Gao Report to Us Congress Re State and Fed Prisons Construction and Operations Cost May 1992
Download original document:
This text is machine-read, and may contain errors. Check the original document to verify accuracy.
_..- - GAO United States General May 1992 Office ____ ._ -----_ _-~_._ STATEANDFEDERAL PRISONS Factors That Affect Construction and Operations Costs 146913 RESTRICTED--Not to be released outside the General Accounting Office unless specifically appro,ved by the Office of Congressio Relations. ----------.GAO/GGD-92-73 _^ R&x-t to Congressional Requesters -.- ..- I Accounting 554362 -WY - GAS3 Acconndng United States General Wahhgtm, Of&e D.C. 20548 General Government Divieion B-248117 May 19,1992 The Honorable Bob Graham United States Senate The Honorable Dennis DeConcini t ,,/’ United States Senate The Honorable Richard Bryan United States Senate ,J The Honorable Joseph Lieberman / United States Senate The Honorable Herb Kohl United States Senate : .,d In response to your joint request, we recently issued a report that compared construction and operations costs for medium security state and federal prisons opened between 1986and 1989and identified opportunities for savings in the federal system.1For the purposes of that report, we aggregateddata for the state and federal prisons in our sample and, except for a few examples, did not include data for individual prisons. After the report wss issued, your offices suggestedthat publishing the cost information for the individual state and federal prisons we sampled and the major reasons for cost differences might encourage some of the higher cost jurisdictions to try to reduce costs. We agreed to prepare a report on the information we obtained for the individual prisons and the factors that contributed to differences in their construction and operations costs. Background The state and federal governments are spending billions for new prison construction to accommodate continuing increases in inmate populations. According to the February 1992 Corrections Compendium, 26 state corrections systems requested a total of $2.3 billion for the 1992-1993fiscal year. Included were requests for 86 new facilities, which would add over 66,000new prison beds2 Texas alone asked for more than $600 million in construction funds to add over 26,000new beds. The Federal Bureau of Prisons (BOP) is in the midst of an unprecedented expansion program that (GAOfGGD-02-3,Oct. --,m.....s_,_ ,_ _ aA “bed” is a generic unit of measure for a prison’s inmate capacity. For example, a MO-bed prison would have a rated capacity of 600 inmates. Page 1 OAOioOD-92-72 Prbon Colt Factore B-248117 will double its 1989prison capacity by 1996at a cost of about $3 billion, In reality, construction costs are only the down payment on a prison’s total cost to society. BOP has estimated that operating a prison over its useful life costs 16 to 20 times its construction costs. Prisons can vary widely in size, design, and costs of construction. There is no universal standard or “cookie cutter” prison design, although some jurisdictions have adopted their own standard layouts. Many factors can influence a prison’s ultimate structure, including its intended capacity, the security level of inmates expected to be housed in the facility, the urgency of need for prison beds, the jurisdiction’s desire to meet the accreditation standards of the American Correctional Association (ACA), budget constraints in the jurisdiction, and the corrections policy and philosophy of the jurisdiction. Resultsin Brief Construction costs varied widely among the medium security state and federal prisons we sampled.At the 36 medium security prisons included in our sample (32 state, 4 federal), construction costs ranged from $11,243to $93,333per bed and averaged$66,374.3The most important factor contributing to differences in prison construction costs per bed was the amount of space provided, measuredin terms of gross square feet (GSF) per inmatem4 This factor accounted for 95 percent of the variability in per bed construction costs for the 36 prisons in our sample. Other factors that m ight have contributed to the cost differences were the type of building structure, the housing area design and layout, whether the facility was designedfor a m ix of security levels, and geographic location. We tested alternative combinations of these factors. We found that none of the combinations explained a significant amount of additional variability a in construction costs beyond that explained by the amount of space provided to each inmate. Although state and federal prison systemsare revising their design standards to allow for more double celling of inmates, we believe all of the five factors identified above will continue to affect differences in prison construction costs after the revisions are fully implemented. 9’he cost per bed is the total cost of the facility divided by the number of inmates that the facility was designed to accommodate. The cost per bed includes costs for all areas of the prison, including housing, recreation, education, and prison industry. ‘Gross square feet is defined by the American Institute of Architects as the sum of the areas of the several floors of a building, measured from the exterior faces of exterior walls or from the centerline of walls separating buildings. The areas of covered walkways, porches, and similar space are multiplied by a factor of 5. GAO/GGD-92-73 Prieon Cost Factors Pa2e 2 ‘i,: B-242117 Operations costs also varied widely at the 23 prisons (21 state, 2 federal) that provided operations cost information, ranging from $22.26to $81.08 per inmate per day (referred to as an inmate day) and averaging$41.93. The key factors that contributed to the operations cost differences were personnel salariesand related expenses,inmate-tostsffratios, and the amount spent on supplies, materials, and food. Objectives,Scope, and Methodology Our objective was to identify the factors that contributed to differences in prison construction and operations costs. We obtained prison construction and operations cost information from the questionnairesdeveloped for our recently issued prison cost report (see footnote 1). The questionnaires were designedto obtain reliable and comparable data for each state and federal prison that met the following criteria: opened between 1986and 1989; new, independent facilities; designedto house adult males; designedfor a population of 200 inmates or more; and in operation for one full year at or near design capacity (operations costs only). We took several steps to ensure that the questionnaireswould obtain sufficient data to perm it meaningful comparisonsdespite the great number and diversity of reporting jurisdictions. In designing the questionnaires,we met with architects, engineers,and cost accountants to identify the key information that would account for differences in design and costs. To encourageparticipation in our study and lessen the burden of responding, we focused the questionnaireson information that (1) was readily available in the states’departments of correctioti and BOP; (2) was, for the most part, consistently defined and captured in standard government cost accounts, and (3) was objective, measurable,and comparable (e.g., size, populations, number of rooms). We pretested the questionnairesat three state corrections departments and BOP to further increase the likelihood that the respondents would understand how to complete them and provide comparable and reliable data. We also followed up with respondents that appeared to have submitted incomplete or erroneous data On the other hand, we did not make a detailed cost reconciliation for each prison, nor did we assess what effect, if any, prison design and construction may have had on enhancing prisoner rehabilitation and the incidence of prison violence. Page 3 GAWGGD-92-73 Prima Coot Fwtmw B=242117 We used ACA’S1990Directory of Juvenile and Adult Correctional Departments,Institutions, Agencies,and Paroling Authorities as our source for states and prisons to receive the questionnaires.We mailed questionnairesto BOP, the District of Columbia, and the 37 states that the Directory identified as building prisons during the target period. This distribution covered 62 state and 4 federal prisons. BOP provided construction cost information for all four facilities built between 1986and 1989which were, for the most part, all designedto house a majority of medium security inmates. These prisons are the Federal Correctional Institution (m) Phoenix, Ark; m Marianna, Fla.; F+CI Sheridan,Oreg,; and FCIMcKean,Pa. BOP’S construction cost information and our analysisdid not include a 126bed temporary dorm itory built at FCI EIOP also provided operations cost Phoenix in 1990at a cost of $6Q8,6QQ. information for the two prisons that had been in operation for at least one year at or near their design capacity. These prisons are FCXPhoenix and FCI Of the 62 questionnairesmailed to state prisons, 11 were not used because we later found that the projects did not meet one or more of our criteria Two states voluntarily completed questionnairesfor prisons that met our selection criteria but that were not listed in ACA’S 1990directory. Of the 63 state prisons we expected to participate, 46 (from 30 states and the District of Colwnbia) returned the construction portion of our questionnaire,and 29 (from 21 states and the District of Columbia) returned the operations cost portion. However, we reduced the operations cost sample to 28 becauseone jurisdiction did not isolate operations costs by individual departments, and thus the questionnaire responsewas not usable. Becausethe four federal prisons built during the defined time frame were designedto house mostly medium security inmates, we 0 reduced the state sample to include only prisons designedto house a majority of medium security inmates. Our final tally was construction cost data from 32 prisons in 20 states and the District of Cohunbia, and operations cost data from 21 prisons in 16 states. A list of the state prisons that reported construction cost information is in appendix III. A list of the state prisons that reported operations cost information is in appendix Iv. To facilitate our analysisof construction costs, we divided the state and federal prisons in our sample into three cost groups-low, medium, and high. When the 36 prisons were arrayed in order of cost per bed from low to high, natural breakpoints occurred between the low cost and medium cost groups and between the medium cost and high cost groups. Pqe 4 GAO/GGD-92-73 Prhon Ckut Flctom B-240117 Construction costs per bed in the $11,243to $24,679range were classified as low cost, those in the $46,007to $73,438range were classified as medium cost, and those in the $83,771to $93,333range were classified as high cost. Similarly, to analyze operations costs, we divided the prisons into low, medium, and high cost groups. We used breakpoints that existed in the daily operations costs per inmate to define the three cost groups. Daily operations costs per inmate in the $22 to $37 range were classified as low cost, those in the $42 to $61 range were classified as medium cost, and those in the $69 to $81 range were classified as high cost. There was no direct relationship between the operations cost groups and the construction cost groups. We used standard statistical techniques to determ ine the relationships between prison construction costs and the factors for which we obtained data. These techniques allowed us to determ ine the amount of variability within different measuresof construction costs that was explainable by each factor and by various combinations of factors. We were able to identify the factors that explained at least 96 percent of the variability of each of the following three measuresof prison construction costs: total construction costs, costs per bed, and costs per GSF.These factors are discussedindividually in the report. Other factors were significant for particular groups of prisons but were not consistent across all of the prisons included in the analysis.For example, housing area design and layout proved to be important in explaining construction costs for state prisons, but not for federal prisons. The results of our statistical analysis must be considered in light of certain lim itations inherent in our study. Becausethe 36 prisons included in the analysis were not randomly selected,we cannot infer that they are representative of the universe of prisons. If additional ‘or tiother set of prisons were included in the analysis,the results m ight be’different. It is also possible that additional factors for which data was not collected may affect prison construction costs. We did our work between ,December1991and March 1992in accordance with generally accepted government auditing standards. Construction Costs VbriedW idely ” Construction costs varied widely among the medium security prisons we sampled.At the 36 prisons, total construction costs ranged from a low of $6,464,644(rated capacity of 312 inmates) to a high of $266,066,796(rated capacity of 2,916inmates). Per bed construction costs ranged from $11,243 Page 6 -D-92-78 P&on Coat Factmu B=242117 to $93,333and averaged $50,374.The cost per bed of the high cost prisons ($87,271) averaged almost five times as much as the cost per bed of the low cost prisons ($17,730).The average per bed cost of the medium cost prisons was $58,282.See figure 1. Flgure 1: Avorago Prlron Conatruttlon Co8ta p8r Bed 100000 Dollan In thousands 60000 60000 70000 60000 60000 40000 20000 2oooo 10000 0 CO81 coet coift prisons prisons ----- The Amount of Space Provided to Inmates Accounted for Most Construction Cost Differences prlronr Weighted average for 36 prisons, $56,374 Of the factors we examined, the amount of space provided, measured in terms of GSFper inmate, accounted for most of the differences we found in prison construction costs per bed. The high cost prisons provided an overall average of 554 GSFper inmate, over two and one-hslf times the average of 215 GSFper inmate provided at the low cost prisons. After testing alternative factors, we found that, when considered independently, the amount of space provided to inmates accounted for 95 percent of the variability in cost per bed. Figure 2 illustrates the close relationship between cost per bed and GSFper inmate. Page 9 GAO/GGD-92.73 Prison Coot Factom L B.248117 Agun 2: Comparlron Botwoen Average per Bad Conrtructlon Cortr and Grou Square Feet per Inmate in State and Fodenl Prhonr loo 90 80 70 60 60 40 30 20 10 0 Dollsrs In thousands Aversgo par brd construction costs 0 100 200 300 400 500 600 700 800 900 1000 Gross squrrrr fort par lnmrtr Low cost prisons Medium cost prisons High cost prisons ‘.Sypeof Structure, Housing Configuration,M ix of Security Levels,and GeographicLocation Also Affected Construction Costs In addition to the amount of space provided to inmates, we examined several other factors in terms of per bed construction costs. One of these factors was the type of building structure. The National Directory of Corrections Construction, published by the National Institute of Justice (April 1988),classifiesprisons into several general types of structures, including an integrated structure (one building); clusters (a number of individual buildings that are interconnected); and campus style (a number of individual buildings that are not connected).6Although construction costs varied for each design style becauseof factors such as size and housing layout, our analysisfound that integrated structures, on average, were the most costly of the three types of structures, followed by clusters and campusstyle. Of the seven high cost prisons in our sample, five were either single buildings or clusters. In contrast, seven of the eight low cost prisons were campus style. The design of prison housing units also contributed to construction cost differences at our sample prisons. The high cost prisons reported that, The prteons in our sample were in these three categories. Other types of structurea described in the directory were hi rise (one building, more than four stories in height); ladder, telephone pole (linear cell blocks arrange -@ii- in parallel off a central connecting corridor); wheel, apok e, or radial (linear cell blocks that emanate from one central contsol area like spokes f&m the hub of a wheel); and courtyard (linear cell blocks interconnected around a central enclosed courtyard). Pnge 7 GMMGQD-92-72 Prbon Coet Factors a B-242117 overall, 90 percent of their beds were designedto be in single cells, less than one percent in multiple occupancy cells, and 10 percent in dorm itories. In contrast, only 4 percent of beds at the low cost prisons were designedto be in single cells, while 60 percent were in multiple occupancy cells and 36 percent in dorm itories. Another factor that contributed to construction cost differences was whether the prison was built to accommodateinmates &om different security levels. Construction costs per bed tended to increase as the percentageof medium security beds declined. Overall, the high cost prisons classified 76 percent of their beds as medium security, compared to 89 percent for the medium cost prisons and 90 percent for the low cost prisons. The geographic location of the prison also affected construction costs. According to the National Institute of Justice and the ACA, prison construction costs tend to be higher in the Northeast and West and lower in the South and M idwest due to significant differences in the cost of materials and prevailing labor rates. The prisoxisin our sample reflected those tendencies.Of the 8 low cost prisons, 6 were in the South, while only 2 of the 21 medium cost and 1 of the 7 high cost prisons were in the South. Conversely,no Northeast prisons were in the low cost group, while five Northeast prisons were in the medium cost group and three in the high cost group. Another indicator of the importance of geographic location is its effect on the cost per GsF.The cost per osFis, in effect, the measure of the amount of space the jurisdiction was able to buy for its money, independent of the number of inmates the prison was designedto house. It encompassessuch cost factors ss site acquisition and preparation as well as materials and labor. To some extent, cost per GSFcould even be a measure of the economic conditions and contracting environment during the period leading up to construction. Cost per GsFat the 36 prisons ranged from $68.06to $216.60and averaged$129.48.We analyzedthe effect of various factors on the cost per GSFand found that about 96 percent of the variability in cost per GSFwas explainable by the national construction cost index. This Index is a surrogate measurefor the state in which the prison is built. The factors that contributed to differences in prison construction costs are discussedin appendix I. CJMUGGD-92-78 Primon Cht Factma Page 9 ,:; a B=248117 Changesin Prison Design Stand&s W iU Affect Future Construction Costs For comparative purposes our cost per bed analyseswere based on a common baseline-the number of inmates the facilities were actually designedto accommodate(referred to as the design capacity or the rated capacity) as reported by the participating jurisdictions. The prisons we sampled were built with design standards that called for housing one inmate in a single cell or two or more inmates in multiple occupancy cells or dorm itories. BOP has recently adopted a lim ited double celling standard (two inmates per cell) for the design of medium security prisons. The new standard allows for double celling in up to 60 percent of cells having 76 or more square feet. This change also increased the rated capacity of existing nor facilities that met the cell size criterion. In practice, BoPfacilities have been double celled extensively for some time and without unmanageable problems. Prison design standards are being revised at the state level as well. In August 1991,the ACA revised its accreditation standards for medium security facilities to perm it double celling and reduced the required space in multiple occupancy and dorm itory housing areas. Some states will likely revise their rated capacities based on the new ACA standards. Further, in January 1992,Attorney General W illiam P. Barr announced an effort to help states lift some cour%orderedprison population ceilings. These are believed by some to unreasonablylim it the number of inmates that may be housed in a prison. To the extent that the new standards increase rated capacity, new prisons that incorporate the new standards will have lower per bed construction costs. Nevertheless,we believe that the factors that affected prison construction costs at the prisons we sampled will continue to signMcantly affect construction costa after the revisions are fully implemented.That is, prison construction costs will continue to be driven in large measureby the amount of spaceprovided to inmates (GSF per inmate), the type of building structure, the housing area design and layout, whether the facility was designedfor a m ix of security levels, and geographic location. PersonnelExpenses Accounted for the Majority of ” OperationsCosts Operations costs also varied widely at the 23 prisons (21 state, 2 federal) that provided operations cost information. Operations costs ranged from $22.26to $81.08per inmate day and averaged$41.93(see fig. 3). The low cost prisons averaged$32.37per inmate day, compared to $46.83for the medium cost prisons and $62.81for the high cost prisons. The single Page 9 GAOJGGD-92.78 FMaon coat Fu?tora B-242117 largest operational expensewas personnel compensation+&aries and related expenses.Personnelcosts ranged from 66 to 93 percent of total operations costs and averaged76 percent. Figure 3: Avongo Opmtlonr Co& p& Inmtio Day _ 100.00 Cc4 In dollrn PO.00 60.00 70.00 60.00 60.00 40.00 -------- 30.00 2o.ao 10.00 0 LoW OOSt p&on0 ----- Medium - High cwt oost prlaonr prlrons Weighted average coat for 23 prieons, $41.83 An important factor in accounting for differences in personnel costs is the staff@ levels of a prison relative to its inmate population (the inmate-to-staff ratio). The prisons that employed more staff relative to a their inmate populations (i.e., those with lower inmate-tc-staff ratios) tended to incur higher personnel costs-and, consequently,higher operations costs. The low cost prisons reported an averageinmate&Maff ratio of 3.13 to 1, compared to 2.71 to 1 for the medium cost group and 1.76 to 1 for the high cost group. Figure 4 shows that as the inmate-tu-staff ratio increases,personnel costs per inmate day decrease. Page 10 GAOAXD-92-72 Prbon Coat Factors B-248117 Flguro 4: Comparhn Between the Inmate-tctWafi Ratio and Pwwonnel Co&to per Inmate Day Inmate-to-atatf ratio Pononnol coata par lnmrtl dry 66.0 62.0 46.0 330 3.26 3.00 44.0 40.0 2.75 2.50 36.0 2.26 32.0 2.00 1.76 1.60 20.0 1.25 16.0 1 .oo 0.76 12.0 0.60 0.26 0 6.0 4.0 0 tort prisons prlronr Amounts Spent for Supplies,Materials, and Food Contributed to Operations Cost Variances Conclusions High cost prisons Modlum Low COIL - Pemonnrl -- Inmate-to-rtalf cortt:prr inmate day ratio Other important factors that contributed to differences in operations costs were expensea for supplies, materials, and food. Although there were notable differences in the amounta spent by individual prisons, the low cost prisons spent an average of $4.76 per inmate day for supplies, materials, and food, compared to $6.24 at the medium cost prisons and $7.22 at the high cost prisons. The Eactorsthat contributed to differences in operations costs are discussed in appendix II. At the 96 medium security prisons included in our sample, per bed construction costs varied widely, ranging from $11,243to $93,333.The amount of space provided, measured in terms of GSF per inmate, accounted for 96 percent of the variability in per bed construction costs. Other factors that might have contributed to the differences were the type of building structure, the housing area design and layout, whether the facility was designed for a mix of security levels, and geographic location. Page11 GAWGGD92-78PrbonCortFa&om B-248117 However, these did not have a sign&ant additional effect when considered in combination with the amount of space provided to inmates. Operations costs also varied at the 23 prisons that provided operations cost information. Per inmate day operations costs ranged from $22.26to $81.08.The factors that contributed to the difYerencesin operations costs were personnel salaries and related expenses,inmate-to-stsff ratios, and the costs of supplies, materials, and food. Through better understanding of the reasonsfor cost differences in various prisons, jurisdictions concerned about the high costs of building and operating prisons can consider less costly alternatives. In designing new prisons, significant economies can be realized by providing less GSF per inmate (consistent with acceptable standards), using lower cost building types, making greater use of dorm itories and multiple occupancy cells in place of single cells, and, for somejurisdictions, selecting lower cost geographic locations. Similarly, designing new prisons to operate with greater inmate-to-staff ratios where appropriate can help hold down personnel costs-the single largest operations cost at a prison. AgencyComments and Our Evaluation We discussedthe contents of this report with JSOPofTiciah, who have overall responsibility for prison construction. They generally agreed with the facts presented. BOP officials informed us that its new design standard for cells in medium security prisons is 76 square feet, a reduction from the 90 squarefeet required under the old standard. This change is expected to be incorporated into BOP’S ofEicial policy guidelines in the near future. No change is anticipated to EIOP’S policy of assumingthat 60 percent of the cells will be double occupancy for purposes of calculating rated capacity. At the Suggestionof BOP officials, we includedthis information in our report, but the revised design standards did not affect our analysis of construction costs for existing facilities that we sampled. We also discussedthe contents of the report with an official of the ACA. He stated that the report presented important information that will be very useful to prison planners. In addition, he suggestedseveral factors that contribute to differences in prison costs. He stated that the intended inmate population, the m ission of the facility, climate, local building codes, and whether the prisons are in heavily unionized or right-to-work state3 can all affect prison construction and/or operations costs. P-e 12 GAWGGD-92-78 Prbon Chat Factora 0 B-248117 In doing our work we took into account most of the factors described by the ACAoffkial as they affected construction costs. For example, the national construction estimator index, used in our analysti of construction costa, was baaed on actual nationwide construction coats and thus accounted for differences in climates, wage rates, and other construction cost variables. Also, in developing our selection criteria, we excluded prisons designedfor less than 200 inmates and prisons with special m issions becausewe wanted to make prisons in our sample comparable and reduce cost distortions. Unless you announce the contents of this report earlier, we plan no further distribution of it until 30 days from ita issue date. We will then make copies available to the Attorney General,the Director of BOP, the states that participated in our study, and other interested parties. Copies will be made available to others upon request. The mqjor contributors to this report are listed in appendix V. If you have any questions about this report, please call me at (202) 6664026. Harold A. Valentine Associate Director, Administration of Justice Issues P4e18 Contents Letter Appendix1 FactorsThat Contributedto DifferencesinPrison Construction Costs Appendix11 FactorsThat Contributed to DifferencesinPrison OperationsCosts AppendixIII The32StatePrisons Submitting QuestionnairesUsed inAnalysisof ConstructionCosts AppendixIV The21StatePrisons Submitting QuestionnairesUsed in Analysis of Operations Costs AppendixV Major Contributorsto ThisReport 16 16 18 21 24 26 Amount of Space Provided to Inmates TSpeof Structure Design and Layout of Housing Units Mix of Secwlty Levels Geographical Location 29 31 Differences in Personnel Costa Inmate-t&StaffRatios Supplies and Services 34 36 38 39 40 GAWGGD-92-78 P&on Page 14 : Coot Factme content4 Tables Figures Table 1.1:Costs per Bed Comparedto Gross SquareFeet per Inmate Table 1.2:Per Bed Construction Costs by Type of Structure Table 1.3:Housing ConfiguxMon Table 1.4:Distribution of Beda by Security he1 Table 1.5:GeographicalAreas and National Construction Estimator Index Comparedto Costs per Bed and Coats per Gross SquareFoot Table II.1: AverageDaily Costa of Operations per Inmate Table 11.2:PersonnelCosts Table 11.3:PersonnelCosts per Inmate Day Table 11.4:Inu~te-t~-Staf’fRatios Table II.6: Daily per Inmate Operational ExpenseaOther Than PersonnelCosts Figure 1: AveragePrison Construction Costaper Bed Figure 2: ComparisonBetween Averageper Bed Construction Costa and Gross SquareFeet per Inmate in State and Federal Prisons Figure 3: Average Operations Costs per Inmate Day Figure 4: ComparisonBetween the Inmate-to-Staff Ratio and PersonnelCosts per Inmate Day Figure 1.1:Building Con@urations Figure 1.2:Typical Prison Housing Layouts Abbreviations ACA BOP FCI GSF American Correctional Association Bureau of Prisons Federal Correctional Institution Gross SquareFoot 17 20 23 26 27 30 32 33 36 36 6 7 10 11 19 22 Appendix I Factors That Contributed to Differences in Prison Construction Costs Officials managingthe acquisition of a new prison can directly influence its cost through their control over the design of the facility and, to some extent, where the facility is built. The 36 prisons (32 state, 4 federal) that participated in our study reported a wide range in construction costs. Total construction costs ranged from a low of $6,404,644(rated capacity 312) to a high of $266$66,796(rated capacity 2,916).Per bed construction costs ranged from $11,243to $93,333-more than an eight-fold difference. This section will provide some insights into the factors that contributed to these differences. To facilitate our analysis of the factors that affected construction costs, we divided the prisons into three cost groups-low, medium, and high. Natural breakpoints existed between the low and medium groups and between the medium and high groups. Table I.1 shows the prisons that comprise each cost group and arrays the prisons in ascending order by cost per bed. This same ascendingorder will be used for the other tables presented in this appendix. Where appropriate, the tables also include totals, weighted averages,’and medians for each cost group and for all 36 prisons. Amount of Space Provided to Inmates The most important factor contributing to differences in prison construction costs per bed was the amount of space provided, measured in terms of gross square feet (GSF) per inmate. Our analysis showed that 96 percent of the variability in the cost per bed was due to the amount of space provided. Table I.1 shows that as the amount of space provided per inmate increases,the per bed costs of the prisons also tend to rise. This increase in costs is especially dramatic when the lowest and highest cost groups are considered, with the cost per bed of the high cost prisons averagingalmost five times as much as the low cost group ($87,271vs. $17,730).The relationship between space and cost is quite striking for these cost groups, with the high cost prisons providing an averageof 664 GSF per inmate, over two and one-half times the averageof 216 GSF per inmate provided at the low cost prisons. ‘To compute the weighted averqjes, the value of each item to be avera8ed (cost per bed, for example) wee multiplIed by its weight (dee@ capacity) and the total of these products divided by the sum of the weight8 (aggregate design capacity for all 86 prison@. Source for wei@ed average formula: Fundamental Statistics for Bueineee and Economics, Third Edition, by John Neter and William Wawem (Boston: Allyn and B Page 16 GAWGGD-92-78 Pr&on Ckmt Factorm 0 Tablo I.1: Cooto par Bad Comparod to Qroor Squaw Foe1par Inmato Prleon name Low coot prhonr state Varner Calhoun Chlppewa McCormick Evans Allendale AR FL MI SC SC SC Craggy NC Winslow Weiahted averaaes AZ Coat par bed Grow quan foot nor Inmato $11,243 13,825 15,625 19,006 19,370 20.277 . 20,720 24,679 $17.730 194 219 185 220 220 220 220 251 215 $45,007 45,424 45,920 47.289 48,793 49,966 50.824 54,206 56,460 58,702 59,013 59,386 62,092 63,411 64,107 64,980 65,517 67,006 67,446 70,188 73,438 $58.282 411 423 260 615 447 597 460 521 413 414 402 540 481 431 588 355 507 404 671 562 627 475 Medium tort prlrone Danville Hill Lorton Avovelles Illinois River FCI Phoenix (BOP) Western Illinois Frackvllle Dayton Arkansas Valley Ross Smithfield Carson City Chuckawalla Correctional Complex Cayuga EC. Brooks Riverfront FCI Marlanna (BOP) EIY FCI Sheridan (BOP) Weiahted averaaes IL IL DC LA IL AZ IL PA OH co OH PA Ml CA IN NY Ml NJ FL NV OR (continued) Page 17 GAO/GGD92-78 Prioon Coat Factarm . &w* 1 Fmtoro!llWCoatrlbute~dtoDlffemncah Prtaonconatnsctloncow . Prlron name’ High coat prlaonr state Northern Old Colony FCI McKean (BOP) Corcoran Mule Creek Eastern Oshkosh Weighted averages NJ MA PA CA CA KY WI Welahted averiwte). 36 Drl#onr Cart per bed Grorr l quaro feet Dar Inmate $83,771 85,203 85,391 87,814 66.277 88,577 93,333 $67,271 556.374 389 565 670 524 624 634 619 554 435 Qtly the “short name”that diStingUlShe8 each facility from others in the same jurisdiction was used in the tables. For example, Arizona State Prison Complex-Winslow is shown as Winslow, and Pennsylvania’sState CorrectionalInstitutionat Frackville is shown as Frackville.Also, the 36 responding prisons are listed in ascending order of contruction costs per bed. The order is retained in the subsequenttables in appendix I. Another factor that we examined in terms of per bed construction costs was the type of building structure. The National Directory of Corrections Construction, published by the National Institute of Justice (April 1988), classified prisons into the following general types (see fig. I.1 for Type of Structure illustration): l l l l l l l integrated structure-one building; building, more than four stories in height; ladder; telephone pole-linear cell blocks arranged parallel to one another off a central connecting corridor; wheel, spoke, or radial-linear cell blocks connected to one central control area like spokes from the hub of a wheel; courtyard-linear cell blocks interconnected around a central enclosed courtyard; clusters-a number of individual buildings that are interconnected, and campus style-a number of individual buildings that are not interconnected. high rise-one Y Page 18 GAWWD-W-78 Prbon Cad Factma flgun I.1: Bulldlng Conflgumtlonr Campus 0 Ladder, telephone I cl 0 I pole Wheel, spoke or radial L I I Clusters Courtyard Page 19 UWGGtD-92-78 Prhon Cast Factors Appentus I hctom That ConMbuted to Dtmreneaa Prtmm coMtrQedon casta tn According to the questionnaires,all of the prisons in &r sample were either clusters, campus style, or integrated structures. Although per bed construction costs varied for each design style becauseof factors such as size and housing layout, the single building and cluster styles tended to be more costly than the campus style. As table I.2 shows, although per bed construction costs varied for each type of structure, integrated structures, on average,were the most costly of the three types of structures, followed by clusters and campus style. Of the seven high cost prisons in our sample, five were single buildings or clusters. In .contrast, seven of the eight low cost prisons were campusstyle. Tablo 1.2:Pot Sod Conrtructlon Coats by Typo of Structure Prlron name Low tort prlsonr State Varner Calhoun Chippewa McCormick Evans Allendale AR FL MI SC SC SC Cwxw NC Winslow Weiclhted averages AZ Co& by type of structure Single q Campur bulldlng rtylo Cluatwr $11,243 $13,825 15,625 19,006 19,370 20,277 20,720 24,679 $18,986 $11,243 Medium co&t prloonr IL IL DC LA IL AZ IL PA OH co OH PA Ml CA Danville Hill Lorton Avovelles Illinois River FCI Phoenix (SOP) Western Illinois Frackville Dayton Arkansas Vallev Ross Smithfield Carson Citv Chuckawalla $45,007 45,424 $45,920 1, $47,289 48,793 49,966 50,824 54,206 56,460 58,702 59,013 59,386 62,092 63,411 (continued) GAO/GGD-92-78 Primn Co& Factam Page 20 ,^ ,‘I’ ; Prloon name Correctional Complex Cayuga EC. Brooks Riverfront FCI Marianna (BOP) EIY FCI Sherldan (BOP) Weiahted averaaes St& IN NY Ml NJ FL NV OR Coat8 by type of atructuro Slnglo Camp448 bulldlng Cluoton 8W 64,107 64,980 65,517 67,006 67.446 70,188 73,436 $61.834 $61,140 $53,445 High coot prlronr Northern Old Colonv FCI McKban (BOP) Corcoran Mule Creek Eastern Oshkosh Welahted averaaes Wolghtrd averages, 96 prlronr Design and Layout of Housing Units NJ MA PA CA CA KY WI $83,771 85.203 $85,391 $87,814 66,277 88,577 wm $84,187 $86,781 $68,311 $73,55!5 $47,129 $64,012 The design and layout of the housing units is another important factor affecting prison construction costs. Table I.3 shows that prisons with higher percentages of cells designed to accommodate a single inmate tend to cost more to build than prisons designed with multiple occupancy cells and dormitories. For example, only about 4 percent of the beds in the low cost prison group are in single cells, compared to about 72 percent for the medium cost prisons and 99 percent for the high cost prisons. In contrast, about 96 percent of the beds in the low cost prisons are either in multiple occupancy cells or dormitories, compared to about 29 percent in the medium cost prisons and 11 percent in the high cost prisons. Figure 1.2 illustrates typical housing layouts as examples of how prison designs csn differ. Page 21 GAOK+GD-92-78 Pttmn Coat Factmu I, Faetzm That Ckmtrtbutedw D@raneamin Pri8on con8tllIedon corta Pigun 1.2:Typltal Prison Housing Layout8 Linear, with Outside Cells Linear, with Inside Cells Dormitory Module/Pod Page 22 MO&GD-92-72 Prbon Coat Factor6 Facton That Contibuted to DlffereneerIn PllaoncanatnlcttonCosta ~ In August 1991,the Federal Bureau of Prisons (EIOP) adopted a limited double celling design standard (two inmates per cell), but the design capacities of the four medium security federal correctional institutions (RI) included in our review were based on a single celling standard in effect when the information was provided. Therefore, the percentagesin the “Single cell” column of table I.3 would be expected to be 100 for each of the federal prisons. However, three of these projects included an adjacent minimum security camp which housed inmates in dormitories.2 For the three rc~~,BOP was unable to separate the construction costs of the medium security prisons from the minimum security camps. Consequently, we showed the prisons and the camps as single units, resulting in the housing configuration percentagesshown in table 1.3. Table 1.3:Housing Conflguratlon (Design) Prlron name Low cost prisons Total bed8 (rated State capaclty) Varner Calhoun Chippewa McCormick Evans Aliendale AR FL MI SC SC SC fWxiy NC Winslow AZ TotaWwelghted averager Cart per bed Slngle cells Beds Percentage Multiple occupancy cells Beds Percentage Dormltorlw Beds Percentage 1,100 768 640 1,104 1,104 1,104 312 650 6,782 $11,243 13,825 15,625 19,006 19,370 20,277 20,720 24,679 $17,730 0 0 0 96 96 96 0 0 268 0 0 0 9 9 9 0 0 4 0 0 640 1,008 1,006 1,008 0 400 4,064 0 0 100 91 91 91 0 62 60 1,100 768 0 0 0 0 312 250 2,430 100 100 0 0 0 0 100 38 36 896 896 400 610 787 518 728 504 498 724 $45,007 45,424 45,920 47,289 48,793 49,966 50,824 54,206 56,460 58,702 896 896 192 78 787 518 726 504 498 724 loo 100 48 13 100 100 100 106 100 100 0 0 0 52 0 0 0 0 0 0 0 0 0 9 0 0 0 0 0 0 0 0 208 480 0 0 0 0 0 0 0 0 52 79 0 0 0 0 0 0 (continued) Medium cost prlsons Danville Hill Lorton Avoyelles Illinois Piver FCI Phoenix (BOP) Western Illinois Frackville Dayton Arkansas Valley IL IL DC LA IL AZ IL PA OH co ?he minimum security camp a&went to FCI Phoenix was built as a separate construction project GAWGGD-92-72 Prison Coat Factan Page 22 .,, 1, Prleon name wet. I3089 Smlthfleid Carson City Chuckawalla Correctional Complex Cayupa EC. Brooks Rlverfront FCI Marlanna (BOP) w FCI Sheridan (BOP) OH PA MI CA IN NY Ml NJ FL NV OR TotaWkel~hted . l vemgw Total bed@ (rated capaolty) 1,268 448 612 2,ooo 716 756 580 462 698 476 752 Yultlplo~~~pancy Sing10 oelle Bedr Peroontago Coat pr bed 11,319 59,013 59,386 62.092 63,411 64,107 64,980 65,517 67,006 67,446 70,168 73,438 $58,282 1,047 428 646 2,916 1,700 500 300 7,537 29,638 1,008 448 612 0 716 0 580 462 550 290 496 Bodr Percontago 0 0 0 1,992 0 0 0 0 0 186 0 2,230 0 0 0 ~100 0 0 0 0 0 39 0 10,983 80 100 100 0 100 0 1’00 100 79 61 66 72 $83,771 85,203 85,391 07,014 88,277 88,577 93,333 $87,271 1,007 428 496 2,524 1,500 500 300 6,755 96 100 77 67 88 loo loo 90 $56,374 18,026 61 Dormltorlee Bedr Percentage 250 0 0 8 0 756 0 0 148 0 266 20 0 0 0 0 100 0 0 21 0 34 lb 2,106 14 40 0 0 0 0 0 0 40 4 0 0 0 0 0 0 0 0 23 13 12 0 0 1 0 0 150 392 200 0 0 742 10 6,334 21 5,270 18 High ooat prloonm Northern Old Colony FCI I&Keen (6OP) Corcoran Mule Creek Eastern Oshkosh NJ MA PA CA CA KY WI TotaWwelghtod avenger Td$Iw$$ded P l voragee, Note: Percentages may add to more than 100 due to rounding. lklix of Security Levels Although each of the prisons in our sample was designed to house a predominantly medium security population, some were also designed to accommodate minimum security and/or maximum security inmatea 83s well. Our analysis found that building a prison to accommodate a mix of security levels tended to add to construction costs. The prisons in table I.4 are listed in ascending order of construction costs per bed. The tab!e shows that construction costa per bed tended to increase as the percentage of medium security beds declined. Overall, the GAMWD-92-78 Page 24 ,. >’ ,( “ /,’ Prlaon Cwt Factarm ‘, prisor~ classified 76 percent of their beds as medium security, compared to 89 percent for the medium cost prisons and 00 percent for the low cost prisons. high cost Tablo 1.4:Dlatrlbutlon of Bedr by Socurlty Lwol Tot81 hod, Mlnlmum (rated State capacity) Bodr Porcentaao Prleon nam3 Low tort prlrono Varner Calhoun Chiopewa McCormick Evans Allendale Craoav AR FL MI SC SC SC NC Winslow AZ Total@/ Pclrcentaaoo 1,100 768 640 1,104 1,104 1,104 312 650 400 6.782 400 36 6 Numkr of bode at oath wourlty Iovol Medium Maximum Beds Porcontaao Bodr Porwntaae 700 768 640 1,608 1,008 1,008 312 650 64 100 100 91 91 91 100 100 6.064 90 896 896 208 610 787 518 728 504 466 724 944 448 360 1,992 716 756 360 441 496 100 100 52 100 100 100 100 100 94 100 75 100 59 loo loo loo 62 95 71 96 98 Q6 s 9 9 266 4 192 48 Other Bode Porwntaao Medium oort prlronr Danvilie Hill Lorton Avovelles Illinois River FCI Phoenix (BOP) Western llllnols Frackvllle Dayton Arkansas Valley Ross Smlthfleld Carson’Citv Chuckawalla Correctional Complex Cayuga EC. Brooks Riverfront FCI Marianna (BOP) * IL IL DC LA IL AZ IL PA OH co OH PA Ml CA IN NY Ml NJ FL 896 896 400 610 787 518 728 504 498 724 1,258 448 612 2,000 716 756 580 462 698 250 20 60 8 10 0 60 10 148 21 192 31 160 28 54 32 6 64 5 21 5 8 (continued) Page 26 ,. ApQendtx I Fwtor~ That Contributed to DlfPereneer in Prloon CoMtnledon co&8 Total beda M lnlmum (rated State capacity) Sedr Percentage NV 478 OR 752 266 34 Prlron name@ EIY FCI Sheridan (BOP) Total@/ prcontagor 18,319 782 1,047 40 Numkr of bodr at aach aecurlty level Medium Maxlmum Other Bed8 Percentage Bedo Percentage Bedo Percentage 266 60 190 40 496 66 5 13,632 89 788 5 117 1 92 30 3 17 2 51 1,024 35 High coot prloonr Northern Old Colony FCI McKean (BOP) Corcoran Mule Creek Eastern Oshkosh NJ MA PA CA CA KY WI Total@/ percantagrr Total8/pwcontager, 36 prlrona 428 646 4 150 23 960 428 496 2,916 1.700 392 200 13 12 1,500 1,500 500 300 500 300 7,537 782 29,638 1,964 10 loo 77 88 loo loo 5,684 75 1,054 14 17 0 7 25,410 86 2,130 7 134 0 ‘As pointed out previously,the prisons are arranged in order from lowest constructioncost per bed to highest constructioncost per bed. See table I.1 for specific cost per bed Informatlon. Geographical Lwation Y Prison construction costs can also be affected by geographic location. According to the National Institute of Justice and the American Correctional Association (ACA), construction costs can vary from one part of the country to another due to sharp contrasts in the cost of materials and prevailing labor rates. For example, according to the National Construction Estimator indexes for m id-1989,construction costs tended to be higher in the Northeast and West and lower in the South and M idwest. The prisons in our sample generally reflected those tendencies. Table 1.6 shows that of the 8 low cost prisons, 6 were in the South, while only 3 of the 21 medium cost and 1 of the 7 high cost prisons were in the South. Conversely,there were no Northeast prisons in the low cost group, while four Northeast prisons were in the medium cost group and three in the high cost group. Severalcompaniespublish construction cost indexes that allow cost estimators to adjust for regional differences in the costs of labor, material, and equipment.Table I.6 shows the 1989“National Construction Phge 26 GAWGGD-92-78 Primon Co& Fmtorm Elstimator”index for each of the states in which the sample of state prisons were located. This index allows interested parties to make cross jurisdictional comparisons of construction ~osts.~Our analysis showed that as the estimator index for the state prisons in our sample increased,the cost per bed also tended to increase.The median index for the low cost prisons is 31, compared to 1.06for the medium cost group and 1.17for the high cost group. Another indication of the importance of geographic location is its effect on cost per GsF.The cost per GsFis, in effect, the measure of the amount of space the jurisdiction was able to buy for its money, independent of the number of inmates the prison was designedto house. Table I.6 shows that the cost per GSF at the 36 prisons ranged from $68.06to $216.60and averaged$129.48.Further, the table shows that as costs per GSF increased, costs per bed also tended to increase. We found that about 96 percent of the variability in cost per GSF was explained by the national construction estimator index, which is a surrogate measure for the state in which the prison is built. Table 1.5:Geogmphlcal Areas and Natlonsl Constructlon Estlmator Index Compared to Costs per Bed and Costs per Gross Square Foot Prison name State Costs per bed U.S. Region Index Costs per GSF Low co8t prlrons Vamer Calhoun Chlppewa McCormick Evans Allendale AR FL MI SC SC SC Cww NC Winslow Weighted average Median index AZ $11,243 13,825 15,625 19,006 19,370 20,277 20,720 24,679 $17,730 South South Midwest South South South South west .83 -90 39 .78 -78 .78 .79 1.01 $58.06 63.22 84.45 86.48 88.14 92.27 94.03 98.20 $82.40 .81 Medlum:cost prlrons Danville Hill IL IL ” $45,007 45,424 Midwest Midwest 1,06 1.06 $109.63 107.31 (continued) The following example Illustrates how the estimator Index works. The Correctional Industrial Complex in Indiana coat about $46,900,000.If the aame prlaon had been built in California in the same year, the Index indicates it would have coat about $60,369,0oo(1.211.92 = 1.315 x ~6,900,000 = $60,363,600).On the other hand, lf the aame facility had been build in South Carolina, the index indicatea it would have cost about $33,900,000(.78 / .92 = .S43 x $46,900,000= $33,923,000). Page 27 GAO/GOD-92-73 Prison Coat Factora l Prbon name St& Lorton Avwelles Illinois River FCI Phoenix (BOP) Western llllnols Frackville Dayton Arkansas Valley DC’ IA IL AZ IL PA OH co OH PA MI CA IN Ross Smlthfleld Carson City Chuckawalla Correctional Complex Cayuw NY EC. Brooks Riverfront FCI Marlanna (BOP) Ml NJ FL NV OR EIY FCI Sheridan (BOP) Welghted average Median index Coot8 per bed 45,920 47.289 48,793 49,966 50,824 54,206 56,460 58,702 59,013 59,386 62,092 63.411 84,107 64,980 65,517 67,006 67,446 70,188 73,438 $58,282 U.S. Region South South Midwest west Midwest Northeast Midwest West Midwest Northeast Midwest West Midwest Northeast Midwest Northeast South West West Index Co&r per GSF .92 .a5 1.06 1 .Ol 1.06 1.07 1.10 1.07 1.10 1.07 99 1.21 -92 1.13 99 1.17 90 1.21 1.04 176.35 76.93 109.12 83.73 110.38 104.06 136.86 141.67 146.85 110.01 129.17 146.96 109.09 183.21 129.17 165.81 10059 124.78 117.05 $122.77 1.06 High oort prlronr, Northern Old Colony FCI McKean (BOP) Corcoran Mule Creek Eastern Oshkosh Welphted average Median index NJ MA PA CA CA KY WI $83,771 85,203 85,391 87,814 88,277 88,577 93,333 $87,271 Northeast Northeast Northeast west West South Midwest 1.17 1.19 1.07 1.21 1.21 .91 1 .Ol $21550 150.69 127.42 167.62 141.56 139.71 150.66 $157.64 1.17 Woightod average, 36 prlsono Mddlrn Index, 36 prlwn8 $56,374 $129.48 1.06 The District of Columbia’s Lorton facility is located in suburban Virginla. Page 28 WGGD-92-72 Primon ihta and I&tom . rs That Contributed to Differences in Prison Operations Costs Operationscostaat our sampleprisons varied signiilcsntly, although not to the extent of the differences in construction costs discussed in appendix I. At our sample of 23 prisons (2 federal, 21 state), operations costs per inmate dsy ranged from $22.26to $81.98,with a weighted average of $41.93. The following tables will show that the operations cost differences were due mostly to differences in salaries and related expenses,staffing levels relative to inmate population, and amounts paid for supplies, materials, food, and services. Because operationscostavaried so widely, we divided the prisons in our sample into three cost groups for analysis purposes- low, medium, and high. We used breakpoints that existed in the daily operations costs per inmate to define the three cost groups. We defined prisons with daily operations cost per inmate in the $22 to $37 range as low cost, those in the $42 to $61 range as medium cost, and those in the $69 to $81 range ss high cost. Table II.1 shows the prisons in each cost group, listed in ascending order by average daily cost’per inmate. There wss no direct relationship between the cost groups described in this section and the construction cost groups discussed in appendix I. Page 29 G,UM3GD-S2-78 Rbon Coat Factora Factmu That Contrtbnted Priwn Op4mttonm Coota to Diibmmcea in Teblo 11.1:Avoraao Dally Coetr of Oueretlonr per Inmato Prlron name Low coat Prleonr Ross State Annual l xpeneer McCormick Chlppewa Calhoun Frackville FCI Marianna (BOP) OH SC MI FL PA FL $13,709,314 9,184,304 9,933,378 9,363,992 11,762,OOO 13,230,154 Crams NC Hutchinson Arkansas Valley FCI Phoenix (BOP) Hill Danville KS co AZ IL IL $129,904,910 lotalo/Melghted average, Avenge dally population 1,688 1,075 946 Costs per inmate Per year Costs per Inmate per day $8,122 $22.25 8,544 23.41 28.77 900 1,000 10,500 11,793 13,069 13,230 36.25 30844.809 288 13.350 36.58 5,372,376~ 400 13,431 36.80 12,694,OOl 14,781,482 12.947.700 13,081,400 935 1,078 944 946 13,576 13,712 13.716 37.20 13,828 10,994 $11,818 $32.37 470 $15,374 $42.12 300 15,460 42.36 794 32.31 35.81 37.57 37.58 37.89 Medium coot prlronr Dayton Al Burruss Mule Creek Smithfield Corcoran Oshkosh Old Colony Cayuna OH GA CA PA CA WI MA NY $7,225,893 4,637,974 50.020.790 7,332,OOO 82,538,576 7,141,779 10,787,163 17,651,991 Totalsiwetlghtedaverage8 $187,33&l 66’ 3.204 15.612 42.77 450 16,293 44.64 4,838 17,060 46.74 399 17,899 49.04 589 950 18,314 18,581 50.18 50.91 11,200 $16,726 $45.93 6 Hlah coat prloone Eastern Northern Riverfront MD NJ NJ TotaWwelahtod avoraclee ToteWwelghted avrrager, 23 prlronr $31,189,074 22,461,OOO 13,QO8,500 1,440 1,037 $21,659 21,660 59.34 470 29,593 81.08 S 87.558.574 2.947 $22.925 $82.81 $394,799,650 25,141 $16,309 $41.93 $59.34 % this table, the 23 responding prisons are listed in ascending order of the daily costs of operations per inmate.This order is used in the subsequenttables in appendix II. I Page 80 GAWGGD-92-78 Primon Co& Factora Appsndlx II Footon That Cantrlbutad Prhon oporatl0M cooto Differencesin PersonnelCosts to Milerenasr Lr The single largest expenseof operating a prison is the cost of personnel compensation.As table II.2 shows, personnel costs ranged from 66 percent to 93 percent of total operations costs, with an an overall averageof 76 percent. Page 31 GMUGGD-92-72 Prison Co& Factore Tabla 11.2:Pereonnol Coat0 Prlron nom0 Low cod prl8ono Ross McCormick Chlppewa Calhoun Frackviiie FCI Marianna (BOP) Craaav Hutchinson Arkansas Valley FCI Phoenix (BOP) Hill Danvilie stmto OH SC MI FL PA FL NC KS co AZ IL IL Totals/weighted averages Poraonnrl cost8 $10,293,742 6,163,171 6,360,376 6.659067 7,943,ooo 6,661,206 3,048,585 4,039,103 9,102,164 9,561,593 6.801 .ooo 8,824,900 $91,476,931 Poreonnol cortr a0 percentage of total cortr 75 67 64 71 68 65 79 75 72 65 66 67 70 Medium tort prlronr Dayton Al Burruss Mule Creek Smithfield Corcoran Oshkosh Old Colony Cavuga OH GA CA PA CA WI MA NY $5,559,613 3,790,863 36,301,226 5,275,OOO 64,682,355 5599,232 10,050,670 14,925,770 $14&l 94,749 TotalaIw~lghkd rvera&r 77 62 77 72 76 78 93 65 79 8 Hlah co81 prlrons Eastern Northern Riverfront Totala/welghted avoragw TotaNwolghtod avoragw, 23 prlaonr MD NJ NJ 2 50,127,160 71 75 80 74 $289,788,640 75 $22,129,160 16,687,OOO 11,111,OOO Differences in personnel costs did not account for all of the variances and were not always consistent with differences in overall costs. For exainple, table II.2 shows that personnel costs comprised 79 percent of total costs at Pago a2 GAOIGGD-92.72 Primon Coat Factora the medium cost prisons, compared to 74 percent at the high cost prisons. Table II.3, however, shows that daily personnel costs per inmate at the high cost prisons were about $10 higher than at the medium cost prisons due to staffing levels relative to inmate populations (see following section). Table 11.3:Pwoonnal Coti av pof Inmab Prloon nrmo Low coat prbno Ross McCormick Chippewa Calhoun Frackviiie FCI Marianna WP) Craaov Hutchinson Arkansas Valley FCI Phoenix (BOPI Hill Danvliie SUtO OH SC MI FL PA FL NC KS CO AZ IL IL TotalWelghtod avorager PoraonnoI cortr Avonga dally populrtlon Poraonnol coata per Inmato day $10,293,742 6,163,171 8,380,378 6658,087 7,943,ooo 1,688 1,075 946 794 900 $16.71 15.71 24.27 22.97 24.18 8661,208 3,046,565 4,039,103 9,102,164 1,000 286 400 935 23.73 29.00 27.67 26.67 9.561.593 8,801,OOO 8,824,900 1,078 944 946 24.30 25.54 25.56 8 91,476,931 10,994 822.88 $5559,613 3.790.883 38,301,226 5,275,OOO 64682,355 5,599,232 10,050,670 14,925,770 470 300 3,204 450 4,836 399 569 950 832.41 34.62 32.75 32.12 36.63 38.45 46.75 43.04 6146,164,749 11,200 836.25 Modlum coot prlronr Dayton Al Burruss Mule Creek Smithfield Corcoran Oshkosh Old Colony Cayuga TotaWwolghted avomger OH GA CA PA CA WI MA NY (continued) Page88 -D-82-78 Prhon ChtFactom 8 lkotmn ‘hat Contributed Prlum op8miOM co8t8 Prktn nom8 stata to DIlYorenw In Average dally population Pamonnel coat8 per lnmata day $22,129,160 16,887,OOO 11.111,000 !,440 1,037 $42.10 44.61 470 64.77 $60,127,100 2,947 $46.60 5289.788.540 25.141 S31.66 Pomonnal coat8 High coat wlaona Eastern Northern Rlverfront TotaWWalghtad average8 TotaWvelghted rvaragaa, 23 Drlaona Inmate-to-StaffRatios MD NJ NJ The stxifEi level of a prison relative to it9 inmate population (inmate-to-staff ratio) is an important factor in accounting for differences in personnel costs. For example, table II.4 shows that the low cost prisons reported an averageinmate-to-staff ratio of 3.13 to 1, compared to 2.71to 1 for the medium cost prisons and 1.76to 1 for the high cost prisons. This clearly shows that the prisons in our sample that employed more staff relative to their inmate populations (i.e., those with lower inmate-to-staff ratios) tended to incur higher operational costs. Page a4 6lAWGGD-92-72 Prbon Coat Factarm Tabla 11.4:Inmate-to-Staff Ratloa Prlaon name Low ooat prlaona Ross stat8 McCormick Chlppewa Calhoun Frackvilie FCI Marianna (BOP) OH SC MI FL PA FL Craggy NC Hutchinson Arkansas Valley FCI Phoenix (BOP) Hill Danvilie KS co AZ IL IL TotaWwalghted avaragaa Authorized atatf Avamga dally Inmate-topopulatlon staff rat108 377 342 217 266 253 295 132 173 301 322 404 428 1,666 1,075 946 794 900 l,ooO 266 400 935 1,078 944 946 4.48 3.14 4.36 2.98 3.56 3.39 2.18 2.31 3.11 3.35 2.34 2.21 3,510 10,994 3.13 220 157 870 267 1,582 188 370 457 470 300 3,204 450 4,838 399 569 950 2.14 1.91 3.68 1.57 3.66 2.12 1.59 2.08 4,131 11,200 2.71 791 566 323 1.82 1.83 1.46 1,682 1,440 1,037 470 2,947 9,323 25,141 Madlum coat prlaona Dayton Al Burruss Mule Creek Smithfield Corcoran Oshkosh Old Colonv Cayuga OH GA CA PA CA WI MA NY TotaWweightad average8 Hlgh coat prlaona Eastern Northern Riverfront TotaWwalghtad avaragea TotaWwalghtad averagea, 23 prlaona Supplies and Services MD NJ NJ most of the prisons in our sample was supplies, material, and food. Although there were notable differences in the amounts spent by Pqe a6 1.711 2.70 individual prisons, the medium and high cost prisons tended to spend more in this category. For example, table II.6 shows that the low cost prisons spent an averageof $4.76per inmate day for supplies, materials, and food, compared to $6.24at the medium cost prisons and $7.22at the high cost prisons. Table II.6 also shows that the high cost prisons spent more than the other cost groups on services.1Servicesat the high cost prisons amounted to $6.47per inmate day, compared to $2.62at the low cost prisons and $.86 at the medium cost prisons. Table II.5: Dally par Inmato OperatIonal Expenwa Other Than Petwonnol Coat8 Rent, Staff communlcatlonr~ Prloon nom0 travel utllltllt, Setvlcer St& SuptWo, miierlal; food Equlpmant Other Total $5.54 7.70 4.50 0.34 Law coat arlaena McCormick Chlppewa Calhoun Frackville FCI Marlanna (BOP) OH SC MI FL PA FL $0.01 0.02 0.09 0.10 0.05 0.43 Cww NC 0.06 Hutchinson Arkansas Valley FCI Phoenix (BOP) Hill Danvllle Weighted averages P@rcentages KS co AZ IL IL ROSS 3.42 0.00 2.02 6.12 2.01 1.74 0.60 0.39 4.22 4.71 4,66 $2.52 26 $3.72 2.57 3.74 5.06 3.63 6.32 4.14 5.95 7.97 6.06 4.32 4.45 $4.75 50 $0,05 0.03 0.10 0.15 0.10 1.04 0.02 0.05 0.43 0.64 0.19 0.30 $0.28 3 $0.00 0.05 0.00 0.58 0.03 0.07 0.00 0.00 0.09 0.16 0.66 0.74 $0.20 2 $0.56 0.69 0,.46 3.65 0.80 1.49 $4.92 4.74 5.35 6.03 5.46 6.36 $0.11 0.23 0.74 1.52 0.32 0.10 $1.96 0.00 1.03 0.02 1.32 0.11 $1.58 $0.16 0.14 0.03 0.45 0.04 0.04 $0.12 1 1.62 0.57 1.44 1.49 2.65 1.62 2.19 1.61 1.71 2.12 2.13 $1.71 18 $0.06 0.02 0.20 0.12 0.27 006 $2.09 2.06 2.24 1.19 1.94 2.45 11.63 12.52 7.57 9.13 10.52 13.27 12.03 12.33 $B.W 100 Medium coot prlronr Dayton Al Burruss Mule Creek Smlthfleld Corcoran Oshkosh u OH GA CA PA CA WI 5 9.71 7.74 10.02 12.52 10.11 10.59 (continued) %ervices include such expenses ae trash dispoeal, laundry and dry cleaning, repair and maintenance of equipment, and medical treatment from outside sources. QAWGGD.92.78 Priooa Co& Facton Page 86 ” 6 Prleon name, state Old Colonv Cavuaa - -. -wWeighted averages Percentages MA NY Staff _ .-.. tnvol Rant, communicntlona. __.~...~. ~~~~ ~, utllltler supplier, material. SOWICOO tooi Equipment 0.12 0.02 $0.19 2 0.21 1.09 $1.86 19 1,08 0.97 $0.86 9 1.60 5.48 $5.24 55 $0.17 0.02 0.00 $0.09 1 $2.57 0.85 1.22 $1.75 11 $7.42 5.31 6.13 $6.47 40 $0.15 $1.78 $2.24 Other Total 0816 0.15 $0.45 5 0.26 0.15 $0.98 10 3.43 7.86 $9.53 $6.05 8.10 8.85 $7.22 45 $0.28 0.41 0.03 $0.29 2 $0.75 0.03 0.07 $0.39 2 $17.24 $5.26 $0.35 $0,57 $10.35 100 Hlgh ooet prleonr Eastern Northern RIverfront Weiohted averaaes Percentages WzfLtz rvoragw, Parcantaaar MD NJ NJ 14.73 16.31 $16.21 100 23 1 Page 27 WGGD-92-78 Pr&n Costa and Pactmu Appendix JII The 32 State Prisons Submitting QuestionnairesUsed in Analysis of Construction Costs state Arizona Arkansas California Colorado District of Columbia Florida Illinois Indiana Kentucky Louisiana Massachusetts Michigan Nevada New Jersey New York North Carolina Ohio Pennsylvania South Carolina Wisconsin Page 813 Prlmn Arizona State Prison Complex - Winslow, Winslow Varner Unit, Grady California State Prison, Corcoran Chuckawalla Valley State Prison, Blythe Mule Creek State Prison. lone Arkansas Valley Correctional Facility, Crowlsy Modular Facility, Lorton, Virginia Calhoun Correctional Institution, Blountstown Danville Correctional Center, Danville Hill Correctional Center, Galesburg Illinois River Correctional Center, Canton Western Illinois Correctional Center, Mt. Sterling Correctional Industrial Complex, Pendleton Eastern Kentucky Correctional Complex, West Liberty Avoyelles Correctional Center, Cottenport Old Colonv Correctional Center. Bridnewater E. C Brooks Regional Facility, Muskegon Carson City Regional Facility, Carson City Chippewa Temporary Correctional Facility, Kincheloe Elv State Prison, Elv Northern State Prison, Newark Riverfront Correctional Facility, Camden Cayuga Correctional Facility, Moravia Craggy Correctional Center, Asheville Dayton Correctional Institution, Dayton Ross Correctional Institution, Chillicothe State Correctional Institution at Frackville, Frackville State Correctional Institution at Smithfield, Huntingdon McCormick Correctional Institution, McCormick Allendale Correctional Institution, Fairfax Evans Correctional Institution, Bennettsville Oshkosh Correctional Institution, Oshkosh GLWGGD-92-78 Prison Comt Facto- The 21 State ‘Prisons Submitting Questionnaires Used in Analysis of Operations Costs state California Colorado Florida Georaia lllinols Kansas Marvland Massachusetts Michigan New Jersey New York North Carolina Ohio Pennsylvania South Carolina Wisconsin Pype 2B Prkn California State Prison, Corcoran Mule Creek State Prison, lone Arkansas Valley Correctional Facility, Crowley Calhoun Correctional Institute, Blountstown Al Burross Correctional ‘Trainina Center, Forsvth Danville Correctional Center, Danville Hill Correctional Center, Galesburg Hutchinson Correctional Work Facility, Hutchinson Eastern Correctional Institution. Westover Old Colony Correctional Center, Bridgewater Chippewa Temporary Correctional Facility, Kincheloe Northern State Prison, Newark Riverfront Correctional Facility, Camden Cayuga Correctional Facility, Moravia Craggy Correctional Center, Ashville Dayton Correctional Institution, Dayton Ross Correctional Institution, Chillicothe State Correctional Institution at Frackville, Frackville State Correctional Institution at Smithfield, Huntingdon McCormick Correctional Institution. McCormick Oshkosh Correctional Institution. Oshkosh GAWGGD-92-72 Prhon Cost Factora AppendixV Major Contributors to This Report General Government Division, Washington, D.C. Rkhard M. Stana,AmWant Director,Administrationof JusticeIsauea JoanneParker,SeniorSocialScienceAnalyst BonnieSteller,SeniorStatitician Kim Wheeler,PublishingAdvisor Los Angeles Regional Office DannyM. Bullock, Evalua~r-in-charge (lanSl9) Page 40 Ordering Information The first copy of each GAO report and testimony is free. Additional copies are $2 each. Orders should be sent to the following address, accompanied by a check or money order made out to the Superintendent of Documents, when necessary. Orders for 100 or more copies to be mailed to a single address are discounted 25 percent. I J.S. General Accounting Office I’.(). 130x 6015 Gaithersburg, MD 20877 Orders may also be placed by calling (202) 2756241. _.. .“, .” “, United States General Accounting Office Wwhington, I).(?. 20548 Official Penalty _ Husiness for Private ,,“....”.“_.I.--- Fir-&Cl&s Mail Postage 8~ Fees Paid GAO Permit No. GlOO lJse $300 .._..._ .. - _ ___...-. .~ .-.“. .^-~ ---