Prison labor effects on the unskilled labor market.

Derrick, Frederick W. ; Scott, Charles E. ; Hutson, Thomas 等

I. Introduction

The prison population has increased nearly five-fold since 1970--over eight percent per year--from 2 in 1,000 adults in 1970 to 9 in 1,000 adults in 1998. This has led to an escalating tax burden and a significant opportunity cost from "warehousing" over 1.9 million inmates. Employing idle inmates "in factories with fences" is getting renewed attention in Congress and state legislatures to offset the expense, to more fully develop and use human capital, and to increase the nation's productivity.

Unfortunately, the impact of prison labor on non-prison labor has been debated over the years based solely on anecdotal evidence. Proponents have touted the social benefits of decreased (net) cost for prisons, improved behavior in prison, the potential for lower recidivism, and rehabilitation through development of work and social skills. Opponents of prison labor from the beginning have pointed out the private costs of lost jobs and lost sales of industries outside the prison.

This paper uses an immigration model and established estimates of elasticities of supply and demand to estimate the labor market implications of prison labor. The prison labor impact estimates provide the first empirical evidence of the impact of prison labor on non-prison labor employment and wage rate. Sections I and II provide discussions of the historical and current contexts, respectively. Sections III and IV include discussions of the methodology and results. Section V includes conclusions and directions for additional research.

II. Historical and Legislative Perspective

Prison labor was designed to "offset the cost of incarceration" in the American penal system. New-gate prison, opened in New York City in 1797, paid nearly all of its expenses during the first five years from production within the prison. The prison was sustained by the prisoners themselves, required no money from taxpayers, and aided in the rehabilitation of the inmates. By 1825, the legislature stipulated that the duty of prison agents was "to cause all the expenses of any kind, to be supported wholly, or as nearly as shall be practicable, by the labor of prisoners" (Morris and Rothman, 1995).

Artisans and businesses bitterly criticized public and private enterprises that used prisoners because of the competition the prison products created. Artisans also feared competition from ex-prisoners who had learned the trade behind bars, actively making it hard for ex-convicts to obtain jobs (Lewis, 1965). Artisans and businesses exerted pressure on legislatures for restrictive laws on prison labor and prison industry. Citizens argued that prisoners were taking the "means of livelihood from local communities" (Walker, 1988). Taxpayers argued such laws would "create a distinction in favor of the mechanic, or a wish to tax the farmer, the merchant, and the professional men, to support the convicts in our State Prisons" (Lewis, 1965). The New York Legislature in 1835 passed legislation that stated inmates would not be taught mechanical trades which competed with domestic labor. The 1842 New York Legislature limited employment of convicts to trades the prisoner had previously learned and practiced prior to conviction. This evolving public attitude toward prison labor was a precursor to the current debate.

The basic issue was how to protect free labor from low cost prison labor. One model to do so, the state-use system, allows the training and use of prisoners for the production of supplies for publicly owned institutions. Free workers viewed this as less competitive than the other forms of prison labor but opposed the selling of excess production in the market. To reduce competition, free laborers argued for diversity in products and low volumes to minimize the impact on any one free producer (Walker, 1988). Other methods argued for reducing competition from prison labor included price controls, limiting the hours or numbers of prisoners, requiring free market wages, and restricting capital improvements (Garvey, 1998).

As unions became politically stronger in the latter half of the nineteenth century, especially during the recessions of the 1870s and 1890s, opposition to prison labor grew. Congress in 1887 outlawed the leasing of federal convicts to farmers and entrepreneurs but did not pass legislation restricting interstate trade in prison production (Morris and Rothman, 1995). At this time, many states depended heavily on the revenue from prisoner leasing. The states refused to house federal prisoners as a result of this legislation forcing Congress to establish the federal prison system in 1891. In spite of these actions, 85 percent of prisoners worked in 1900. However, leasing of state prisoners ended in the years preceding World War I when it was no longer economically viable (Walker, 1988).

The first significant federal legislation limiting state use of prisoners came in 1929 with the Hawes-Cooper Act. It mandated that "prison-made goods transported from one state to another be subject to the laws of the destination state" and allowed individual states to ban the sale of prisoner-made goods within the state (Pencavel, 1967). This act effectively completed the transition of the market for prison goods from the open market in the early nineteenth century to a state-use system. (1)

At the Federal level, Federal Prison Industries Inc. (FPI)--trade name UNICOR--was created in 1934. FPI, which employs federal prison labor to produce goods for sale only to federal agencies, is designed for rehabilitation, not profit. Eight of ten statutory legislative mandates are public policy oriented. The mandates are to: employ the greatest number of prisoners reasonably possible; concentrate on labor intensive manufacturing; diversify industries so that no private industry faces undue competition; diversify products so sales are widely dispersed; minimize competition with private industry and free labor; limit market share for any specific product; sell products to federal and other government institutions; provide opportunity for prisoners to learn skills useful in the free market; sell products at no more than the current market prices; and operate in a self-sustaining economic manner (Schwalb, 1994).

Congress further restricted prison labor in 1935 with the Ashurst-Sumners Act making "shipping prisoner-made goods to a state where state law prohibited the receipt, possession, sale or use of such goods a federal offense." This effectively ended the industrial prison era by banning interstate commerce in prison made goods. The Walsh-Healy Act of 1936 "banned convict labor on federal procurement contracts in the manufacturing, production, or furnishing of any materials, supplies, articles or equipment used in government contracts where the amount thereof exceeds $10,000." (2) In 1940, Congress added further restrictions with the Sumners-Ashurst Act that made it a federal crime to "knowingly transport convict-made goods in interstate commerce for private uses, regardless of the laws in the states" (Orlando, 1967). These actions, in concert with a shift in the structure of corrections from work to rehabilitation through counseling, training, and education, led to further reductions in prison industries.

Increasing prison populations and prison violence led to increasing costs of incarceration and renewed interest in employing prisoners in the 1970's. Congress relaxed its stance on prison labor in 1979 with the Percy Amendment to the Justice Improvement Act which created the Private Sector/Prison Industry Enhancement Certification program (PIE) in seven states. PIE involves private entrepreneurs and the federal government working cooperatively in prison-based joint ventures. PIE programs must pay the prevailing wage; must show that prison labor does not displace private workers or local businesses; and must limit deductions for room and board, family support, and victim compensation to less than 80%. The Amendment loosens some restrictions on "the interstate sale of prison made goods to the federal government" (Walker, 1988). PIE was expanded by Congress in 1984 and 1990 to certificates for 50 localities. (3) During the past twenty years, more than thirty states have enacted laws permitting the use of convict labor by private enterprises. (4)

III. Current Prison Labor Context

Prison labor is small relative to the national economy--1.6 billion dollars for prison industry output in 1997 relative to GDP of over 8 trillion dollars. Thus, total inmate production is less than two hours worth of United States GDR The 76,519 inmates working in prison industry jobs in 1997 amounted to 0.06 percent of the United States labor force (Miller, Shelton, and Petersik, 1998). In addition, over ninety percent of prisoners working have institutional support jobs--laundry, maintenance, cooking, and other tasks--done for the prison where the need for them is created by the prison itself.

Opponents of prison labor focus on the private costs born by the competing companies, manufacturing associations, and labor. They argue that low wage prisoners undermine competitive unskilled and semi-skilled labor markets and decrease living standards of those who remain employed by reducing wage rates. Unskilled or semi-skilled laborers would be the most at risk since most prison jobs are unskilled or semi-skilled.

Industry critics point out that prison industries are exempt from the Fair Labor Standards Act that dictates minimum hourly wages and imposes constraints on employer behavior. Prisoners have no means of filing a grievance or voicing complaints concerning hiring, firing, or reassignment. Flanagan et al (1988) notes prison employment offers workers "sub-minimum wages, no required health benefits, no unions, no vacation time, no absenteeism, and no overtime." In addition, federal mandatory sourcing requirements provide federal prison industries employers a monopoly, further undermining competitive markets. (5)

Supporters of prison labor argue that the danger from low prison wages is overstated for a number of reasons. The value of the prisoner marginal product is low due to a lack of job skills, low socialization skills, high labor/capital ratio, and high turnover rates with forty percent of state prisoners released in less than three years (Ripley, 2002). Costs are higher due to inefficient shipping and production time lost due to security--especially that lost due to lockdowns and the need to carefully control tools (Kling, 1999). (6) Economies of scale are often restricted due to limited market access. In addition, they argue that unskilled and semi-skilled jobs are being transferred out of the country to developing countries. To the extent that prison labor repatriates these jobs, there will be minimal negative employment effect of prison labor. (7)

Prison laborers may develop a work mentality including time accounting, productivity, and economic reward, which improve employment opportunities; increased future earnings; improved behavior in prison, and lower recidivism. If recidivism is reduced by five percent, assuming the average level of crime of $35,000 per year per criminal, the net present value of the savings is approximately $11,000 per inmate released (Kling and Krueger, 1999). Reynolds (1996) projected that the employment of 25% of the prisoners would decrease taxpayer costs by 2.4 billion dollars, approximately 10%. As financially important as the issue is and as heated as the arguments are, it is surprising there are no estimates of the prison labor impact on non-prison labor.

IV. Methodology

Prison labor impact is similar to the impact of increases in labor supply resulting from increased free trade and immigration. Immigrants in the United States, while small in number relative to the host population, are likely to make up a significant portion of the low income and low skilled workers. Prisoner demographics are similar to those of immigrants in that they tend to be younger, less educated, and a larger percentage male than the population.

Immigration models as reviewed by Borjas (1994) and Friedburg & Hunt (1995) commonly treat immigrants as separate inputs into production. One can then derive a system of demand equations or its reduced form for immigrant and host labor and estimate the elasticity of substitution. Without wage data, the comparative static implications for the change in wages and employment can be derived from the reduced form assuming a competitive market. This approach is often implemented by identifying metropolitan areas as labor markets and by comparing labor market outcomes between high and low immigrant cities or counties.

Alternatively, one can estimate the labor market implications of prison labor based on the elasticities of supply and demand. We model the competitive labor market following Freeman (1999) as:

(1) Log S = A + [epsilon]Log W

where S is the quantity of labor supplied and W is the wage rate

(2) Log D = B - [eta]Log W

where D is the quantity of labor demanded.

In this double logarithmic functional form, [epsilon] and [eta] are the absolute values of the wage elasticities of supply and demand respectively. Equating S and D to the equilibrium quantity of labor (Q) gives the market clearing wage and quantity of labor:

(3) Log W = (B - A) / ([epsilon] + [eta])

(4) Log Q = ([epsilon] B + [eta] A) / ([epsilon] + [eta]).

The employment of previously unemployed prison labor resources increases the supply of labor by the number of prisoners employed, increasing output and employment for society. However, the quantity of free labor employed and the free labor wage rate will decrease. (8) If the supply of labor increases by one percentage point due to increased prisoner employment opportunities:

* Free labor employment will decrease by [epsilon] / ([epsilon] + [eta]) percent;

* Prison labor employment would increase by ([eta] - [epsilon]) / ([epsilon] + [eta]) percent of the labor force;

* The net change in the quantity of labor employed is [eta] / ([epsilon] + [eta]) percent and results in an increase in total output for society;

* The free labor wage rate will drop by 1 / ([epsilon] + [eta]) percent.

The greater the elasticity of demand for labor, the larger will be the gains to the country and the smaller the losses to free labor. We use demand elasticities of -.20 and -.31 based on the consistent findings in prior labor market studies. A review of studies by Hamermesh in the Handbook of Labor Economics (1986) concludes that the labor demand elasticity is in the range -.15 to -.50 with the more elastic estimates being found in studies of earlier time periods. The elasticity of demand for low skilled workers is typically estimated to be -0.10 in minimum wage analyses.

The more inelastic the supply of labor, the greater will be the employment and output gains to the country and the smaller the loss of jobs for free labor. Two estimates of the labor supply elasticity are also used, 0.20 and 0.30. The two estimates used are consistent with Pencavel's survey of the literature in the Handbook of Labor Economics (1986) which concludes "that the elasticities of hours worked with respect to wages are very small" and that the behavioral responses "appear to be of a relatively small order of magnitude."

The impact of employing prisoners is estimated using the above elasticities assuming two prison employment levels. The impact of the 2000 prison employment of 85,432 (Correctional Industries Association, 2001) is shown in Tables 1 and 2. The impact of expanding employment to twenty-five percent of the 1,443,175 (US. Department of Justice, 2002) prisoners, the long term goal of prison industry managers is shown in Table 3. The base of comparison for these impacts is the 16,425,356 unskilled civilian labor force(Bureau of Labor Statistics, 1999), the likely affected group. Comparisons against the total civilian labor force would show markedly less impact. Prison wages range considerably across state, industry and type of work, but comparing the total revenue with wages indicates that wages are, on average, very low.

Two alternative marginal rates of technical substitution (MRTS) of prison labor for free labor are used: 1 to 1 (prison labor comparable to civilian labor) and 3 to 1 (prison labor one third as productive). The former provides a worst-case scenario for displaced workers. The latter adjusts for prisoners having less education and work experience, and being transient. An estimate that output per prison employee is roughly one third that of the typical worker is consistent with immigrants in the immigration research and with the ratio of the percentage of aggregate output provided by prisons to the percentage of employment that prisoners constitute.

These assumptions provide three scenarios. The first assumes an MRTS of 1 to 1 and the current level of prison employment. The second assumes an MRTS 3 to 1 and the current level of prison employment. The final scenario assumes an MRTS of 3 to 1 and the long term goal of 25% employment of prisoners. The resulting estimates are short-term estimates of the impact of employing prisoners since price and wage adjustments and capital flows will moderate the impact over time. In addition, expanded employment will lead to greater consumption and associated growth in labor demand.

V. Empirical Results

The current employment of prisoners, MRTS of 1 to 1 results, presented in Table 1, provide estimates for the worst case private costs argued by opponents of prison labor. The net increase in employment per prison job is shown in the top portion of Table 1. For each 100 prison jobs, society nets between 40 and 61 new jobs. The remaining prison jobs result from crowding out of free labor. These assumptions suggest that the 85,432 prisoners currently employed reduce free employment by between 33,503 and 51,254 jobs out of a total 16,425,356 unskilled civilian labor employment as shown in the middle of Table 1. To the extent that demand is elastic and supply is inelastic, allowing the additional employment to create its own demand, there is more limited crowding out. To the extent that labor demand is inelastic, and supply is elastic, there is more crowding out, with non-prison workers dropping out as wages adjust to the increased labor supply.

Unions and other non-prison labor are especially concerned about the wage effect of prison employment. Prison wages are typically below the market wage by a considerable amount increasing the potential for labor substitution. Supply and demand models assume that the employer will employ up to where the wage equals the marginal revenue product of the last worker. (9) As shown in the lower portion of Table 1, the more inelastic the supply and demand are, the larger the wage adjustment will be. Neither quantity demanded nor quantity supplied responds quickly to additional labor, requiring larger wage adjustments to restore equilibrium. Unskilled workers would pay the maximum financial price for the additional prison employment. Depending on the elasticities, wages decrease from .85% to 1.30%. At a wage of $6 per hour this reduction is 5.1 to 7.8 cents per hour or $102 to $156 per 2,000-hour work year. At a wage of $10 per hour, the reduction would be 8.4 to 13.0 cents per hour or $168 to $260 per year.

One way to account for the probable lower productivity of prison labor is to count each prisoner as a fraction of a civilian worker added to the labor force. Doing this by using an MRTS of 3 to 1 leads to 100 prison jobs creating between 80 and 87 new jobs, crowding out only 20 to 13 free jobs, respectively, as shown in Table 2. Current prison employment would cost free unskilled workers between 11,056 and 16,916 jobs--between 0.28% and 0.43%. It also would cost labor from 1.7 to 2.6 cents per hour from a $6 wage or 2.8 to 4.3 from a $10 dollar wage. Annual costs would be between $34 and $51 or $56 and $86, respectively.

Table 3 results assume expansion of prison labor to 25% of prisoners (360,794 prisoners employed) and an MRTS of 3 to 1. This would be nearly a four-fold increase over current prison employment. The net relative employment results are the same as scenario 2: between 80 and 87 new jobs per 100 prison jobs and 20 to 13 free workers, respectively, displaced. With current unskilled civilian employment, this would result in between 46,691 to 71,437 displaced unskilled free workers. Wage reductions would be larger, ranging from decreases of 1.19% to 1.81%. This is 7.1 to 10.9 cents per hour for a $6 wage or 11.9 to 18.1 cents per hour for a $10 wage. On an annual basis, reductions would be from $142 to $218 or from $237 to $362, respectively.

At current prison employment and wage levels, there is an excess supply of prison laborers, providing productivity incentives for the successful applicants and the ability for employers to choose the most productive workers. If prison employment is expanded, this would lead to less motivated and less productive workers being drawn into the process. This would lead to a decline in the (already estimated to be 1/3) productivity, effectively decreasing the crowding out effects of a given increase in prison employment.

VI. Conclusions

These results are consistent with the immigration literature that finds little impact, at most, on wages or unemployment. Lalonde and Topel (1991) note that the large flow of immigrants during the 1970's amounted to only 2.5 million extra persons in the labor force. They found relatively small and quickly dissipating impacts of immigration, even on close substitutes. This labor inflow is relatively small compared to the increase of over 20 million extra persons in the labor force as a result of the baby boom and the increased labor force participation of young women. Thus the increase to 315,418 prison workers would be relatively small compared to shocks the labor market has already received. This suggests that there is sufficient labor mobility, including low wage labor, and/or growth to assure that cross-sectional or time series effects of prison labor will be minimal.

Our results suggest that even a return of prison labor to the free market would have little employment impact on free laborers. Employing every inmate for 2,000 hours (50 weeks of 40 hours per week) per year at a value of production of the minimum wage would account for only 0.2 percent of GDP, which is less than the typical "statistical discrepancies" (Kling and Krueger, 1999). This estimate of the prison employment impact is likely to be inflated as it is unreasonable to assume full employment, no decline in marginal productivity and all new production. The estimates also overestimate the impact of prison labor to the extent that prison industries purchase materials and equipment, providing additional employment and income feedbacks locally. These secondary impacts through business-to-business purchases and tax revenue impacts will mitigate the negative direct job implications on the local community.

The political debate over employment of prisoners has often been heated with accusations that employing prisoners takes away the livelihood of free workers and gives employment to those on whom society has passed judgment. Unions fear the artificially low prison wages and the preferential access to government markets by prison labor employers. Our results suggest that current prison employment has, and that a near four-fold increase in prison employment would have, little impact on the labor market or the overall economy. Paraphrasing Lalonde and Topel (1991)." .. in short, [prison labor is] rather easily absorbed into the American labor market."

TABLE 1

Current Prison Employment and MRTS of 1 to 1

Net Increase in Total Employment per Prisoner
Employed

 Demand Elasticity

 0.31 0.20

Supply 0.30 0.51 0.40
Elasticity 0.20 0.61 0.50

(Job gain per Prison job)

Change in Non-Prison Employment (jobs)

 Demand Elasticity

 0.31 0.20

Supply 0.30 -42,016 -51,259
Elasticity 0.20 -33,503 -42,716

Wage Change

 Demand Elasticity

 0.31 0.20

Supply 0.30 -0.85% -1.04%
Elasticity 0.20 -1.02% -1.30%

TABLE 2

Current employment and MRTS of 3 to 1

Net Increase in Employment per Prison Job

 Demand Elasticity

 0.31 0.20

Supply 0.30 0.84 0.80
Elasticity 0.20 0.87 0.84

(Job gain per Prison job)

Change in Non-Prison Employment (jobs)

 Demand Elasticity

 0.31 0.20

Supply 0.30 -13,865 -16,916
Elasticity 0.20 -11,056 -14,096

Wage Change Assuming Partial Substitution

 Demand Elasticity

 0.31 0.20

Supply 0.30 -0.28% -0.34%
Elasticity 0.20 -0.34% -0.43%

TABLE 3

25% Prisoners employed and MRTS of 3 to 1

Net Increase in Employment per Prison Job

 Demand Elasticity

 0.31 0.20

Supply 0.30 0.84 0.80
Elasticity 0.20 0.87 0.84

(Job gain per Prison job)

Change in Non-Prison Employment (jobs)

 Demand Elasticity

 0.31 0.20

Supply 0.30 -58,555 -71,437
Elasticity 0.20 -46,691 -59,531

Wage Change Assuming Partial Substitution

 Demand Elasticity

 0.31 0.20

Supply 0.30 -1.19% -1.45%
Elasticity 0.20 -1.42% -1.81%

Notes

(1.) The Hawes-Cooper Act was intentionally deleted in 1995 in the rewriting of legislation. Thus, it is no longer applicable and state legislation in 7 states which refer to its enforcement powers are moot.

(2.) Congress's goal was to limit the amount of profit from using prison labor by limiting economic profit available.

(3.) Locality is broadly defined. For example, the state of Texas is one locality; the state of Utah and Utah County have separate certificates; and counties may have certificates even though the state does not as in New Hampshire.

(4.) Garvey (1998) argues that PIE is as restrictive as the Hawes-Cooper and Ashurst-Somners Acts since it replaces one restriction with another. The PIE restriction is that one must pay the same wage for work when productivity in the prison will be lower than on the outside. With lower productivity and comparable wages, Garvey argues that companies have no incentive to invest in PIE. Thus, he argues for a return to the contract system and offers ways of overcoming the challenges of potential abuse, stereotyping, and the principle of less eligibility in a contract system.

(5.) FPI officials argue that mandatory sourcing--having the right to first refusal supplying federal needs--is necessary to overcome the economic disadvantages of low productivity labor and security costs. The National Performance Review headed by Vice President Gore recommended the deletion of the mandatory sourcing requirement.

(6.) A normal eight hour shift in Maryland contains approximately four and a half hours of production. The remaining time is lost to shift changes by guards and supervisors, counts of prisoners as they move between prison sections, tool distribution in the morning and after lunch, and tool collection before lunch and at the end of the day.

(7.) Katz and Krueger (1999) note a major reason for the low unemployment rate in the late 1990's was the 150% increase in the number of men in prison since 1985. The number of adult males in prison equals 2.3 percent of the male labor force. They conclude that incarceration has lowered the unemployment rate by 2.5% and raised wage rates.

(8.) Since the measure of number of workers added is the number of prison jobs, the assumption is that each prison job is occupied by a prisoner. Free market additions to the labor market leave open whether the added workers or the current participants in the labor force are "crowded out."

(9.) This describes what happens in the non-prison labor market. The probability that the prisoner wage will equal their marginal product is quite low as prison industries are often limited by space or other politically-determined restrictions.

References

Borjas, George J. (December 1994). "The Economics of Immigration," Journal of Economic Literature, XXXII, 1667-1717.

Bureau of Labor Statistics. "2000 National Cross-Industry Estimates, http:www.BLS.Gov/ OES_DL.html#1999.

Correctional Industries Association. (2001). Correctional Industries Association Directory, Baltimore.

Flanagan, J. Timothy, Terrence Thornberry, Kathleen Maquire and Edmund McGarrell. (January 29, 1988). "The Effect of Prison Industry Employment on Offender Behavior: Final Report of the Prison Industry Research Project," The Hindelang Criminal Justice Research Center, State University of New York at Albany, Albany, New York.

Friedburg, Rachel M. and Jennifer Hunt. (Spring 1995). "The Impact of Immigrants on Host Country Wages, Employment and Growth," Journal of Economic Perspectives, 9(2), 23-44.

Freeman, Richard. (May 21, 1999). "Making the Most from Prison, Labor," National Symposium on the Economics of Inmate Labor Force Participation, George Washington University.

Garvey, Stephen P. (1998). "Freeing Prisoners Labor," Stanford Law Review.

Hamermesh, Daniel J. (1986). "The Demand for Labor in the Long Run," in Handbook of Labor Economics, Volume I, Edited by O. Ashenfelder and R. Layard, Elsenvier Science Publishers BV, 429-471.

Katz, Lawrence and Alan B. Krueger. (1999). "The High Pressure U.S. Labor Market of the 1990s, Brookings Papers on Economic Activity.

Kling, Jeffrey. (1999). "The Effect of Prison Sentence Length on the Subsequent Employment and Earnings of Federal Criminal Defendants," Woodrow Wilson Economics Discussion paper 208, Princeton University.

Kling, Jeffrey, and Alan Krueger. (May 21, 1999). "Costs, Benefits and Distributional Consequences of Inmate Labor," National Symposium on the Economics of Inmate Labor Force Participation, George Washington University.

LaLonde, Robert J. and Robert H. Topel. (1991). "Labor Market Adjustments to Increased Immigration," in John M. Abowd and Richard B. Freeman, eds. Immigration, Trade, and Labor Markets, University of Chicago Press, 167-99.

Lewis, David. (1965). "From Newgate to Dannemoa: The Rise of the Penitentiary in New York." New York: Cornell University Press.

Miller, Rod, Mary Shelton, and Tom Petersik. (April 1998). "Inmate Labor in America's Correctional Facilities: A Preliminary Report of the American Bar Association's Subcommittee on Correctional Industries," Community Resource Services.

Morris, Norval and David Rothman. (1995). "The Oxford History of the Prison," New York, Oxford University Press.

Orlando, Lewis, (1967), "The Development of American Prisons Customs," Monticlair, NJ: Paterson Smith.

Pencavel, John. (1986). "Labor Supply of Men: A Survey," in Handbook of Labor Economics, Volume I, Edited by O. Ashenfelter and R. Layard, Elsevier Science Publishers BV, p. 3-102.

Reynolds, Morgan. (September 1996). "Factories Behind Bars," NCPA Policy Report No. 206.

Ripley, Amanda. (January 21, 2002). "Outside the Gates," Time.

Schwalb, Steven. (1994). "The State of Correction: Proceedings of the American Correctional Association Conference," American Correctional Association.

US Department of Justice Bulletin. (April, 2003). "Prison and Jail Inmates at Midyear 2002," (NCJ 198877).

Walker, Donald. (1988). "Penology for Profit," College Station Texas: Texas A&M University Press.

Frederick W. Derrick, Charles E. Scott and Thomas Hutson *

Loyola College in Maryland, Baltimore, MD