the security level of prisoners. We found the factor has little effect on cost, however. Please see the Appendix for details. 30 See Prison and Jail Inmates at Midyear and State Expenditure Report. The author thanks Jim Meade of Lattimore, Black, Morgan and Cain for gathering and codifying the cost data. Department of corrections budgets do include other items such as probation and parole expenses. There is no reason to think these other expenses are in any way related to privatization.
It was also noted in Section C that union power can affect a prison's costs in a number of ways. By controlling for the presence of state "right to work laws," we can help control for the presence of union power. In a state with a right to work law, it is illegal for a union to require that a firm hire only union members. In states with right to work laws, employers not only enjoy greater freedom in staffing, but they presumably operate in an environment where labor monopolies (unions) are not terribly powerful. The impact of a right to work law is captured with a "dummy" variable that takes on a value of "1" if the state is a right to work state and a value of "0" if it is not. This variable should account for the total effect of union strength. We expect to find a negative relationship between right to work laws and per-prisoner cost. In other words, states with freer employment laws and weaker unions are likely to be less burdened by the cost of labor monopoly.
The model also controls for a prison's ability to utilize space efficiently - its density. Our variable for density is "crowding" data obtained from the U.S. Bureau of Justice Statistics. In a recent survey, the Bureau asked each state to rate its prisons' capacities based on three measures: what the prison was built to handle, what the prison was rated to handle and what the prison could handle. Each state's prison population was then taken as a percentage of the three capacity measures to create three different measures of crowding. Each state's lowest and highest measure of crowding was averaged to obtain the most objective measure possible. As noted above, it is reasonable to expect that as crowding (density) increases, prisons will be able to utilize their resources more efficiently. A negative relationship between crowding and per-prisoner spending is a reasonable expectation. That is, as crowding goes up, per-prisoner cost goes down.
Finally, for the variable of interest, percent of state prisoners in private custody, we expect a negative relationship between cost and the extent of privatization.33 That is, as the percent of prisoners privately housed increases, per-prisoner cost should decrease. Furthermore, it is reasonable to expect the relationship to be non-linear. There are two reasons for these expectations. Some savings are likely to be achieved simply because of better private sector management. As the number of prisoners under such management increases, one would expect per-prisoner costs to decline (though perhaps with diminishing marginal savings). Another important source of cost saving from privatization is competition. The credible threat of contract termination should induce both public and private prisons to carefully mind costs. Here, diminishing marginal cost savings are likely. Once the threat of contract termination has been well established, increased privatization seems unlikely to have the same marginal impact as at lower levels of privatization. Such diminishing marginal cost savings can be captured by taking the natural log of the percent of prisoners privately housed.
When all other factors are held constant, a state with five percent of its prisoners in private custody can expect to save an additional $423 per-prisoner per year by increasing privatization by one percent (see Table 2). A state with 45 percent of its prisoners already in private custody, however, will save an additional $55 per-prisoner per year by upping its private prison population by one percent.36 Estimated total savings are quite staggering. A 45 percent privatized state spends about $10,000 less annually per-prisoner than a state with no privatization. In other words, an unprivatized state that chooses to privately house 45 percent of its prisoners can cut its perprisoner budget by a third!
Notes:
Median Per-prisoner Sending is $29,937. Parentheses indicate negative value (savings). The other explanatory variables have no less of an interesting effect on cost. For example, when all other factors are held constant, every extra dollar earned by an entry-level state police officer (as an indicator of the market wage for prison guards) leads to an increase in 92 cents in a state's annual per-prisoner cost. The magnitude of right to work legislation was particularly interesting. All else being equal, the presence within a state of a right to work law reduces annual per-prisoner cost by over $9,000. This is strong evidence of the costly nature of union power.
Finally, the study finds that crowding tends to reduce costs. All else being equal, every one percent increase in a state's ratio of prison population to capacity, lowers annual per-prisoner cost by $147. Prisons reduce per-prisoner non-capital costs by increasing density. This finding must be weighed against quality, which may very well suffer when prisons are over-crowded. On the other hand, an attempt to reduce crowding by prematurely releasing prisoners to parole status may endanger public safety.
IV. Conclusion
Other factors being equal, those states with a strong commitment to privatization in prison services spend significantly less per-prisoner per year than states without privatization. The data indicate, for example, that annual per-prisoner spending in a five percent privatized state was $4,084 less in 2001 than in a state with all-public prisons. A 45 percent privatized state, meanwhile, was estimated to spend $9,660 less per year per-prisoner than a non privatized state. This is an annual savings of nearly one-third of per-prisoner cost.
Rio Grande Foundation
P.O. Box 2015
Tijeras, NM 87059
505 286-2030
www.riograndefoundation.org
Home | About Us | Projects | Articles | Contact Us..................................................................................© Copyright 2012 Corplan Corrections, Inc....