prison privatization

Social Science Research 42 (2013) 596–610

Contents lists available at SciVerse ScienceDirect

Social Science Research

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / s s r e s e a r c h

Prisons, jobs and privatization: The impact of prisons on employment growth in rural US counties, 1997–2004

0049-089X/$ - see front matter � 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ssresearch.2012.12.008

⇑ Corresponding author. Fax: +1 509 335 6419. E-mail address: ghooks@wsu.edu (G. Hooks).

Shaun Genter a, Gregory Hooks b,⇑, Clayton Mosher b a Tacoma Community College, Building 18, 6501 South 19th Street Tacoma, WA 98466, United States b Department of Sociology, 204 Wilson-Short Hall, Washington State University, Pullman, WA 99164-4020, United States

a r t i c l e i n f o

Article history: Received 30 September 2010 Revised 7 June 2012 Accepted 12 December 2012 Available online 11 January 2013

Keywords: Incarceration Prison expansion Community economic development Rural employment Privatization

a b s t r a c t

In this study of prison privatization we draw on the insights of a recent body of literature that challenges a widespread belief that prisons help to spur employment growth in local communities. We look to these studies to provide an empirically and theoretically grounded approach to addressing our research question: what are the benefits, if any, to employment growth in states that have privatized some of their prisons, compared to states with only public prisons? Our research makes use of a large, national, and compre- hensive dataset. By examining the employment contributions of prisons, as recent research has done, we were able to corroborate the general findings of this research. To study prison privatization we distinguish between states in which privatization has grown rapidly and those states in which privatization has grown slowly (or not at all). Our findings lend sup- port to recent research that finds prisons do not improve job prospects for those commu- nities that host them. We contribute to this literature by demonstrating that new prisons in states in which privatization is surging impede employment growth in the host commu- nity. To explain this we highlight the significant reduction in prison staffing – in both pri- vate and public prisons – where privatization is growing quickly.

� 2012 Elsevier Inc. All rights reserved.

1. Introduction

This article focuses on the intersection of three trends in criminal justice and local economic development. First, in abso- lute and relative terms, the United States has dramatically expanded the number of prisons and prisoners. Fueled by ‘‘get tough on crime’’ policies (Dyer, 2000), incarceration rates soared. In 2010, over 1.6 million individuals were incarcerated, with an incarceration rate of approximately 500 per 100,000 (US Department of Justice, 2012). While this incarceration rate is slightly lower than the peak (506 per 100,000 in 2007), this represents a dramatic increase relative to the 1970s and one of the highest in the world. To accommodate this growth, correctional officials filled prisons beyond their capacity (Besser and Hanson, 2005), converted old buildings to prisons, transferred inmates to correctional institutions in other states, and con- structed new facilities (Lawrence and Travis, 2004). Second, over this same period, state and local governments have priv- atized a number of public services – including corrections Gaes et al., 2004; Hallett, 2006; Logan, 1990; Shichor, 1995). Third, a disproportionate number of new prisons were built in nonmetropolitan counties. Rural counties and communities aggres- sively campaigned to attract prisons. As Besser and Hanson (2005:2) note, ‘‘what had been viewed as a LULU [locally unde- sirable land use] became a last resort for promoting economic development.’’

The sharp recession that began in 2008 – and the enormous pressure placed on state and local budgets – has stalled and may lead to a reversal of the incarceration trend (Coleman, 2008; Lotke, 2008). Even before this recession, the local benefits

S. Genter et al. / Social Science Research 42 (2013) 596–610 597

of prison expansion have been called into question (Besser and Hanson, 2005; Gilmore, 2007; Glasmeier and Farrigan, 2007; Hooks et al., 2004, 2010; King et al., 2004). The pervasive finding in recent research – that prisons bring few if any economic benefits – contradicts the hype surrounding prisons. But this finding is consistent with Eason’s (2010) assertion that prison- building has compounded problems for struggling rural communities. Even if the rate of incarceration declines, the privati- zation of prisons will likely continue (Burnett, 2011; Caputo, 2011; Vardon, 2011). In fact, Corrections Corporation of America (CCA) recently extended an offer to buy prisons from cash-strapped states. This offer is contingent on a commitment to have CCA manage the prisons for 20 years and that the facility is at least 90% full over the period (Kirkham, 2012). Whereas supporters of privatization tout savings for the public sector and attractive employment options for nearby resi- dents, critics dispute these claims and stress the profits flowing to out-of-state corporations. Our research is focused on priv- atized prisons and employment. To the extent that privatization delivers on the promise of lowering costs, it does so by reducing labor costs. As such, it is possible that private prisons generate fewer secure and well-paying jobs than comparable prisons managed by the public sector. In turn, the counties hosting private prisons are less likely to experience a net increase in employment when a prison opens.

To establish a baseline for comparison, we use measures and methods that have been employed to examine the relation- ship between prisons and local economic trends (Hooks et al., 2004, 2010). We then distinguish between states in which the privatization of prisons has grown rapidly and those states in which privatization has grown slowly. Consistent with trends in the literature, our empirical research provides no evidence that prisons contribute to local economic growth. Specific to private prisons, our analyses provide evidence that they have a negative impact on host counties. Where privatization is increasing at a rapid rate, we find a negative and significant relationship between prisons and local job growth.

2. Private prisons and local employment trends

‘‘Privatization’’ refers to the transfer of state managed and owned bureaucracies – and the public function they serve – to private management and ownership (see Sclar, 2000). In important respects, critics and proponents of privatization agree about the mechanisms through which privatization reduces costs. Public bureaucracies are labor intensive, the savings (if any) will come from reducing outlays for labor. The executives of firms bidding on privatization contracts receive lucrative compensation (Mattera et al., 2003). As such, cost saving must come from other levels of the organization: fewer employees, fewer full-time employees and lower salaries.

‘‘Market fundamentalism’’ refers to a strong adherence to an idealized free market and a belief that markets should be sheltered from political intervention and operate ‘‘freely’’ (see Somers and Block, 2005). For fundamentalists, unfettered markets offer the most efficient, highest quality, and most cost effective means of responding to human needs and desires. Unleashing the pressures of supply and demand ‘‘enforces efficiency by virtue of [the ability of buyers] to cease transacting with one provider in favor of another’’ (Hallett, 2006: 134). For proponents of privatization, the pressure of competition and avoidance of suffocating civil service rules insures that privatization will lower costs while maintaining or improving service delivery. Because private corporations can hire and fire personnel with fewer constraints, privatization allows for more nim- ble and responsive management than is possible with an ossified public bureaucracy.

Skeptics doubt that private prisons will display creativity and fresh thinking. Managers of private prisons are constrained by corporate executives and by a board of directors focused on profitability and growth. Moreover, all prisons must conform to the complex legal and political framework that constrains prison management (Ogle, 1999, p. 588). These cross-cutting pressures will likely result in private prisons mimicking publicly managed prisons. In a study sponsored by the Department of Justice, Austin and Coventry (2001) found little evidence of organizational innovation. With regard to budgetary savings, private prisons achieved ‘‘modest cost savings, at least initially, by making modest reductions in staffing patterns, fringe ben- efits, and other labor-related costs. . .. The promises of 20-percent savings in operational costs have simply not materialized’’ (Austin and Coventry, 2001, p. 58).

Critics are not surprised to find few cost savings. But problems go further. This is not the first time that prisons have been privately run in the United States – and the record is disturbing (Durham, 1993; see also ACLU of Ohio, 2011; Walker, 1994). ‘‘History shows that privately operated prison facilities were plagued by problems associated with the quest for higher earn- ings. The profit motive produced such abominable conditions and exploitation of the inmates that public agencies were forced to assume responsibility’’ (Austin and Coventry, 2001, p. 17). With the emphasis on reducing labor costs, there is no reason to expect that contemporary private prisons will avoid such abuses. For the present purposes, this larger debate over prison privatization provides a context for considering local employment impacts. If the defenders are correct when they paint a positive view of working conditions, then host counties should be better off for housing a private prison (or at a minimum, no worse off). If, however, the critics are correct when projecting fewer jobs – and if these jobs pay less and are less secure than at public prisons – then host counties may be significantly worse off for housing a private prison.

2.1. Prisons and employment growth

As incarceration increased in the 1980s and 1990s, economic developers and policymakers espoused the virtues of this ‘‘industry’’ and promoted it to solve local economic problems. Communities, especially rural communities, competed fiercely to ‘‘win’’ prisons by offering large, public subsidies (Beale, 1997; Glasmeier and Farrigan, 2007; King et al., 2004). Academics,

598 S. Genter et al. / Social Science Research 42 (2013) 596–610

policy makers, and journalists of the 1980s and 1990s generally supported the economic claims of proponents, entrenching a conventional wisdom that a prison provided direct benefits in terms of employment opportunities at the prison. In addition, proponents claim a number of indirect benefits, including increased demand for local goods and services, increased tax rev- enues for local governments, upgrades of the local infrastructure, and population increase (creating opportunities for in- creased funding from a variety of federal programs) (Glasmeier and Farrigan, 2007, p. 277). This conventional wisdom was reinforced by commissioned impact studies designed to generate findings favorable to prison expansion (McShane et al., 1992) and by a host of academic studies that concluded prisons generated jobs for local communities (Abrams and Lyons, 1987; Carlson, 1992, 1991; Sechrest, 1992, 1991; Shichor, 1992). Indeed, the view that prisons boost local economies was so widespread that opponents conceded the point, even as they condemned economic development plans rooted in hu- man suffering (see Hooks et al., 2004).

Recent studies have challenged these optimistic claims (see Hooks et al., 2004; McShane et al., 1992). Most important, prisons do not generate local economic benefits (King et al., 2004), and in some cases can be harmful (Besser and Hanson, 2005; Gilmore, 2007; Hooks et al., 2004, 2010). Furthermore, simply ‘‘identifying that a prison creates employment without tracing the recipients of that employment provides little useful information for citizens of a potential host community.. . . When prisons are promoted to a community as a solution for economic distress, the belief by the community is that not only will the prison bring jobs, but also that the current residents will fill the jobs’’ (King et al., 2004, p. 457). But this is often not the case. Thies (1998) found that workers from outside of the county typically occupy new prison jobs, effectively narrowing opportunities for local residents (see also Beale, 1997; Gilmore, 2007). Partly, this can be explained by seniority privileges, especially where corrections personnel are unionized (King et al., 2004; see also Dao, 1997; Thies, 1998). Nor are these dynamics limited to high-wage jobs; prisons generate surprisingly few low-wage jobs. Janitorial positions are out of reach because they are typically filled by the prisoners themselves. In fact, a number of prisons provide janitorial, landscaping and similar services to government, church and other community groups – paying well below the minimum wage. This can result in direct competition between local residents and inmates for jobs (King et al., 2004:474; see also Blankenship and Yanarella, 2004).

Recent literature casts doubt on the local benefits of large public infrastructure outlays more generally, e.g., sports venues and casinos. Such projects are often highly visible and politically popular, but empirical evidence casts doubt on claims of local benefits (Wolman and Spitzley, 1996). When local and county governments invest in prison construction they may be left with fewer resources to invest in other services that are more effective (Hooks et al., 2004). Furthermore, prisons gen- erate few linkages with the host community and thus leak important local dollars that are essential to job creation. During the construction phase, local firms and workers may be squeezed out. Prisons are large facilities. Especially in rural counties, local construction firms and workers typically lack the expertise to compete on such projects (see Blankenship and Yanarella, 2004; Hooks et al., 2004; King et al., 2004). The exclusion of local workers may continue after the prison opens. Gilmore (2007) notes that relatively few corrections officers live in the host community, thereby diluting their contributions to the local economy (see also King et al., 2004).

2.2. Privatization and the management of prisons

The case for privatizing prisons rests on claims of efficiency, innovation, and flexibility. Corrections Corporation of Amer- ica, the largest private prison company in the United States, housing some 75,000 inmates in more than 60 facilities, trum- pets these purported benefits on its website: ‘‘CCA benefits America by protecting public safety, employing the best people in solid careers, rehabilitating inmates, giving back to communities, and bringing innovative security to government correc- tions – all while consistently saving hardworking taxpayers’ dollars.. . . We extend life-changing career opportunities, where you can make a difference’’ (Corrections Corporation of America, 2010). Evaluating whether or not privatization has deliv- ered on these promises has sparked scholarly debate on many aspects of private prisons, including ethics, race, politics, and inequality (ACLU of Ohio, 2011; Shichor, 1998, Shichor, 1995; Logan, 1990; Hallett, 2006; Price, 2006; Price and Riccucci, 2005). In reviewing the literature, we gleaned important lessons on two issues directly related to our research: (1) the effects of prison privatization on labor, and (2) the generation of an economic climate of competition and the extent to which this may trigger a ‘‘race to the bottom.’’ Specific to prisons, proponents of privatizing them believe that competition leads private prisons to be more efficient than their public counterparts; they can, therefore, provide an equal or better ‘‘service’’ at a lower cost. Detractors emphasize the tension between profit chasing and cost containment. Because private companies are behol- den to their shareholders, they privilege profit over the well-being of employees, prisoners or taxpayers.

Approximately 70% of prison budgets are allocated to labor (Donahue, 1989; Gaes et al., 2004). According to Gaes et al. (2004), the purchase of materials required for construction and operations offers few opportunities for dramatic reductions in spending. Based on these facts, most observers – including proponents of privatization (see for example Logan, 1990) – agree that the major cost savings of private prisons must come from labor expenses. A recent comparison of public and pri- vate prisons (Austin and Coventry, 2001, p. iii) has shown that the total cost savings of private prisons ‘‘is only about 1%, most of which comes from cutting labor costs.’’ Since the focus on trimming labor costs is taken as a given, how do proponents and detractors envision the effect of free market discipline on the operations of private prisons? What are the consequences for job growth in host counties?

We found several studies asserting that private prisons have generated significantly more local jobs than public prisons – if only for the wrong reasons. When comparing public and private prisons, higher turnover rates and lower morale will likely

S. Genter et al. / Social Science Research 42 (2013) 596–610 599

be found at private prisons. In turn, because it leads to more job openings being advertised, high turnover may give the appearance of a net growth in jobs available to local residents (Shichor, 1995; see also Camp and Gaes, 2002; Hallett, 2006).

According to proponents, private corporations are deft at personnel management and ‘‘can cut costs without cutting sal- aries. . .[through] [a]dequate and appropriate staffing, better working conditions, and more efficient procedures.’’ In turn, this superior personnel management can ‘‘improve productivity and morale, decrease absenteeism and turnover, and reduce expensive reliance on overtime’’ (Logan, 1990, p. 81, emphasis added). Some authors have argued that the introduction of private prisons can discipline public prisons into becoming more efficient and flexible (McFarland et al., 2002, p. 6). That is, the innovative approach to personnel management adopted by private prisons will be adopted by the remaining public prisons, thereby providing indirect benefits.1 Assuming for the moment that Logan is correct about the attractive salaries and working conditions at private prisons, the only other avenue for controlling labor costs is through a net reduction in staff, leav- ing residents of host communities with fewer opportunities to secure employment. Critics of privatization dispute the assertion that private prisons pay higher salaries to employees, but they are in agreement that private prisons will have a relatively smal- ler workforce (see for example, ACLU of Ohio, 2011). The issue here as it relates to impacts on the host county should be obvi- ous: job opportunities for locals, which are limited even for public prisons, are in even shorter supply in private prisons.

2.3. Patterns of prison location and management, public and private

The US Department of Justice (various) has conducted a census of adult prisons at various points in time. We compiled data from these censuses to track the trend towards privatization and the consequences for staffing patterns. Table 1 sorts states based on the rate of privatization from 1995 to 2000.

Although rapid privatization (above the 75th percentile) is not specific to one region, no southeastern state and no New England state experienced rapid privatization. Nor is it the case that only small and less populated states rapidly turned to privatization (or vice versa).

In recent decades, new prisons were disproportionately sited in rural areas (Beale, 1997; Hooks et al., 2010). Eason (2010) has documented the minority population was significantly higher in prison towns when compared to rural communities without prisons. Fig. 1 returns to this issue, comparing patterns in states maintaining publicly managed prisons with pat- terns in states shifting towards privatization (see Table 1).

Fig. 1 contrasts counties with and without a new prison (between 1990 and 1997), and it provides separate contrasts based on rates of prison privatization (see Table 1). On several sociodemographic measures, differences are modest. In states undergoing a slow shift towards privatization, the percentage of adults with a high school degree is nearly identical in coun- ties with and without a new prison. In states rapidly shifting to privatized prisons, the prevalence of high school degrees is slightly lower than in states privatizing more slowly, but there is little difference between those counties with and without a new prison. Where privatization was occurring at a slow rate, approximately 12–13% of the population possessed the BA degree. Where privatization is occurring at a rapid pace, new prisons were located in counties with a comparable rate of BA degrees. However, in counties with no new prison, the percentage of adults with a BA degree was significantly higher (15.8%).

When the focus shifts to poverty, small differences are in evidence. Counties that hosted a new prison have a poverty rate approximately 2 percentage points higher in both cases: 14.2% versus 16.1% where privatization is occurring slowly and 15.3% versus 17.5% where privatization is occurring rapidly. Consistent with Eason’s research (2010), the most striking con- trast involves the minority population in a county. Where privatization is occurring rapidly, new prisons were located in counties with a significantly higher nonwhite population (17.0%) than counties with no new prison (12.6%). Even more strik- ing is the pattern where privatization is occurring slowly (or not at all). Eleven percent (11%) of residents were nonwhite in counties with no new prison. However, where new prisons were built between 1990 and 1997, over 23% of the residents were nonwhite. Given these differences in the counties hosting prisons, our analyses of employment trends will revisit these issues (see below).

The focus now shifts to employment at prison, contrasting private and public prisons on this score. As of 2005 – and excluding employment at federal facilities – nearly 150,000 people were employed at maximum security facilities (see Fig. 2). One hundred twenty five thousand (125,000) people were employed at medium security facilities; while minimum security facilities accounted for another 67,000 jobs. Private prisons account for a small fraction of these jobs: 1% at maxi- mum security prisons and 6% at medium security prisons. While private prisons accounted for 21% of all employment at minimum security prisons, the total number of jobs is significantly smaller at these facilities. In total (including all security levels), prisons employed approximately 233,000 people in 2005; private prisons accounted for 21,700 or 6.5% of these jobs.

Given that the vast majority of prisons and most jobs at prisons remain in the public sector, we focus our attention on the possibility that the shift toward privatization is changing staffing patterns at public prisons. As noted (see above),

1 To control costs, Phoenix opened up bidding for garbage collection. Ultimately, the city-run ‘‘Public Works’’ won the contract with a margin of $6 million over the nearest private competitor, demonstrating the ability of public agencies to compete with private firms on price. For proponents, Phoenix provides an example of competition keeping costs down while pushing public bureaucracies to provide services with greater flexibility and efficiency (Osborne and Gaebler 1992). Given our focus on employment prospects, it is notable that these savings were made possible by using technology requiring only one worker per truck (instead of three), resulting in reduced service and fewer jobs (Gaes 2005, p. 86).

Table 1 Prison privatization trends 1995–2000 (by state). Sources: US Department of Justice, 1998, 2004

Rapid growth in privatization (above 18% growth between 1995 and 2000, 75th percentile)

Slower growth in privatization (below 18% growth between 1995 and 2000, 75th percentile)

Public prisons, only

Arizona Alabama Delaware California Arkansas Nebraska Colorado Connecticut New Hampshire New Jersey Florida Rhode island Oklahoma Georgia Vermont Pennsylvania Idaho Washington South Dakota Illinois Texas Indiana Utah Iowa Wyoming Kansas

Kentucky Louisiana Maine Maryland Massachusetts Michigan Minnesota Mississippi Missouri Montana Nevada New Mexico New York North Carolina North Dakota Ohio Oregon South Carolina Tennessee Virginia West Virginia Wisconsin

600 S. Genter et al. / Social Science Research 42 (2013) 596–610

proponents of privatization argue that allowing private corporations to perform public functions will serve to discipline pub- lic agencies and induce streamlined staffing.

Fig. 3 charts employment trends by management type and the growth in privatization between 1995 and 2000. Because we are focused on employment trends, this chart tabulates the number of jobs per 100 inmates. It comes as no surprise that privatized maximum and minimum security prisons generate significantly fewer jobs than publicly-managed facilities. With regards to minimum security facilities, the difference between public and private facilities is modest.2 For the present pur- poses, contrasts among public prisons are especially important.

Regardless of security level, states without private prisons have the highest levels of staffing, generating 35.67 positions per 100 prisoners at maximum security facilities, 37.52 at medium-security prisons and 30.24 at minimum security prisons. In states where privatization grew slowly (below the 75th percentile) between 1995 and 2000, there is no clear pattern of decline in staffing trends. At maximum and minimum security facilities, staffing levels are nearly identical to those found in states with only public facilities. A significant difference is found at medium security facility: states with only public prisons generated roughly 25% more jobs per 100 inmates (37.52 versus 30.03) than in states in which privatization grew slowly. When the focus turns to states undergoing rapid privatization (above the 75th percentile – or an 18% increase between 1995 and 2000), contrasts are clear. When compared to states with only public prisons, public maximum security prisons in states undergoing rapid privatization generated 22 percent (22%) fewer jobs per 100 inmates (35.67 versus 29.17). Com- parable contrasts for medium and minimum security facilities are 37 percent (37%, 37.52 versus 27.40) and 26 percent (26%, 30.24 versus 23.92), respectively.

Given that the vast majority of prison employment is found in maximum and medium security facilities that are publicly- managed (see Fig. 2), the sharp contrast among public prisons has implications for local employment impacts. We have undertaken empirical research to shed light on debates over privatization and the impact of prisons on host counties. Given

2 Many of the privately managed minimum security facilities are work release facilities (Feeley, 2002). The relatively high staffing levels may reflect an aggressive deployment of prison labor to perform work in the community and generate income for the corporation managing the facility. Future research is warranted to examine these issues.

Fig. 1. Selected characteristics of counties by rate of prison privatization (see Table 1) and prison expansion in county (1990–1997). Sources: Data on prison privatization and prison growth from US Department of Justice, 1998, 2004. Data on county characteristics from US Department of Commerce, 1995

Fig. 2. Full-time employment at public and private prisons in the United States, non-federal prisons in the United States in 2005. Source: US Department of Justice, 2005

S. Genter et al. / Social Science Research 42 (2013) 596–610 601

the trends revealed in Figs. 2 and 3, we anticipate finding that employment trends are worse in the context of rapid privati- zation. As noted, recent literature has examined the effects of prisons on local areas and found that prisons, on the whole, have a modest impact on local employment. To the extent that prisons impact employment in host counties, that impact is negative. Should we find that privatization contributes to a significant and negative impact on local employment, this re- search can further our understanding of why the prison building boom provided so few benefits to host counties.

Fig. 3. Full-time jobs per 100 inmates in 2005 (fulltime jobs per 100 inmates in 2005 was computed for each state, management type and security level. This measure was calculated by dividing the number of inmates (100s) by the number of full-time employees), by Management Type and Growth of Prison Privatization between 1995 and 2000 (computed for each state by subtracting the percent of state prisons operated by private firms in 1995 from the comparable measure in 2000). The 75th percentile (or 18% growth between 1995 and 2000) demarcates slower and rapid growth. In Fig. 3, staffing trends in states with only public prisons are contrasted to other states. In models of interest (see Table 3–5), counties in these states are included with those where growth was below the 75th percentile) (non-Federal Prisons in the United States). Sources: US Department of Justice, 1998, 2004, 2005

602 S. Genter et al. / Social Science Research 42 (2013) 596–610

3. Data and methods

To examine the employment impacts of prisons, we made use of a dataset that includes detailed county-level employ- ment information and other economic measures, as well as information on prison construction. According to Hooks et al. (2004, 2010), counties (1) are preferred over larger units of analysis such as states or regions because these larger units often obscure within unit variation, and (2) provide stable boundaries over time, which facilitates comparative analysis. With a county-level focus, however, it is possible to lose sight of macro-level political and economic interventions or processes that in part determine county level conditions. For this reason, and following Hooks et al. (2004, 2010), we include controls for regional economic processes and spatial autocorrelation (see below).

We employ a panel design – a 7-year period for the dependent variable (employment growth except construction, 1997– 2004), as well as the lagged measure of this same variable (1990–1997). For this research, established prisons refer to those founded prior to 1991; new prisons opened between 1991 and 1997. We identified states reporting a rapid shift towards privatization between 1995 and 2000, at or above the 75th percentile, i.e., eighteen percent (18%) or faster growth in priv- atization (see below for details on data sources and measurement). When employing a panel design to study change, Halaby (2004; see also Finkel, 1995) makes a persuasive case for employing the method of first difference, i.e., the dependent and independent measures are change scores. This approach models directly the change under investigation and provides some protection in the case of omitted variables (see Finkel, 1995: 5). For this reason, we created change scores for the indepen- dent and control variables. To reduce heteroscedasticity, we employ logarithmic transformations for each dependent and independent measure. Change scores include values of 0 and negative numbers. Because the log is undefined for these val- ues, we calculated the log for the absolute value, plus one (1). Where the observed change score was negative, this result was multiplied by negative one (�1).

3.1. Dependent variable

Because jobs are generally viewed as the most important benefit in contests to attract new prisons, we focus on employ- ment growth. The Bureau of Economic Analysis (BEA) (2008) compiles detailed information from federal sources as well as

S. Genter et al. / Social Science Research 42 (2013) 596–610 603

from state reports (i.e., Social Security contributions, unemployment compensation, etc.). Using the BEA data, we developed a measure of employment growth (1997–2004). To control for employment stimulation due to construction, we subtract con- struction employment from total employment. Because we have employment data for each year, we measure the average annual change in employment over the period. That is, we measured the change in employment for each year and calculated the average across this 7-year period. This average annual change score is superior to a simple change score using the begin- ning and ending year of the period because information from each year is included and thus potential influence of any one exceptional year is moderated by the trends of all the years in the period (log transformation).

3.2. Prisons

We have compiled data on adult (state level, public and private) correctional facilities built in the United States prior to 2000 (US Department of Justice, 2000). We exclude federal prisons from these analyses. By focusing solely on state prisons, our analyses highlight the impacts of state-level decisions to adopt or eschew privatization. Prior research contrasts the eco- nomic effects of newly constructed prisons and established prisons, thereby distinguishing between immediate benefits (or drawbacks) and long-term benefits (or drawbacks) of prison construction. We consider a prison to be new prison if it was constructed between 1991 and 1997. Established prisons are those built prior to 1991 (see Hooks et al., 2004, 2010).

3.3. Control variables

Agglomeration effects are among the most persistent of findings in regional analyses: larger and more populated regions tend to be centers of subsequent growth, and in a similar vein, regions with greater human capital and prior economic activ- ity tend to grow more rapidly. Our analyses include several lagged measures of agglomeration, including employment growth in the preceding 7-year period (1991–1997). We include a separate measure of change in construction employment (1991– 1997) in order to control for the local benefits that potentially flow from large construction projects (including, but not lim- ited to, prison construction) (Bartik 1991, 1994). In the North American Industrial Classification (NAIC) system, the two-digit classification of construction (23) encompasses commercial, industrial, and residential projects. For the years prior to 1997, we used equivalent codes from the Standard Industrial Classification (SIC) system to construct this variable (for details on NAIC and SIC and conversion between them, see US Census Bureau, 2008).

The importance of infrastructure to economic development is well-documented, with air transportation increasingly sig- nificant over the period of interest. Counties with an airport have grown more quickly than those lacking one (Irwin and Ka- sarda, 1991). The change in commercial aircraft activity (1980–1990) compiled from US Federal Aviation Administration reports, (1981, 1990) provides a measure of infrastructural development. Commercial banking assets are also an important infrastructural resource that facilitates local growth. As such, change in bank deposits (1980–1990) in a county is included in our models, with the expectation that this variable is positively associated with economic growth.

Research into local taxation has become increasingly nuanced in recent years. Instead of assuming that employers will vote with their feet to escape relatively high local taxes, recent research (Brueckner, 2000; Wilson, 2000) also considers the possibility that residents and employers will be attracted to public goods made possible by higher levels of taxation. To control for impacts of local taxation rates, change in per capita property taxes (1977–1987), is included. To control for the potentially positive role of enhanced local fiscal capacity, we included change in revenues available to local governments (1977–1987). Although rates of taxation and levels of general revenue are closely intertwined, we measured taxation on a per capita basis and absolute levels of general revenue. We found only modest levels of correlation and no evidence of collinearity.

Our models include measures of change in percent of population with a high school degree (1980–1990) and change in per- cent of population with a college (BA) degree (1980–1990). We anticipate both measures are positively associated with employment growth, but the change in college graduation rates have a greater impact (Young et al., 2004).

To control for employment structure, we draw on Lobao et al.’s (1999) four measures of change in a sector’s concentration in a county between 1980 and 1990: core manufacturing, core non-manufacturing, competitive, and state sector. The core sector (durable manufacturing and non-manufacturing industries such as producer services) is characterized by high wages and benefits and greater job security; the competitive sector (such as retail services) is characterized by low wages, few benefits, and lower job security. The state sector includes employment in public administration, health, education, and welfare. We use the two-digit SIC codes to develop the four measures, each of which refers to change in the percentage of county employ- ment in the sector. Although core manufacturing provided stable and well-paying employment for much of the postwar per- iod, the period we are investigating is one of deindustrialization. For this reason, we expect that change in core manufacturing will be inversely related to employment growth. Because service industries grew rapidly over the period, change in the remaining sectors, core non-manufacturing, competitive sector and state sector will likely have a positive im- pact on employment growth.

County economic growth may be influenced by regional economic processes. We control for regional context by inserting a dummy variable for each Census division, omitting the North Central division.

604 S. Genter et al. / Social Science Research 42 (2013) 596–610

3.4. Spatial autocorrelation

Growth or decline in one county likely influences economic growth in nearby counties (e.g., commuting to work across county boundaries or the expansion of service industries to support economic activities in nearby counties). We address spa- tial autocorrelation with a procedure developed by Land and Deane (1992). To calculate the spatial effects term ‘‘each place is treated successively as the point of reference, and the sum of quotients of the [dependent measure] of every other place divided by its distance from the reference point is computed’’ (Land and Deane, 1992, p. 227). Distances are computed using a standard trigonometric function and latitude and longitude coordinates internal to each county. As the spatial effects term is endogenous, it is regressed on all dependent variables in the model plus a set of instrumental variables. In turn, the instru- mented form of the spatial effects term is used in the model of interest. We use the unemployment rate in 1970 and existing housing units in 1950 as instrumental variables.

4. Results

Table 2 reports summary statistics (means and standard deviations, logged and untransformed values) for the variables included reported analyses. Because the rate of prison privatization (by state) is a central feature of our analyses, we have included summary statistics for all (2242) non-metropolitan counties,3 for counties in states with rapid growth in private pris- ons (at or above the 75th percentile, 487 counties), and for counties in states with slower growth in privatization (1755 coun- ties). A quick glance at the dependent variable reveals that there is no difference in employment growth between states undergoing rapid versus slower privatization. In like fashion, employment growth over the 1990–1997 period was also very similar in the two sets of states.

For counties in states undergoing rapid privatization, a mean value of 0.08 prisons per county is found for established prisons (built in 1990 or earlier) and the mean value for new prisons (built between 1991 and 1997) was 0.08 per county. Conversely, for counties in states with little or no privatization, the mean value of established prisons was 0.11 prisons in 1990 while the mean value for new prisons was 0.04 prisons per county. These mean values suggest that most rural counties do not house a prison – but they also demonstrate that, on average, states undergoing rapid privatization had fewer prisons in 1990 but were adding prisons at a faster pace from 1991 to 1997. Freemont County (Colorado) for example – a state expe- riencing rapid privatization (see Table 1) – had eight prisons as of 1990 and added four more prisons between 1991 and 1997.

Table 3 presents the findings of three regression analyses that examine the determinants of total employment growth in rural counties. The full sample includes all rural counties in the contiguous 48 states. The second model, includes 487 rural counties in states that rapidly shifted towards private prisons between 1995 and 2000. Conversely, the third model is re- stricted to the 1755 counties in states that avoided privatization or shifted more gradually towards private management of prisons.

For the most part, control variables performed as expected across the three models. The Land-Deane spatial effects term fails to attain statistical significance in any of the models, suggesting that the analyses are not compromised by an unmea- sured spatial process. Coefficients for lagged measures of change in employment are positive and significant. It is also the case that the change in construction employment (lagged) is positive and significant in each model. While not specific to prisons, this suggests that large construction projects have a positive impact on county employment trends. Contrary to expectations, neither per capita taxes nor commercial bank deposits achieved statistical significance; and we found an in- verse relationship between growth in commercial aircraft activity and employment. We also found positive contributions to employment growth in general revenues of local government. Enhanced educational attainment contributed to employ- ment growth, but this positive role is only in evidence for growth in BA degrees (not high school graduation rates). With regard to industrial segmentation, core manufacturing returned negative and statistically significant results in each model. Core non-manufacturing and competitive sector, on the other hand, failed to attain statistical significance. However, growth in the state sector’s share of the overall economy stimulated job growth.

Overall, Table 3 corroborates prior research into the impacts of prisons on local employment trends. These analyses offer no evidence that prisons contribute to employment growth. In states undergoing rapid privatization, there is evidence that prisons impede local employment growth. For counties in these states, established prisons do not contribute to growth, and new prisons are inversely and significantly related to employment growth. Overall, these findings suggest that privatization is likely to result in negative effects on employment growth in host communities.

Table 3 points to an interaction between privatization trends in a state and the local impact of prisons. To examine this interaction in greater detail, we created a dummy variable based on privatization trends from 1995 to 2000: this variable equals zero (0) for states reporting little or moderate shift towards privatization (below 75th percentile), it is coded one (1) for counties in states that rapidly embraced private management of prisons (at or above 75th percentile). Because this is a dummy variable, we created four slope-dummy interaction terms (see Hamilton, 1992; Jorgenson, 2006): (1) established prisons in states with rapid growth in privatized prisons; (2) established prisons in states undergoing a slower growth in priv-

3 As our dataset includes all nonmetropolitan counties, a case could be made that inferential statistics are not needed. However, the various measures we employ are collected on 1 day during a given reporting year. In this sense, we do not have data on each day and county; and our data may reflect sampling error. For this reason, we have opted to include tests of significance.

Table 2 Descriptive statistics. (mean and standard deviation)a

All non- metropolitan countiesb

Rapid growth in privatization (at or above 75th percentile)

Slower growth in privatization (below 75th percentile)

Number of states 48 10 32 Number of non-metropolitan counties 2242 487 1755

Dependent variable Employment (1997–2004) except

construction, count 9.28 (0.02) 9.28 (0.02) 9.28 (0.02)

Independent variables New Prison (1991–1997), count 0.05 (0.18) 0.08 (0.24) 0.04 (0.16) Established Prison (pre-1991), count 0.11 (0.29) 0.08 (0.27) 0.11 (0.29)

Control variables Employment (1990–1997) except

construction, count 10.27 (0.02) 10.27 (0.02) 10.27 (0.02)

Construction employment (1990–1997), count

8.34 (0.01) 8.34 (0.01) 8.34 (0.01)

Percentage of labor force w/BA (1980–1990) 0.10 (0.04) 0.10 (0.06) 0.10 (0.04) Percentage of labor force w/HS diploma

(1980–1990) 0.10 (0.03) 0.09 (0.04) 0.10 (0.03)

Core manufacturing (1980–1990), percentage

0.02 (0.02) 0.02 (0.02) 0.02 (0.02)

Core non-manufacturing (1980–1990), percentage

0.02 (0.02) 0.02 (0.02) 0.02 (0.02)

Competitive sector (1980–1990), percentage 0.04 (0.03) 0.05 (0.04) 0.04 (0.03) State sector (1980–1990), percentage 0.02 (0.02) 0.03 (0.03) 0.02 (0.02) Commercial aircraft activity (1980–1990),

flights per year �0.01 (0.03) �0.00 (0.04) �0.01 (0.03)

Commercial bank deposits (1980–1990), $1982 millions

2.05 (2.17) 1.54 (2.32) 2.20 (2.10)

Per capita property tax (1977–1987), $1982 millions

0.19 (0.16) 0.26 (0.22) 0.17 (0.14)

General revenues of local government (1977–1987), $1982 millions

1.56 (1.05) 1.72 (1.06) 1.52 (1.04)

a All values are natural logarithms of observed change scores. As change scores include zero and negative values, the natural logarithm is based on the absolute value of the observed value plus one (1). When the observed score was negative, the result was multiplied by negative one (�1).

b Excludes district of Columbia, Alaska and Hawaii and individual counties with missing data for one or more variables included in models of interest (see Tables 3 and 4).

S. Genter et al. / Social Science Research 42 (2013) 596–610 605

atized prisons; (3) new prisons in states with rapid growth in privatized prisons; (4) new prisons in states experiencing slower growth in privatized prisons.

When reviewing Table 4, a review of control variables and the variables of interest reinforce the trends discussed above. The Land-Deane term is not significant, and the remaining control variables performed similar to Table 3. Prisons do not con- tribute to employment growth under any circumstances; in states shifting rapidly towards privatized prison management, new prisons impede employment growth.4

As noted, new prisons are not randomly distributed across counties. Instead, it appears that new prisons went to rela- tively disadvantaged counties (Eason, 2010, see also Fig. 1). We pursued analyses to assess the degree to which job growth also varied along these lines. In the analyses summarized in Table 5, we employed the same measures and modeling proce- dures as employed in Table 4 – with one difference. We divided the sample into two based on several sociodemographic characteristics (above and below the median).

The results in Table 5 are noteworthy. Consistent with the trends reported in Table 4, there is no evidence that prisons (established or new) contribute to local job growth. Instead, where privatization is proceeding rapidly, these results suggest that new prisons impede local employment growth (again consistent with Table 4). But Table 5 introduces a new wrinkle. It appears that the negative impact of new prisons is not concentrated in relatively disadvantaged counties. Instead, in the con- text of a rapid shift toward privatization, new prisons impede employment growth for the counties with less poverty, higher educational attainment and a lower percentage of minority residents. As discussed below, this suggests that the negative impacts of prison privatization are felt most directly in counties with higher paying and more secure jobs.

4 The analyses summarized in Table 4 divides states at the 75th percentile based on the pace of privatization. In other analyses (not reported here), we divided states at center of the distribution (median and mean) and performed comparable analyses. The results were similar to those discussed in this paper. That is, for counties in states shifting towards privatization at a more rapid pace, prisons are inversely related to employment growth. Results available upon request.

Table 3 Determinants of employment growth in nonmetropolitan counties, 1997–2004 (two-stage least squares, n = 2242).a

Determinant All non-metropolitan countiesb

Rapid growth in privatization (at or above 75th percentile)

Slower growth in privatization (below 75th percentile)

Land-Deane spatial effects term �0.04 (0.05) 0.02 (0.12) �0.04 (0.04) New prison (1991–1997) �0.26 (0.23) �0.84** (0.30) �0.00 (0.32) Established prison (prior to 1991) �0.06 (0.18) �0.36 (0.36) �0.09 (0.21) Employment (1991–1997 w/o

construction) 86.79*** (99.3) 118.68*** (16.79) 81.80*** (11.43)

Construction employment (1991– 1997)

39.73*** (5.73) 25.31* (11.26) 46.34*** (5.59)

% of labor force w/BA (1980–1990) 8.51*** (1.77) 5.02* (2.43) 8.52*** (2.16) % of labor force w/HS diploma (1980–

1990) 0.96 (1.68) �1.41 (3.08) 1.98 (1.96)

% Core manufacturing (1980–1990) �5.28* (2.18) �10.83** (3.91) �4.90* (2.33) % Core non-manufacturing (1980–

1990) 1.69 (1.95) 1.16 (3.22) 1.52 (2.47)

% Competitive sector (1980–1990) 0.43 (1.30) 1.57 (2.04) 0.67 (1.60) % State sector (1980–1990) 6.47* (2.54) 7.46* (4.77) 5.96* (2.76) Commercial aircraft activity (1980–

1990) �3.11* (1.58) �2.81* (2.52) �3.15# (1.85)

Commercial bank deposits (1980– 1990)

�0.01 (0.03) 0.02 (0.03) �0.02 (0.03)

Per capita property tax (1977–1987) �0.52 (0.38) �0.38 (0.49) �0.35 (0.45) General revenues of local government

(1977–1987) 0.12* (0.05) 0.35*** (0.10) 0.05 (0.06)

Constant �294.20** (114.11) �502.89** (162.49) �298.58* (136.88)

Adjusted R2 0.42 0.62 0.38 n = 2247 n = 487 n = 1755

# p < 0.10 two-tailed test. * p < 0.05 two-tailed test. ** p < 0.01 two-tailed test. *** p < 0.001 two-tailed test.

a Coefficients and standard errors have been multiplied by 100 to improve readability.

606 S. Genter et al. / Social Science Research 42 (2013) 596–610

5. Discussion

Private operation of some prisons induces change across a state’s criminal justice system. Proponents of privatization anticipate a positive sum outcome due to this system-wide change: lower costs to the taxpayers, the same or improved ser- vice delivery, and improved salary and working conditions. Opponents of privatization characterize these systemic changes as a race to the bottom, wherein salaries, job security and working conditions decline as privatization gains momentum.

Whether or not they bring jobs, critics object to the subsidization of private prison construction under the banner of eco- nomic development. Private prisons have received extensive economic subsidies from local, state, and the federal govern- ments. A study of 60 private prisons with a capacity of 500 or more beds found that a total of $628 million in tax-free bonds and other government issued securities were used to finance the prisons, that more than one-third received property tax abatements or other tax reductions, and that 23% received various infrastructure subsidies, such as water, sewer, or util- ity hookups, access roads, or other publicly financed improvements (Mattera and Khan, 2001; see also Shichor, 1995).

Where privatization is moving rapidly, we found a negative relationship between a new prison and local job growth. This finding is consistent with prior research that many of the jobs promised by private prisons simply do not materialize. For example, Wackenhut Corporation (now the GEO group) secured a contract to build and operate a correctional facility in East Mississippi in 1997, with officials predicting the facility would create up to 350 jobs. However, as of 2005, the prison em- ployed only 220 people (American Federation of State, County, and Municipal Employees, 2010). Further, non-residents fill many jobs at new prisons (Gilmore, 2007; Mestas, 2010). It is not only the case that private prisons generate surprisingly few jobs. But they are less likely to contribute to employment stability because of extremely high turnover (with low wages and salaries no doubt contributing to this trend). Nationally, it is estimated that annual employee turnover in private prisons is 52% (American Federation of State, County, and Municipal Employees, 2010) compared to estimates of 12–25% in public pris- ons (Minor et al., 2009). In Texas, annual staff turnover rates of 90% were found in private prisons, compared to 24% for pub- licly employed correctional officers in that state (Texas Senate Committee on Criminal Justice, 2008). With regard to wages, the US Department of Labor (2010) reports that in 2008, the median annual wage for correctional officers employed by the federal government was $50,830, $38,850 for officers employed by state governments, and $37,510 for those employed by local governments. But, officers employed in privately operated prisons earned only $28,790.

Critics – calling attention to anemic job creation, depressed wages, high turnover and public subsidies – anticipate zero- sum dynamics: corporations have profited and their executives have earned exorbitant salaries (Mattera et al., 2003). But privatization has yielded few savings for the public sector, contributed to an erosion of service delivery, and undermined

Table 4 Determinants of employment growth in nonmetropolitan counties with slope dummy interaction terms, 1997–2004 (two-stage least squares, n = 2242).a

Determinant With zero-order measure of privatization context

With slope dummy interaction terms

Land-Deane spatial effects term �0.04 (0.05) �0.03 (0.05) New prison (1991–1997) �0.25 (0.23) – Established prison (prior to 1991) �0.06 (0.18) – Rapid growth in prison privatization (75th percentile or above,

dummy variable) �0.06 (0.14) –

Rapid growth in private prisons (75th percentile or above) New Prison (1991–1997) – �0.67** (0.25)

Established prison (prior to 1991) – �0.23 (0.27)

Slower growth in private prisons (below 75th percentile or above) New Prison (1991–1997) – �0.01 (0.28) Established Prison (prior to 1991) – 0.01 (0.20) Employment (1991–1997 w/o construction) 86.76*** (9.92) 87.02*** (9.95) Construction employment (1991–1997) 39.64*** (5.69) 39.45*** (5.75) % of labor force w/BA (1980–1990) 8.61*** (1.75) 8.68*** (1.79) % of labor force w/HS diploma (1980–1990) 0.89 (1.65) 0.94 (1.68) % Core manufacturing (1980–1990) �5.43* (2.12) �5.70* (2.16) % Core non-manufacturing (1980–1990) 1.72 (1.98) 1.61 (1.95) % Competitive sector (1980–1990) 0.49 (1.34) 0.47 (1.30) % State sector (1980–1990) 6.61** (2.46) 6.73** (2.55) Commercial aircraft activity (1980–1990) �3.11* (1.58) �3.05# (1.56) Commercial bank deposits (1980–1990) �0.01 (0.03) �0.01 (0.03) Per capita property tax (1977–1987) �0.49 (0.37) �0.49 (0.38) General revenues of local government (1977–1987) 0.12* (0.05) 0.13* (0.05)

Constant �293.62** (113.55) �294.90** (113.92)

Adjusted R2 0.42 0.42

# p < 0.10 two-tailed test. * p < 0.05 two-tailed test. ** p < 0.01 two-tailed test. *** p < 0.001 two-tailed test.

a Coefficients and standard errors have been multiplied by 100 to improve readability.

Table 5 County characteristics and determinants of employment growth, 1997–2004 (two-stage least squares, n = 2242).*

Faster growth in private prisons (at or above 75th percentile)

Slower growth in private prisons (below 75th percentile)

Established prison, as of 1997

New prison, 1997–2004

Established prison, as of 1997

New prison, 1997–2004

Relatively disadvantaged counties Percentage of adult populations with high school

degree below median N/S N/S N/S N/S

Percentage of adult populations with bachelor’s degree below median

N/S N/S N/S N/S

Non-white population above median N/S N/S N/S N/S Percentage of families in poverty above median N/S N/S N/S N/S

Relatively advantaged counties Percentage of adult populations with high school above

median N/S Negative N/S N/S

Percentage of adult populations with bachelor’s degree above median

N/S Negative N/S N/S

Non-white population below median N/S Negative N/S N/S Percentage of families in poverty below median N/S Negative N/S N/S

* After dividing counties at the median for selected characteristics (percentage of adult populations with high school degree, percentage of adult popu- lations with BA degree, non-white population [percentage], and percentage of families in poverty), separate analyses were performed. Modeling followed the same procedures and included all variables presented in Table 4 (see above).

S. Genter et al. / Social Science Research 42 (2013) 596–610 607

working conditions. Our findings cannot speak to the full range of outcomes – but they do point towards another dimension of zero-sum outcomes. The counties hosting prisons are placed at risk by their failure to expand employment and by the depressed pay and high turnover that is characteristic of private prisons. Given that these negative findings are prevalent

608 S. Genter et al. / Social Science Research 42 (2013) 596–610

among relatively advantaged counties (as measured by minority population, education, and poverty levels, see Table 5), it appears that the quality and availability of middle-class jobs are placed at risk.

6. Conclusion

Most studies of prison privatization have focused on cost efficiencies (or inefficiencies) and comparisons of the ‘‘quality’’ of imprisonment with that found in public facilities. Our research broadens the scope to consider local employment impacts. Consistent with recent studies, we find that prisons do not contribute to growth. Furthermore, we demonstrate that a state’s decision to privatize some of its prisons does not improve employment prospects for the host county. In fact, it impedes them.

Proponents of privatization have long stressed its potential to induce a transformation of public agencies and services these agencies provide (McFarland et al., 2002; Osborne and Gaebler, 1992). However, few studies have empirically exam- ined this assertion. Our research into county employment impact of prisons takes initial steps towards assessing this claim; our findings challenge these optimistic claims. That is, in states undergoing a rapid shift towards privatization, prisons run by public agencies have reduced staffing to levels similar to private prisons (see Fig. 3, especially trends in medium and max- imum security facilities). In turn, we provide evidence that these shifts in staffing have consequences for the rural counties hosting prisons. Specifically, new prisons in states undergoing a rapid shift towards privatization are inversely related to employment growth.

These findings offer preliminary evidence bearing on an important but understudied phenomena. But this research is not offered as the final statement. On the contrary, this research may be of greater value for the questions that it poses. In the context of aggressive efforts to privatize prisons in states around the country (Burnett, 2011; Caputo, 2011; Vardon, 2011), research into the impacts of such shifts is especially valuable. In addition to local employment impacts that we have exam- ined, research into the quality and effectiveness of prison operations (escapes, riots, assaults and recidivism) would be sali- ent. Given the findings that we present, the contrast should not be simply public versus private prisons (as is common), but the contrasts should consider the larger administrative context, including the extent and pace of shifts to privatization. While our focus is on corrections, when other aspects of public administration and other public services are examined, it would also be advantageous to consider the larger context.

The direct effects offer the most obvious explanation for the inverse relationship between local employment growth and prisons in the context of rapid privatization, i.e., downward pressure on salary, benefits and staffing levels at prisons. But a potentially complementary explanation, one emphasizing indirect mechanisms, is also plausible. Hooks et al. (2004) suggest that prisons might impede employment growth due to the opportunity cost of a misguided investment. One of the ironies of the prison expansion boom of the 1980s and 1990s is that the resources that were used to fund this effort were often di- verted from programs, such as education, that have a well documented and positive impact on local employment opportu- nities (Crookston and Hooks, 2012). According to Hallett (2002:375, emphasis in the original), the incarceration boom diverted public resources ‘‘toward prisons and away from public programs in education and childcare.’’ Coming to terms with the impacts of privatization is of growing importance. Facing budget pressures, a growing number of states are contracting out public services, including corrections. Our findings suggest these trends will have deleterious consequences for rural counties and their efforts to expand and retain middle-class jobs.

Acknowledgments

We would like to thank Peter Wagner of the Prison Policy Initiative (http://www.prisonpolicy.org/) for providing data and guidance on prison data. Thomas Familia and Katie Vendermark provided valuable assistance in data management. Gregory Hooks gratefully acknowledges the support of the Open Society Institute (Soros Justice Fellowship Program).

References

Abrams, Kathleen S., Lyons, William, 1987. Impact of Correctional Facilities on Land Values and Public Safety. Joint Center for Environmental and Urban Problems, Florida International University, Miami, Fla.

American Civil Liberties Union of Ohio, 2011. Prison for Profit: A Look at Prison Privatization. Cleveland, OH: American Civil Liberties Union of Ohio. <http:// www.acluohio.org/issues/CriminalJustice/PrisonsForProfit2011_04.pdf> (accessed 19.04.11).

American Federation of State, County, and Municipal Employees, 2010. Prison Privatization: Don’t Be a Prisoner to Empty Promises. <http:// www.afscme.org/publications/1253.cfm> (accessed 01.10.10).

Austin, James, Coventry, Garry, 2001. Emerging Issues on Privatized Prisons. U.S. Department of Justice, Washington, D.C.. Bartik, Timothy, 1991. Who Benefits from State and Local Economic Development Policies? Kalamazoo, MI, W.E. Upjohn Institute for Employment Research. Bartik, Timothy, 1994. The effects of metropolitan job growth on the size distribution of family income. Journal of Regional Science 34, 483–501. Beale, Calvin L., 1997. Rural prisons: an update. Rural Development Perspectives 11, 25–27. Besser, Terry L., Hanson, Margaret M., 2005. Development of last resort: impact of new state prisons on small town economies. Journal of the Community

Development Society 35, 1–15. Blankenship, Susan E., Yanarella, Ernest J., 2004. Prison recruitment as a policy tool of local economic development: a critical evaluation. Contemporary

Justice Review 7, 183–198. Burnett, John, 2011. Private Prison Promises Leave Texas Towns In Trouble. National Public Radio (March 28, 2011). <http://www.npr.org/2011/03/28/

134855801/private-prison-promises-leave-texas-towns-in-trouble> (accessed 19.04.11). Camp, Scott D., Gaes, Gerald G., 2002. Growth and quality of U.S. private prisons: evidence from a national survey. Criminology and Public Policy 1, 427–449.

S. Genter et al. / Social Science Research 42 (2013) 596–610 609

Caputo, Mark, 2011. Florida Lawmakers Look to Privatize Prisons. Miami Herald (March 29, 2011). <http://www.miamiherald.com/2011/03/29/2139213/ florida-lawmakers-look-to-privatize.html> (accessed 19.04.11).

Carlson, Katherine A., 1991. What happens and what counts: resident assessments of prison impacts on their communities. Humboldt Journal of Social Relations 17, 211–237.

Carlson, Katherine A., 1992. Doing good and looking bad: a case study of prison/community relations. Crime and Delinquency 38, 56–69. Corrections Corporation of America, 2010. Corrections Corporation of America. <http://www.correctionscorp.com> (accessed 01.10.10). Crookston, Andrew., Hooks, Gregory., 2012. Community colleges, budget cuts and jobs: the impact of community colleges on employment growth in rural

U.S. counties, 1976–2004. Sociology of Education 85, 350–372. Dao, James, 1997. Prison in the Forest; The Itinerant Jailer Is and Adirondack Plan’s Best Booster. The New York Times, September 16. Donahue, John.D., 1989. The privatization decision: public ends, private means. Basic Books. Durham, Alexis, 1993. The future of correctional privatization: lessons from the past. In: Bowman, G., Hakim, S., Seidenstat, P. (Eds.), Privatizing Correctional

Institutions. Transaction Publishers, New Brunswick, NJ, pp. 33–49. Dyer, Joel., 2000. The perpetual prisoner machine: how america profits from crime. Boulder, CO: Westview Press. Eason, John, 2010. Mapping prison proliferation: region, rurality, race, and disadvantage in prison placement. Social Science Research 39, 1015–1028. Feeley, Malcolm, 2002. Entrepreneurs of punishment: the legacy of privatization. Punishment & Society, 321–344. Gaes, Gerald G., 2005. Prison privatization in Florida: promise, premise, and performance. Criminology and Public Policy 4, 83–88. Gaes, Gerald G., Camp, Scott D., Nelson, Julianne B., Saylor, William G., 2004. Measuring Prison Performance: Government Privatization and Accountability.

Altamira Press. Gilmore, Ruth Wilson, 2007. Golden Gulag: Prisons, Surplus, Crisis, and Opposition in Globalizing California. University of California Press. Glasmeier, Amy K., Farrigan, Tracey, 2007. The economic impacts of the prison development boom on persistently poor rural places. International Regional

Science Review 30, 274–299. Halaby, Charles N., 2004. Panel models in sociological research: theory into practice. Annual Review of Sociology 30, 507–544. Hallett, Michael A., 2002. Race, crime and for-profit imprisonment. Punishment and Society 4, 369–393. Hallett, Michael A., 2006. Private Prisons in America: A Critical Race Perspective. University of Illinois Press, Chicago. Hooks, Gregory, Mosher, Clayton, Rotolo, Thomas, Lobao, Linda, 2004. The prison industry: carceral expansion and employment in U.S. Counties, 1969–

1994. Social Science Quarterly 85, 37–57. Hooks, Gregory, Mosher, Clayton, Genter, Shaun, Rotolo, Thomas., Lobao, Linda, 2010. Revisiting the impact of prison building on job growth: education,

incarceration, and county-level employment, 1976–2004. Social Science Quarterly 91, 228–244. King, Ryan S., Mauer, Marc, Huling, Tracy, 2004. An analysis of the economics of prison siting in rural communities. Criminology and Public Policy 3, 453–

480. Kirkham, Chris, 2012. Private Prison Corporation Offers Cash In Exchange For State Prisons. Huffington Post (February 14). <http://

www.huffingtonpost.com/2012/02/14/private-prisons-buying-state-prisons_n_1272143.html> (accessed 06.06.12). Lawrence, Sarah, Jeremy Travis. 2004. The new landscape of imprisonment: mapping America’s Prison Expansion. Washington, DC: Urban Institute. Lobao, Linda, Rulli, Jamie, Brown, Lawrence A., 1999. Macrolevel theory and local-level inequality: industrial structure, institutional arrangements, and the

political economy of redistribution, 1970 and 1990. Annals of the Association of American Geographers 89 (4), 571–601. Logan, Charles H., 1990. Private Prisons: Cons and Pros. Oxford University Press. Mattera, Phillip, Khan, Mafruza, 2001. Jail Breaks: Economic Development Studies Given to Private Prisons. Good Jobs First. <http://www.goodjobsfirst.org/

search/searchres.cfm> (accessed 18.06.08). Mattera, Phillip, Khan, Mafruza, Nathan, Stephen, 2003. Corrections Corporation of America: A Critical Look at its First Twenty Years. Grassroots Leadership,

Charlotte, NC. McFarland, McGowan, Chris, O’Toole, Tom, 2002. Prisons, Privatization, and Public Values. <http://government.cce.cornell.edu/doc/html/

PrisonsPrivatization.htm> (accessed 01.01.10). McShane, Marilyn, Williams III, Frank P., Wagoner, Carl P., 1992. Prison impact studies: some comments on methodological rigor. Crime and Delinquency,

38105–38120. Mestas, A., 2010. Walsenberg Struggling with Prison Closure. March 7. <http://www.mcclatchy.com> (accessed 07.06.10). Minor, K., Dawson-Edwards, C., Wells, J., Griffith, C., Angel, E., 2009. Understanding staff perceptions of turnover in corrections. Professional Issues in

Criminal Justice 4, 43–57. Ogle, Robbin., 1999. Prison privatization: an environmental catch-22. Justice Quarterly 16, 579–600. Osborne, David, Ted, Gaebler, 1992. Reinventing Government: How the Entrepreneurial Spirit is Transforming the Public Sector. Addison-Wesley Publishing

Company. Price, Byron E., Riccucci, Norma M., 2005. Exploring the determinants of decisions to privatize state prisons. American Review of Public Administration 35,

223–235. Price, Byron E., 2006. Merchandizing Prisoners: Who Really Pays for Prison Privatization? Praeger Press. Sclar, Elliott D., 2000. You Don’t Always Get What You Pay For: The Economics of Privatization. Cornell University Press, Ithica, NY. Sechrest, Dale K., 1991. Understanding the corrections and community response to prison siting. Humboldt Journal of Social Relations 17, 1–16. Sechrest, Dale.K., 1992. Locating prisons: open versus closed approaches to siting. Crime and Delinquency 38, 88–104. Shichor, David., 1992. Myths and realities in prison siting. Crime and Delinquency 38, 70–87. Shichor, David, 1995. Punishment for Profit: Private Prisons/Public Concerns. Sage, Thousand Oaks, CA. Somers, Margaret, Fred Block, 2005. From Poverty to Perversity: ideas, Markets, and Institutions over 200 Years of Welfare Debate. American Sociological

Review 70, 260–287. Texas Senate Committee on Criminal Justice, 2008. Interim Report to the 81st Legislature. <http://www.senate.state.tx.us/75r/senate/commit/c590/

c590_80.htm> (accessed 28.08.10). Thies, Jeanie S., 1998. The Big House in a Small Town: The Economic and Social Impacts of a Correctional Facility on it Host Community. Dissertation.

University of Missouri. US Bureau of Economic Analysis, 2008. County Wage and Salary Summary CA34, 1969-2006. Electronic datafile. Washington, D.C.: Bureau of Economic

Analysis. <http://www.bea.gov/regional/reis/ca34.cfm> (accessed 01.10.10). US Department of Commerce, 1995. USA Counties 1994 on CD-ROM. Department of Commerce, Washington, DC. US Department of Justice, 1998. Census of State and Federal Adult Correctional Facilities, 1995 [Computer file]. ICPSR, Ann Arbor, MI. US Department of Justice, 2004. Census of State and Federal Adult Correctional Facilities, 2000 [Computer file]. ICPSR, Ann Arbor, MI [producer and

distributor]. US Department of Justice, 2005. Census of State and Federal Adult Correctional Facilities, 2005 [Computer file]. ICPSR, Ann Arbor, MI. US Department of Justice, 2012. Prisoners in 2010. <http://www.bjs.gov/content/pub/pdf/p10.pdf> (retrieved 06.06.12). US Department of Labor, 2010. Occupational Outlook Handbook, 2010-2011 Edition. <http://www.bls.gov/oco/ocos156.htm> (retrieved 30.08.10). US Federal Aviation Administration, 1981. Air Traffic Activity. Government Printing Office, Washington, DC. US Federal Aviation Administration, 1990. Air Traffic Activity. Government Printing Office, Washington, DC. Vardon, Joe, 2011. Kasich’s Beginning: An Adventurous, Divisive 100 Days. The Columbus Dispatch (April 19, 2011). <http://www.dispatchpolitics.com/live/

content/local_news/stories/2011/04/19/copy/an-adventurous-divisive-100-days.html?adsec=politics&sid=101> (accessed 19.04.11). Walker, Donald, 1994. Privatization in corrections. In: Kratcoski, P. (Ed.), Correctional Counseling and Treatment. Waveland Press, Prospect Heights, IL, pp.

570–585.

610 S. Genter et al. / Social Science Research 42 (2013) 596–610

Wolman, Harold, Spitzley, David, 1996. The politics of local economic development. Economic Development Quarterly 10, 115–150. Young, Andrew, Levy, Daniel, Higgins, Matthew, 2004. Many Types of Human Capital and Many Roles in U.S. Growth: Evidence from County-Level

Educational Attainment Data. In: Prepared for Presentation at the CESifo/Harvard University-PEPG Conference on ‘‘Schooling and Human Capital Formation in the Global Economy: Revisiting the Equity-Efficiency Quandary.’’ September, Munich, Germany. <http://129.3.20.41/eps/pe/papers/0403/ 0403002.pdf> (accessed 01.10.10).