Contents | Feedback | Search | DRCNet Home Page | Join DRCNet
DRCNet Library | Schaffer Library | Government Publications | GAO Publications
(Letter Report, 07/19/95, GAO/GGD-95-182). Pursuant to a congressional request, GAO examined the Defender Services program, focusing on: (1) the causes of the program's increased workload and costs; and (2) whether death penalty resource centers (DPRC) have reduced the federal costs of representing indigent defendants in death penalty cases. GAO found that: (1) the Defender Services program represents individual defendants rather than cases and during fiscal years (FY) 1990 through 1993 there was an average of 1.4 defendants per criminal case filed; (2) the Defender Services program workload has grown and costs have increased because the number of defendants have increased and the cases have become more complex, resulting in more work for attorneys in each case; (3) there was insufficient data to determine why overall program costs doubled and DPRC costs tripled between FY 1990 and FY 1993; (4) federal public defender and community defender organization representations cost less than panel attorney representations, although the cost advantage has gradually declined; and (5) death penalty representations accounted for 0.6 percent of all panel attorney representations closed in FY 1993, but accounted for 8.7 percent of all panel attorney payments. --------------------------- Indexing Terms ----------------------------- REPORTNUM: GGD-95-182 TITLE: Federal Criminal Justice: Cost of Providing Court-Appointed Attorneys Is Rising, but Causes Are Unclear DATE: 07/19/95 SUBJECT: Federal courts Court costs Public defenders Defendants Lawyers Crimes or offenses Criminal procedure Trials IDENTIFIER: Federal Defender Services Program ************************************************************************** * This file contains an ASCII representation of the text of a GAO * * report. Delineations within the text indicating chapter titles, * * headings, and bullets are preserved. Major divisions and subdivisions * * of the text, such as Chapters, Sections, and Appendixes, are * * identified by double and single lines. The numbers on the right end * * of these lines indicate the position of each of the subsections in the * * document outline. These numbers do NOT correspond with the page * * numbers of the printed product. * * * * No attempt has been made to display graphic images, although figure * * captions are reproduced. Tables are included, but may not resemble * * those in the printed version. * * * * A printed copy of this report may be obtained from the GAO Document * * Distribution Facility by calling (202) 512-6000, by faxing your * * request to (301) 258-4066, or by writing to P.O. Box 6015, * * Gaithersburg, MD 20884-6015. We are unable to accept electronic orders * * for printed documents at this time. * ************************************************************************** Cover ================================================================ COVER Report to the Ranking Minority Member, Subcommittee on Commerce, Justice, and State, the Judiciary, and Related Agencies, Senate Committee on Appropriations July 1995 FEDERAL CRIMINAL JUSTICE - COST OF PROVIDING COURT-APPOINTED ATTORNEYS IS RISING, BUT CAUSES ARE UNCLEAR GAO/GGD-95-182 Defender Services Program Abbreviations =============================================================== ABBREV AOUSC - Administrative Office of the U.S. Courts CDO - Community Defender Organization DPRC - Death Penalty Resource Center FDO - Federal Defender Organization FPD - Federal Public Defender (Organization) Letter =============================================================== LETTER B-259805 July 19, 1995 The Honorable Ernest F. Hollings Ranking Minority Member Subcommittee on Commerce, Justice, and State, the Judiciary, and Related Agencies Committee on Appropriations United States Senate Dear Senator Hollings: The Federal Defender Services program provides legal counsel for eligible federal criminal defendants who are unable to pay for these services themselves. Between fiscal years 1990 and 1993, the program's costs more than doubled from about $122 million to almost $252 million.\1 During this same period, the number of Defender Services representations closed rose 23 percent. In federal district courts the number of criminal cases filed rose less than 1 percent, and the number of defendants grew about 3 percent. At the same time, the costs to operate Death Penalty Resource Centers (DPRCs)--whose representations in federal courts are funded by grants from the Defender Services program--almost tripled from about $6 million to $18 million, and the number of DPRC representations more than doubled from 491 to 1,014.\2 These centers were created in part to provide a source of expertise and help to reduce or contain the costs of death penalty representations. You asked us to examine a number of aspects of the Defender Services program, including the causes of the program's increased workload\3 and costs (as measured by program budget obligations) and whether the DPRCs have reduced the federal costs of representing indigent defendants in death penalty cases. As we informed your office early in our evaluation, the Administrative Office of the U.S. Courts (AOUSC) did not have adequate data for us to evaluate the direct impact on program workload and costs of factors that AOUSC had suggested as possible causes of increases. Moreover, the lack of comparable data for both before and after the establishment of the DPRCs, including data on the impact of changes in the complexity of death penalty litigation, precluded meaningful assessment of whether the centers have reduced or contained the costs of representing indigent defendants in death penalty cases. Therefore, as agreed, our objectives were to provide information on (1) the reasons that Defender Services workload has grown faster than district court criminal cases; (2) the data available to assess the causes of increased Defender Services workload and costs; (3) the comparative costs of representations provided by federal defender organizations and by private attorneys directly assigned by federal judges; (4) the additional costs of paying higher standard hourly rates to private, court-appointed attorneys in all or parts of 16 districts; (5) the comparative costs of federally funded private attorney and DPRC death penalty representations; and (6) the potential causes of increased DPRC workload and costs. -------------------- \1 All dollar figures in this report are expressed in current dollars, unadjusted for inflation. See app. I. \2 The Judicial Conference Committee on Defender Services has recently changed the name of the DPRCs to Post-Conviction Defender Organizations. Throughout this report, we refer to them by their former designation, DPRCs. \3 Throughout this report, we measure program workload in terms of representations closed (see page 6 for more details). BACKGROUND ------------------------------------------------------------ Letter :1 The Criminal Justice Act of 1964\4 required the federal judiciary to provide for the legal representation of eligible federal criminal defendants who were financially unable to afford their own attorneys.\5 In response, the federal judiciary created the Federal Defender Services program. In August 1993, AOUSC reported that about 85 percent of all criminal cases prosecuted in federal courts required court- appointed legal counsel. The Defender Services Committee of the Judicial Conference of the United States\6 provides overall policy direction and guidance of the program, and the Defender Services Division of AOUSC provides administrative and program support. Legal services for eligible defendants in the nation's 94 federal district courts and 12 circuit courts of appeals are provided through a mixed system, which at the end of fiscal year 1994 included 45 Federal Public Defender Organizations (FPDs) serving 54 districts; 10 Community Defender Organizations (CDOs) serving 11 districts; and private "panel" attorneys chosen from a list, or panel, maintained by the district courts. Together, FPDs and CDOs are generally referred to as Federal Defender Organizations (FDOs). As shown in table 1, the number of FPDs and CDOs has gradually increased since 1990. FPD attorneys are federal employees; CDO attorneys are not. CDOs are private organizations funded by federal defender services grants. FPD and CDO attorneys are salaried; panel attorneys receive an hourly rate for their services that varies by district. Higher panel attorney hourly rates are generally paid for death penalty cases in all districts. Panel attorneys generally provided all court- appointed representations in the 29 districts that had neither a FPD or CDO, and by statute panel attorneys must receive a "substantial" number of representation appointments in districts with such organizations. Table 1 Number of Federal Public and Community Defender Organizations and Number of Districts Served, Fiscal Years 1990 to 1994 District District District Fiscal s s s year Number served\a Number served\a Number served\a ------ ------ -------- ------ -------- ------ -------- 1990 37 39 6 7 43 46 1991 38 40 6 7 44 47 1992 41 47 9 10 49 57 1993 42 47 9 10 51 57 1994 45 54 10 11 55 65 ------------------------------------------------------------ \a Some FPDs and CDOs serve more than one district. Source: AOUSC. To date, death penalty representations have generally been handled by panel attorneys (including some pro bono representations)\7 or the DPRCs. Occasionally, FDOs have also been appointed in such cases.\8 At the end of fiscal year 1994 there were 20 DPRCs serving 50 federal judicial districts. The DPRCs are specialized CDOs that provide legal services--through direct representation and/or consultation and support services to panel attorneys--for persons appealing state death penalty convictions and/or sentences. The DPRCs are also authorized to represent defendants in federal capital prosecutions and in appeals of federal death penalty convictions. Under the grant agreement between each center and the Judicial Conference, use of each DPRC's federal funds is to be limited to federal death penalty habeas corpus\9 cases and the defense of those charged with capital crimes in federal district courts. A DPRC's nonfederal funds must be used to support representations in state court proceedings. -------------------- \4 Public Law 88-455 (1964), codified at 18 U.S.C. 3006A. \5 In addition to financial requirements, eligibility for court- appointed counsel is also dependent upon the offense with which the defendant is charged. For example, defendants charged with certain misdemeanors are not provided court-appointed attorneys unless the U.S. Magistrate or the court determines that the interests of justice require it. \6 The Judicial Conference is a group of federal judges, chaired by the Chief Justice of the United States, who serve as the central policymaking body for the federal court system. \7 In pro bono cases, attorneys represent defendants without cost as a public service. \8 In late 1994, the Ninth Circuit Judicial Council approved a pilot plan to establish a death penalty unit in the FPD for the Central District of California to handle death penalty habeas corpus cases. The Council has sought Judicial Conference approval for this pilot project. \9 Federal habeas corpus proceedings are civil proceedings in which convicted persons raise a challenge in federal court to their conviction and/or sentence on the grounds that they are in custody in violation of the Constitution or laws of the United States. RESULTS IN BRIEF ------------------------------------------------------------ Letter :2 Defender Services representations cannot be compared directly to criminal case filings in district courts because Defender Services' workload also includes representations, such as appeals (reported in appellate court statistics) and habeas corpus proceedings (reported in district court civil case statistics), that are not included in district court criminal caseload statistics. Moreover, Defender Services represents individual defendants, not cases, and during fiscal years 1990 through 1993 there was an average of 1.4 defendants per criminal case filed. AOUSC has maintained that overall program workload has grown and costs have increased because (1) criminal cases, especially drug cases, now involve more defendants; (2) more defendants apply for and receive court-appointed attorneys; (3) more defendants are being tried in federal courts; and (4) the cases are more complex, principally because of changes in federal sentencing guidelines and mandatory minimum sentencing statutes,\10 resulting in more work for attorneys on each case. While each of these factors may have had some effect, we were not able to determine to what extent they individually or collectively accounted for the doubling of the overall program costs or the tripling of DPRC costs from fiscal years 1990 through 1993. This is because AOUSC collected little consistent, national empirical data that could be used to make such determinations. The data available on the costs of representations by federal defender organizations and panel attorneys were not precisely comparable. For example, FPD and CDO data included the costs of training provided to panel attorneys. Also, FPD and CDO cost data were not available by type of case (such as appeals) or major criminal offense category (such as drugs), but such data were available for panel attorneys. Nevertheless, prior studies by the Judicial Conference and AOUSC and our analysis of available data suggested that, overall, FPD and CDO attorney representations cost less than panel attorney representations, although the cost advantage has gradually declined. However, the average cost per representation closed can vary among districts, among FPDs and CDOs, and from year to year depending upon changes in the proportion of expensive cases in the total workload. In fiscal year 1993, for example, each panel attorney drug or fraud representation (at an average cost of more than $3,000) had as much impact on program costs as three immigration representations. The additional costs of paying higher standard hourly rates to panel attorneys in the 16 districts authorized to pay such rates has been a relatively small part of total panel attorney costs. In 78 of the 94 districts, panel attorneys are paid a standard hourly rate of $40 for each out-of-court hour and $60 for each in-court hour.\11 In 1986, Congress authorized higher rates of up to $75 per hour for both in-court and out-of-court hours in specific circuits or districts, if the Judicial Conference determined it was justified. As of March 1995, the Judicial Conference had approved higher standard rates for 89 districts. However, the Judicial Conference, at congressional direction, has limited actual payment of the higher rates--$60, $70, or $75 for both in-court and out-of-court hours--to all or parts of only 16 districts. We estimated (on the basis of available data) that the additional cost of paying these higher standard hourly rates from fiscal years 1991 through 1993 was at least $33.5 million, about 10.2 percent of total panel attorney obligations during the period. Our estimate excluded most appeals representations because the AOUSC database did not designate the district of origin for at least 85 percent of all appeals representations. Death penalty representations have been expensive relative to other types of Defender Services representations. In 1993, the average DPRC cost per representation\12 was about $17,200 and the average panel attorney cost per representation closed was about $37,000.\13 The 246 panel attorney death penalty representations closed in fiscal year 1993 were 0.6 percent of all panel attorney representations closed, but they accounted for 8.7 percent of the $104.3 million in 1993 panel attorney payments approved by the courts. Because of missing data, we could not compare costs in districts with and without DPRCs to determine if panel attorney death penalty representations were less expensive in districts with DPRCs, or whether the cost per representation had grown less in districts with DPRCs. According to Defender Services officials, DPRC costs have increased because more DPRCs have been created, more death penalty cases are in the courts, and the cases are becoming more complex. The number of DPRCs increased from 14 in fiscal year 1989 to 20 in 1994. The number of DPRC death penalty cases also increased from 313 to 1,086, but data on case complexity were limited. AOUSC has embarked on a number of initiatives to address the data deficiencies identified in this report and to improve the information provided to Congress for appropriations and oversight. -------------------- \10 By statute, a minimum prison sentence must be imposed for certain offenses when the statutorily specified criteria have been met. \11 This basic statutory rate has not increased since 1984. \12 DPRC costs per representation include the cost of all cases during the year, including cases pending at the end of the year. \13 According to AOUSC, death penalty representations are more likely than other types of representations to extend over 2 fiscal years or more. Thus, the total cost of death penalty representations may be more than we report, since our cost is based on the total of all vouchers approved for payment in a single fiscal year. SCOPE AND METHODOLOGY ------------------------------------------------------------ Letter :3 To meet our objectives, we reviewed prior studies by the Judicial Conference, AOUSC, and consultants. We met with federal judges and AOUSC officials, members of AOUSC's Defender Services Advisory Group, defender services attorneys, and other officials in six judgmentally selected judicial districts. To try to identify potential causes of workload and cost increases and to compare FDO and panel attorney costs per representation, we reviewed previous analyses and available AOUSC data on FDO and panel attorney workload and costs. For FDOs, AOUSC provided data for fiscal years 1990 through the first 6 months of 1994 on type of representation; type of disposition; representations opened, closed, and pending; and in-court hours per representation closed. For the same period, we analyzed the panel attorney voucher automated database, which included data by type of representation and criminal offense on (1) total in-court and out-of-court compensation requested and approved, and (2) the in-court and out-of-court hourly rates paid for representations in which the panel attorney had requested payment. To provide comparable workload data for FPDs, CDOs, and panel attorneys, we used representations closed because this was the only workload measure available for panel attorneys. Although we did not verify the accuracy of the data we received, we did some edit checks and a distributional analysis of the data, which raised questions about the accuracy of some of the data in the panel attorney database, including the hourly rates recorded in the database. We have qualified our conclusions to take this possibility of error into account. We also used data from AOUSC's master criminal file database and other sources to determine total district court criminal workload and to try to assess the impact of such variables as the number of trials on Defender Services workload. Our analysis of data on criminal trials and cases with multiple defendants used data on total district court workload because the database did not reliably identify that subset of cases in which defendants had court-appointed attorneys. This limitation is noted in our analysis and conclusions. Certain information, including data on cases involving mandatory minimum sentences and federal prosecutions of state-developed cases, was either unavailable or too incomplete for analysis. Also, because data reported by FDOs and panel attorneys differed, workload and cost comparisons between the two were limited. To determine the additional costs of paying panel attorneys the higher standard hourly rates, we used AOUSC's database to calculate the difference in compensation at the higher and lower standard hourly rates. We calculated the in-court and out-of-court compensation approved for those representations in the 16 districts compensated at the higher standard hourly rates. We then estimated the in-court and out-of-court compensation that would have been requested at the lower standard hourly rates of $40 out-of-court and $60 in-court. We subtracted the lower rate estimate from the actual amounts approved for payment at the higher standard rates. The difference was the estimated additional costs of paying the higher standard rates. As explained in more detail in appendix I, this produced a conservative estimate of the additional costs of paying the higher standard hourly rates in the 16 districts. We used DPRC quarterly reports for fiscal years 1990 through the first half of 1994 to examine trends in DPRC workload and costs. However, due to data limitations we could not fully assess whether DPRCs had lowered the costs of death penalty representations. The Defender Services program has been reviewed by various sources, and where appropriate we have incorporated the results of these studies into our work. To determine what actions the Judiciary has underway to improve Defender Services program data and analyses and to control costs, we obtained copies of AOUSC and Judicial Conference reports, directives, and other documents addressing these issues, and we talked to AOUSC officials. We did our work primarily in Washington, D.C.; New York; and Detroit, between April 1993 and February 1995 in accordance with generally accepted government auditing standards. Our objectives, scope, and methodology are discussed in greater detail in appendix I. AOUSC provided written comments on a draft of this report, which are discussed on pages 28-29 and printed in full in appendix III. DEFENDER SERVICES REPRESENTATIONS ARE NOT SYNONYMOUS WITH DISTRICT COURT CRIMINAL FILINGS ------------------------------------------------------------ Letter :4 Data on Defender Services representations are not fully comparable to district court criminal filings or terminations (case closings). In fiscal year 1993, 46,786 criminal cases involving 65,653 defendants were filed in federal district courts.\14 In the same year, district courts closed (disposed) 44,800 cases involving 64,048 defendants. Since Defender Services represents defendants, not cases, the district court defendant data are more appropriate than case data in comparing district court and Defender Services workloads. Defender Services attorneys closed 78,016 representations in fiscal year 1993, which was 13,968 more than the number of defendants reported as disposed in district courts. Of this additional Defender Services workload, at least 9,478 representations (about 68 percent) were not reported in district court criminal workload statistics. The Defender Services 1993 workload, for example, included 6,126 appeals, reported in appellate court statistics; 1,028 habeas corpus proceedings, reported in district court civil statistics; and 2,324 bail proceedings, which Defender Services reported as a separate workload category, but district courts did not. Criminal representations appeared to be the category of Defender Services representations most comparable to district court criminal workload. In fiscal year 1993, Defender Services closed 54,907 criminal representations, or 9,141 less than the 64,048 criminal defendants disposed in federal district courts in 1993. The 9,141 additional defendants reported in district court statistics may have paid for their own attorneys rather than having court-appointed attorneys. However, we found no consistent data on the number of defendants in federal courts who received court-appointed attorneys. -------------------- \14 Totals include traffic cases and defendants as well as transfers between districts. Traffic cases in federal courts arise from traffic violations on federal property. TOTAL DEFENDER SERVICES REPRESENTATIONS CLOSED, TOTAL BUDGET OBLIGATIONS, AND COSTS PER REPRESENTATION CLOSED HAVE GROWN ------------------------------------------------------------ Letter :5 Defender Services appropriations, obligations, representations closed, and costs per representation closed have generally grown in recent years. As shown in table 2, from fiscal years 1990 through 1994, total budget obligations grew about 118 percent from $122.5 million to $266.7 million.\15 According to AOUSC, the growth in representations closed and obligations was less in fiscal year 1994 than in recent years, but it was not clear at the time whether this was the beginning of a long-term trend or a short-term phenomenon. The Judiciary initially requested $387 million for fiscal year 1994, which included $14 million to extend the standard higher panel attorney rates beyond the 16 districts in which they were being paid. Congress rejected this expansion and appropriated $280 million for fiscal year 1994. Defender Services requested $290.3 million for fiscal year 1995, an increase of 3.7 percent. Congress appropriated $250 million for fiscal year 1995.\16 The Judiciary has requested $295.8 million for fiscal year 1996--an increase of 18 percent over the 1995 appropriation and 1.9 percent more than the 1995 budget request. Table 2 Federal Defender Services Budget Obligations by Major Activity, Fiscal Years 1990-1994 (in Thousands of Current Dollars) Percen t change , Program 1990- activity 1990 1991 1992 1993 1994 1994 ------------ ---------- ---------- ---------- ---------- ---------- ------ Federal $38,243 $47,912 $64,992 $78,798 $90,224 135.9% public defenders Community 10,617 12,733\a 17,861 21,865 24,360 129.4 defenders Death 6,000 9,183 11,524 17,758 18,730 212.2 Penalty Resource Centers\b Panel 62,901 85,463 119,204 123,500 122,858 95.3 attorneys Other 4,700 5,800 8,290 9,700 10,552 124.5 program costs ================================================================================ Total $122,461 $161,091 $221,871 $251,621 $266,724 117.8% -------------------------------------------------------------------------------- \a Reflects recoveries of $1,200,000 from fiscal year 1991 obligations that are not shown in the fiscal year 1993 Budget Appendix. \b As noted in its fiscal year 1996 budget request, Defender Services has renamed these organizations Post-Conviction Defender Organizations. Source: AOUSC and Budget Appendix. -------------------- \15 The Defender Services program costs cited in this report exclude AOUSC costs, such as the operations of the Defender Services Division and processing panel attorney payments. Also excluded are district and courts of appeals costs for reviewing and approving panel attorney vouchers; such costs are included in local court budgets. An AOUSC analysis estimated such annual administrative costs at $10.68 million in fiscal year 1992. \16 Defender Services funds are available until expended. In addition to its fiscal year 1995 appropriation, Defender Services had available $36.5 million in unobligated appropriations from prior years to apply to its fiscal year 1995 expenses. OVERALL CHANGE IN NUMBER OF REPRESENTATIONS CLOSED AND COSTS FOR PANEL ATTORNEYS AND DEFENDER ORGANIZATIONS VARIED ---------------------------------------------------------- Letter :5.1 As shown in figure 1, from fiscal years 1990 to 1993 growth in the number of representations closed, total obligations, and the average cost per representation closed\17 for panel attorneys, FPDs, and CDOs varied. To provide comparable data for defender organizations and panel attorneys, our workload figures are based on representations closed, because data for ongoing representations were not available for panel attorneys (see app. I). While panel attorney workload grew more than that of FPDs or CDOs, panel attorney costs per representation closed grew less than FPDs or CDOs. Compensation approved for panel attorney fees grew about 88 percent, while total panel attorney budget obligations--a figure an AOUSC official told us included payments for investigators and experts in addition to attorney fees--grew about 96 percent during the period. Figure 1: Percent Change in Total Representations Closed, Budget Obligations, and Average Cost per Representation Closed for FPDs, CDOs, and Panel Attorneys, Fiscal Years 1990-1993 (See figure in printed edition.) Note: Growth in panel attorney budget obligations excluding death penalty representations was based on total panel attorney obligations less attorney fees in death penalty representations. We did not have data on nonattorney fee costs, such as transcripts and investigators, in death penalty representations. Source: GAO analysis of AOUSC data. -------------------- \17 Cost per representation closed for FPDs and CDOs was determined by dividing total budget obligations for these organizations in each fiscal year by the total number of representations closed. Cost per representation closed for panel attorneys was determined by dividing the total amount of attorney compensation approved for payment in each fiscal year by the number of representations closed. REPRESENTATIONS CLOSED FOR FPDS, CDOS, AND PANEL ATTORNEYS INCREASED ------------------------------------------------------------ Letter :6 Defender Services representations closed annually increased by about 23 percent from 63,504 to 78,016 between fiscal years 1990 and 1993. Of this total, panel attorney representations closed increased about 33 percent from 28,575 to 38,005, or about twice as fast as FDOs' workload, which increased from 34,929 to 40,011, about 15 percent. Within this FDO total, FPD representations closed grew 28 percent and CDO representations closed declined 12 percent. Most of the CDO decline was due to a drop in immigration representations closed at the CDO in the Southern District of California. Excluding this district's workload, CDO representations closed declined about 5.5 percent. Although still a very small percentage of total panel attorney workload (less than 1 percent), panel attorney death penalty representations closed grew from 9 to 246. (More detail is found in the tables in app. II.) As shown in figure 2, from fiscal years 1990 through 1993 criminal representations closed were by far the largest proportion of total representations closed for both FDOs and panel attorneys, followed by "other" and appeals. The number of FDO and panel attorney criminal representations was about even in fiscal year 1990, but FDO criminal representations increased 8 percent during the period while panel attorney criminal representations increased almost 23 percent. Consequently, by 1993 panel attorney criminal representations closed exceeded those of FDOs by 3,567. The growth in panel attorney appeals representations closed (118 percent) was also greater than that for FDOs (54 percent). From fiscal years 1991 through 1993, panel attorneys closed at least 75 percent more appeals representations each year than did FDOs. On the other hand, the number of FDO "other" representations closed was more than twice that of panel attorneys each year for 1990 through 1993, reflecting the much higher number of bail and probation/parole revocation proceedings handled by FPDs and CDOs. Figure 2: Representations Closed by Federal Defender Organizations (FDOs) and Panel Attorneys, Fiscal Years 1990-1993 (See figure in printed edition.) Note 1: Panel attorney data exclude death penalty representations. Numbers for this figure are shown in appendix II, table II.1. Note 2: "Other" includes habeas corpus and bail proceedings, motions to reduce sentence, and probation and parole revocation proceedings. Source: AOUSC. We disaggregated criminal representations closed into five major offense categories--weapons, immigration, drugs, fraud, and other--to examine changes in representations closed by type of offense and type of Defender Services attorneys. As shown in figure 3, within each major offense category, FPDs, CDOs, and panel attorneys had different growth rates. For some offenses there were fewer representations closed in 1993 than in 1990. However, we could not determine the reasons for the variations shown in figure 3. Figure 3: Percent Change in Criminal Representations Closed by Type of Defender Services Attorney and Major Offense Category, Fiscal Years 1990-1993 (See figure in printed edition.) Note: FPD fraud representations closed increased by less than one-half of 1 percent. Source: AOUSC. DETERMINING THE CAUSES OF DEFENDER SERVICES COST GROWTH IS DIFFICULT ------------------------------------------------------------ Letter :7 Changes in Defender Services costs are basically the result of two factors--changes in the number of representations combined with changes in the average cost of each representation. If the number of representations increases, program costs will also increase even if the average cost per representation remains unchanged. Conversely, costs will rise if the cost per representation increases, even though the number of representations remains unchanged. ESTIMATED IMPACT OF INCREASED REPRESENTATIONS CLOSED ON PROGRAM COSTS ---------------------------------------------------------- Letter :7.1 To determine the impact of increased workload from fiscal years 1990 through 1993--as measured by representations closed--on fiscal year 1993 FPD, CDO, and panel attorney costs,\18 we estimated what fiscal year 1993 costs would have been if the average cost per representation closed in 1993 were the same as the 1990 average cost (unadjusted for inflation).\19 As shown in figure 4, about 26 percent of the increase in FPD costs and 38 percent of the increase in panel attorney costs between fiscal years 1990 and 1993 could be attributed to the increase in representations closed in 1993. However, CDO representations closed declined from 11,706 in 1990 to 10,299 in 1993; consequently, none of the increase in CDO costs could be attributed to increased representations closed.\20 Figure 4: Estimated Percent of the Total Increase in FPD, CDO, and Panel Attorney Costs Between Fiscal Years 1990 and 1993 Resulting From Increased Representations Closed or From Other Factors (See figure in printed edition.) Note: Workload was measured by representations closed. Source: GAO analysis of AOUSC data. However, as the panel attorney data in figures 5 and 6 indicate, these overall estimates can mask considerable variation by type of representation.\21 For some types of panel attorney representations closed--appeals, other, death penalty--from about 69 to 92 percent of the increase in costs could be attributed to increased numbers of representations closed (fig. 5). Within criminal representations, increased representations closed accounted for more than half of the increase in the costs of weapons, drugs, and fraud representations closed, but they accounted for only about 13 percent of the increase in the costs of immigration representations closed (see fig. 6). Other criminal representations closed declined 7 percent between fiscal years 1990 and 1993. Thus, all of the increase in costs for these representations must have been the result of factors that have increased the average cost per representation closed. As discussed later in this report, Defender Services officials and AOUSC have suggested a number of factors, such as the sentencing guidelines and a higher number of defendants charged with mandatory minimum offenses, that could have contributed to higher average costs per representation closed. Figure 5: Estimated Percent of the Total Increase in Panel Attorney Costs Between Fiscal Years 1990 and 1993 Resulting From Increased Representations Closed or from Other Factors (by Type of Representation) (See figure in printed edition.) Note: Workload was measured by representations closed. Source: GAO analysis of AOUSC data. Figure 6: Estimated Percent of the Total Increase in Costs of Panel Attorney Criminal Representations Closed Between Fiscal Years 1990 and 1993 Resulting From Increased Representations Closed or From Other Factors (by Major Offense Category) (See figure in printed edition.) Note: Workload was measured by representations closed. Source: GAO analysis of AOUSC data. -------------------- \18 Costs for FPDs and CDOs are based on total obligations each fiscal year divided by the number of representations closed. Because we could not separately identify those costs associated with FPD and CDO representations closed, our average cost per representation closed included the cost of all representations during the fiscal year--new, closed, and pending at the end of the fiscal year--as well as the costs of training provided for panel attorneys. Panel attorney cost data reflect only the costs of attorney fees for those representations for which panel attorneys have requested payment. Panel attorney costs exclude such costs as transcripts, psychiatrists, or investigators--costs that are included in FPD and CDO budgets, and, thus, FPD and CDO cost per representation. (See app I.) \19 For reasons described more fully in appendix I, our 1993 estimates are based on 1990 costs per representation, unadjusted for inflation. The result may somewhat underestimate the impact of increased representations closed on program costs, especially for FPDs and CDOs. \20 The percentage of increased costs attributed to factors other than increased representations closed is the difference between 100 percent and the percent of increased costs attributed to increased representations closed. For panel attorneys, for example, we assumed about 62 percent (100 percent minus 38 percent) of total increased costs were the result of factors other than increased representations closed. \21 Data limitations precluded a similar analysis for FPDs and CDOs. COST PER REPRESENTATION CLOSED HAS GROWN FOR ALL TYPES OF DEFENDER SERVICES ATTORNEYS ---------------------------------------------------------- Letter :7.2 For each fiscal year 1990 through 1993, panel attorney costs per representation closed were the highest and CDO costs the lowest.\22 (See fig. 7.) Costs per representation closed for all three types of Defender Services attorneys increased each year. Although CDOs still had the lowest cost per representation in 1993), the 134-percent growth in CDO costs per representation had narrowed the gap between CDOs and the other two types of attorneys. In 1990, the FPD cost per representation closed was about $300 lower than that of panel attorneys. By 1993, the cost advantage had narrowed to about $100 per representation closed. However, if the average panel attorney cost per representation is calculated excluding death penalty representations, FPD average cost per representation closed in fiscal year 1993 was about $130 more than the panel attorney average. Including death penalty representations in calculating the overall panel attorney average cost per representation added about $10 per representation closed in fiscal year 1990 and about $223 in fiscal year 1993. Figure 7: Average Cost per Representation Closed for FPDs, CDOs, and Panel Attorneys, Fiscal Years 1990 Through 1993 (See figure in printed edition.) Note: Actual costs for this figure are shown in appendix II, table II.6. Source: GAO analysis of AOUSC data. One possible reason for the generally higher panel attorney cost per representation closed may be the higher average number of panel attorney in-court hours per representation closed (see fig. 8). We could not determine why panel attorneys apparently expended more in-court hours per representation closed than did FPDs and CDOs. However, these additional in-court hours can affect program costs. In 78 of the 94 districts the hourly rate for in-court hours is $20 more than the rate for out-of-court hours. Figure 8: Average Number of In-Court Hours per Representation Closed for FPDs, CDOs, and Panel Attorneys, Fiscal Years 1990-1993 (See figure in printed edition.) Note: Panel attorney data exclude death penalty representations. Source: AOUSC. -------------------- \22 Based on aggregate data for all CDOs. CDO costs per representation closed are higher when the workload and costs of the San Diego CDO are excluded from the analysis. All CDO cost and representation data in this section include San Diego. LIMITED DATA WERE AVAILABLE TO VERIFY THE POTENTIAL CAUSES OF INCREASED COSTS PER REPRESENTATION CLOSED ---------------------------------------------------------- Letter :7.3 Overall, between fiscal years 1990 and 1993, we estimated that factors other than increased representations closed accounted for about 74 percent of FPD cost growth, 62 percent of panel attorney growth, and 100 percent of CDO cost growth. Defender Services officials and Judicial Conference and AOUSC studies offered a number of reasons why the cost per representation closed had steadily increased for FPDs, CDOs, and panel attorneys. These reasons included implementation of the federal sentencing guidelines, which require additional attorney time; changing prosecutorial priorities that increased the number of complex drug and multiple defendant cases prosecuted in federal court, which were more likely to result in trials for one or more defendants; and greater numbers of offenses carrying mandatory minimum sentences, which were more likely to result in trials. These arguments were supported by some anecdotal evidence and such aggregate empirical data as total district court criminal caseload statistics showing a rising number of criminal trials and multiple defendant cases. However, it was difficult to document that the reasons cited were in fact the major contributing factors to rising workload and/or costs. This was largely due to the absence of consistent, reliable data on such potential causes of rising program costs as the number of Defender Services cases involving multiple defendants or defendants charged with mandatory minimum offenses; the percentage of such cases that go to trial; or the number of cases investigated and developed by state agencies and handed off to federal prosecutors because federal law and the sentencing guidelines provided longer sentences. A more detailed analysis of the data sources available to measure the impact of these factors is included in appendix I. As discussed in the following sections, there is some evidence that certain of these factors may have affected program costs. IMPACT OF SENTENCING GUIDELINES -------------------------------------------------------- Letter :7.3.1 Recent Defender Services budget submissions have noted that the federal sentencing guidelines, which became effective November 1, 1987,\23 "have profoundly altered the nature of the sentencing process and dramatically increased federal and panel attorney workloads." Contributing to the impact on both attorney workload and program costs has been the need to consider the accuracy of all facts in the judicial proceedings that could affect the potential sentence under the guidelines; the many guidelines amendments (434 between November 1987 and November 1991, for example); the increasing complexity of the guidelines; more defendants choosing to go to trial under the guidelines; and the increased number of sentencing appeals. Our August 1992 sentencing guidelines report\24 concluded that although empirical data and reliable work measurements did not exist, available statistics and our interviews suggested that the guidelines had apparently increased criminal justice workload. Though Defender Services estimated that the sentencing guidelines implementation had increased the time required to provide representation by about 25 to 50 percent in most cases, AOUSC officials could not provide data to substantiate this observation. In addition, subsequent to the sentencing guidelines implementation, there has been a rise in criminal appeals. The authorizing legislation for the sentencing guidelines, the Sentencing Reform Act of 1984, provided for appeals of sentences imposed by U.S. District Courts. For example, the act provided that a defendant or the federal government may appeal a sentence on the basis that the sentence was imposed as a result of an incorrect application of the guidelines. Generally, appeals of sentences were not common prior to the Sentencing Reform Act's appellate review provisions. There were 2,534 such sentencing appeals in fiscal year 1992, or almost 23 percent of all criminal appeals. Defender Services appeals representations closed more than doubled from about 3,300 in 1987 to over 6,700 in 1993, but we could not determine how much of this increase was attributable to appeals of sentence only. The increase in the number of appeals and cost per appeals representation closed has affected program costs. From fiscal years 1990 through 1993, the cost per panel attorney appeals representation closed rose 24 percent from $2,567 to $3,187,\25 while the number of such appeals rose 118 percent from 1,825 to 3,977. During the period, the cost per panel attorney appeals representation closed was at least 20 percent more than criminal representations closed and more than twice as much as "other" representations closed. -------------------- \23 The guidelines apply to all federal criminal offenses committed on or after November 1, 1987. However, the guidelines were not implemented nationally until after January 1989, when a U.S. Supreme Court decision upheld their constitutionality. Mistretta v. United States, 488 U.S. 361 (1989). \24 Sentencing Guidelines: Central Questions Remain Unanswered (GAO/GGD 92-93, Aug. 14, 1992). \25 Data to calculate the cost of FPD or CDO appeals representations closed were not available. NUMBER AND LENGTH OF TRIALS CAN AFFECT COSTS -------------------------------------------------------- Letter :7.3.2 The number and length of criminal trials can also affect program costs. The fiscal year 1994 Defender Services congressional budget submission noted that the number of trials had increased, particularly for defendants in multiple defendant cases or those charged with offenses requiring a mandatory minimum prison term upon conviction. Although specific data on trials involving defendants represented by Defender Service attorneys were not available, we did review general AOUSC statistical data on overall trial trends.\26 The number of criminal jury trials increased more than total criminal trials (9.5 percent versus 2.9 percent) between statistical years 1990 and 1993. Jury trials with 4 or more defendants increased about 15 percent (from 425 to 488) during the same period, though they tended to be a small and stable percentage (between 8.4 and 8.8 percent) of total jury trials. Longer criminal jury trials, those requiring 6 to 10 or 11 to 20 days, increased 22.9 and 17.3 percent, respectively, between 1990 and 1993.\27 But such trials have remained fairly constant as a percent of criminal jury trials. The number of trials lasting 20 days or more has fluctuated by both district and criminal offense. The in-court hours recorded by FPDs, CDOs, and panel attorneys can be one rough measure of trial time for Defender Services attorneys. While average in-court time from fiscal years 1990 through 1993 has generally increased for FPDs, CDOs, and panel attorneys, no clear pattern emerged. For example, while average number of in-court hours for FPDs and CDOs in drug offense representations closed generally increased, it generally declined for panel attorneys. In average hours per representation closed, the largest increase for panel attorneys was for fraud offenses (from an average of 7 to 11 hours); the largest for FPDs was fraud offenses (3.2 to 4.6); and the largest for CDOs was for "other" offenses (2.1 to 4.3). Panel attorneys consistently reported more in-court hours for all types of representations than did FPDs or CDOs, but we could not determine why this was so. Fiscal year 1993 data from the CDO for the Southern District of California in San Diego provided some indication that representations for defendants who had been charged with an offense carrying a mandatory minimum prison term were more likely to be disposed of by trial than other criminal representations. However, the CDO's attorneys spent about the same amount of in-court and out-of-court time on trials that did and did not involve defendants charged with mandatory minimum offenses. The San Diego data indicated that attorneys spent more time negotiating guilty pleas in multiple defendant drug cases that involved mandatory minimums than in multiple defendant drug cases that did not. -------------------- \26 AOUSC provided trial data by statistical year (July 1 to June 30). AOUSC's trial database defines a trial as any contested proceeding before a judge at which evidence is introduced. \27 Trials lasting 6 to 10 days rose from 607 to 746. Trials lasting 11 to 20 days rose from 214 to 251. PROSECUTORIAL ACTIVITIES AND OTHER FACTORS -------------------------------------------------------- Letter :7.3.3 The Judiciary also has stated that workload and costs of Defender Service attorneys in federal criminal cases have increased as a result of Justice Department prosecutorial initiatives, such as the organized crime and drug enforcement task forces. Although specific data were not available, the Judiciary pointed out that these investigations often lead to cases that can be rather complex and expensive to prosecute and defend. Operation Weed and Seed is a multiple agency approach to combatting violent crime, drugs, and gangs in high-crime areas. Project Triggerlock targets dangerous, repeat offenders for prosecution in federal court as opposed to state court to take advantage of federal mandatory minimum penalties for firearms offenses. According to the Judiciary, both initiatives have increased the complexity of cases and the subsequent costs of Defender Services attorneys. However, the courts do not maintain detailed data on the number of criminal cases resulting from these initiatives. Our analysis showed that from fiscal years 1990 through 1993, the number of weapons representations closed\28 by Defender Services attorneys increased by 75 percent to 3,279. Data were not available to determine how many of these resulted from Operation Triggerlock investigations and/or carried mandatory minimum sentences. Various sources have also cited the growing number of multiple defendant cases as a major contributor to increased program costs. Multiple defendant cases can be more complex and time-consuming than other cases, particularly when they involve sophisticated drug trafficking organizations. Conflict of interest concerns have led FPDs and CDOs generally to represent only one defendant in multiple defendant cases. The remaining defendants are generally represented by panel attorneys, whom our analysis and that of AOUSC showed were, overall, more costly per representation closed than FPD and CDO lawyers. Statistics were not available on the number of representations by Defender Services attorneys in multiple defendant cases. However, AOUSC does maintain national data on multiple defendant cases. These data showed that the number of multiple defendant cases increased by about 9 percent between statistical years 1990 and 1993 to about 8,100. However, as a percent of all criminal cases, multiple defendant cases remained fairly constant (between about 19 and 20 percent). The total number of multiple defendant drug cases also increased; cases involving two to five defendants accounted for most of the increase. In 1992 almost 4,000 cases, or 36 percent of all drug cases, involved multiple defendants. However, there were no reliable data available showing multiple defendant drug cases by type of counsel. The number of Defender Services representations closed for drug offenses increased by 39 percent from 15,271 to 21,270 between fiscal years 1990 and 1993. Drug representations closed were 43 percent of all panel attorney criminal representations closed during the period and, excluding death penalty representations closed, generally were the most costly, rising about 18 percent to an average of $3,152 in fiscal year 1993.\29 Similar cost data were not available for FDOs. -------------------- \28 This is a representation in which the defendant is charged with a federal firearms violation, such as the use or possession of a firearm during commission of a felony. \29 By 1993 the average cost of fraud representations was $3,394, more than that for drug representations. However, fraud representations were less than 10 percent of all panel attorney criminal representations in 1993 and, thus, had a much smaller impact on program costs than drug representations. THE COSTS OF PAYING HIGHER HOURLY RATES IN 16 DISTRICTS ------------------------------------------------------------ Letter :8 The current standard panel attorney rates of $40 for each out-of-court hour and $60 for each in-court hour were set in 1984. On the basis of its assessment of such factors as attorney compensation and expenses, the Judicial Conference has approved higher rates in 89 districts (as of March 1995). However, at congressional direction, the Judicial Conference has limited the actual payment of higher standard rates--a single rate of $60, $70, or $75 per hour for both in-court and out-of-court hours--to all or parts of the 16 districts for which the higher standard rates were approved in January 1990. Had the Judiciary paid the full amount panel attorneys requested in these 16 districts during fiscal years 1991 through 1993, the additional costs of paying the higher rates would have been about $40.4 million, about 12 percent of total panel attorney obligations during the period. However, with some regularity judges approved less than the total compensation requested. Based on the amounts judges approved for payment, a more accurate measure, the additional costs were about $33.5 million, or about 10.2 percent of total panel attorney obligations during the period. Because they exclude virtually all appeals representations, these estimates probably underestimate the additional costs. Appeals were excluded because virtually all appeals representations in the database--at least 98 percent in each year for 1990 through 1993 and 85 percent in the first 6 months of 1994--were identified only by the appeals circuit and, thus, could not be associated with a specific district (see app. I). Because the jurisdictions of 11 of the 12 circuit courts of appeals cover several states and a number of districts, we could not use the circuit identifier for our analysis in 11 of the 12 circuits. DATA LIMITATIONS PRECLUDED FULL ASSESSMENT OF INCREASE IN DPRC WORKLOAD AND COSTS ------------------------------------------------------------ Letter :9 DPRCs were created in part to help reduce or contain the cost of death penalty representations. Budget requests and appropriations for the centers have increased sharply since the establishment of the first one in 1988. From fiscal years 1989 to 1994 the number of centers increased from 14 to 20, and the amounts appropriated for the centers more than tripled, from $5,890,000 to $19,800,000.\30 During the period, the overall cost per case increased from $9,521 to $17,247, up about 81 percent (in current dollars). (See table 3.) Table 3 Total Workload, Total Costs, and Cost per Case for DPRCs and Panel Attorney Death Penalty Cases, Fiscal Years 1988- 1994 Fi sc al ye Total Total Cost per Total Cost per ar DPRCs\a cases Total cost\b case\c cases Total cost\b case\c -- -------- -------- ------------ ---------- -------- ------------ ---------- 19 3 NA $90,000 NA NA NA NA 88 19 14 313 $2,980,000 $9,521 NA NA NA 89 19 15 491 $5,640,000 $11,487 9 $299,288 $33,254 90 19 16 670 $11,540,000 $17,224 121 $2,444,894 $20,206 91 19 19 828 $11,540,000 $13,937 156 $5,398,658 $34,607 92 19 19 1,014 $18,065,000 $17,816 246 $9,072,071 $36,878 93 19 20 1,086 $18,730,000 $17,247 NA NA NA 94 ------------------------------------------------------------------------------------ Legend: NA = not available. \a Total number at end of fiscal year. \b Total budget obligations for DPRCs. Panel attorney obligations reflect attorney fees only. \c Rounded to the nearest whole number. Total "cases" and cost per "case" for DPRCs are based on representations pending at beginning of the fiscal year plus new appointments. Total panel attorney "cases" and cost per "case" are based on representations closed during the fiscal year. Source: Judiciary's fiscal year 1996 Congressional Budget Justifications for Defender Services (DPRC data) and other AOUSC data (panel attorneys). AOUSC offered three major reasons for the increases in total costs and costs per representation: more operational centers, increased caseload, and increased complexity of death penalty litigation. While the number of death penalty cases in federal courts has increased, the lack of a reliable death penalty identifier in the database precluded an accurate measure of the increase. Almost all death penalty cases in federal courts arise from state death penalty convictions. Between the end of calendar years 1989 and 1993, the total number of persons sentenced to death, including those convicted and sentenced in state courts, rose from 2,186 to 2,785. As of March 1995, only 6 persons had been convicted and sentenced to death in federal courts. AOUSC asserted that the services provided by the centers actually lower federal death penalty litigation costs by encouraging competent private attorneys to immediately accept death penalty cases; providing consultation, investigative, and other services to private attorneys; and using expert, salaried center attorneys to directly represent defendants/prisoners. Due to data limitations, we could not fully assess the validity of these assertions. Comparable empirical data on case complexity and reliable baseline data concerning death penalty litigation costs prior to implementation of the resource centers were not available, and much of the available cost information was anecdotal. Our analysis of aggregate data from quarterly reports was limited because (1) comparable reports were available only for fiscal years 1991, 1992, and 1993; (2) the 1992 and 1993 reports included centers not operational in 1991; and (3) complete data were not available from all operational centers. These data limitations in turn precluded a full assessment of the costs of death penalty litigation. The absence of a reliable district identifier in panel attorney death penalty representations precluded a cost comparison of such representations in districts with and without DPRCs to determine if the costs were less in districts with DPRCs. -------------------- \30 The fiscal year 1994 appropriation was $265,000 less than the 1993 appropriation of $20,065,000. The fiscal year 1994 request was $30,559,000. EFFORTS TO IMPROVE PROGRAM DATA AND OPERATIONS ----------------------------------------------------------- Letter :10 In its fiscal year 1995 congressional budget submission, the Judiciary acknowledged that it did not have sufficient data or analyses to explain the causes of increased panel attorney costs and announced steps to improve Defender Services program operations, enhance data collection and analysis, and reduce costs. According to AOUSC officials, many of these initiatives were undertaken in response to congressional concerns and/or our inquiries. Several were recommended by the Economy Subcommittee of the Judicial Conference's Budget Committee or the Defender Services Committee. Among the initiatives are (1) the creation of a nine-person program analysis office within the Defender Services Division of AOUSC; (2) efforts to develop Defender Services case weights and FDO work measurement formulas and improve available analytical data by requiring additional financial and statistical data, including FDO reporting of out-of-court time; (3) implementation of new "death penalty" panel attorney vouchers to aid improved cost analysis; (4) continuation of a study to assess the impact of the federal sentencing guidelines on program costs; and (5) improved audits of FPDs, CDOs, and DPRCs. The Judiciary also announced steps to control FDO costs, such as limitations on space alterations, salary increases, and travel. The Judicial Conference's Committee on Defender Services has established a Death Penalty Representation subcommittee (generally known as the Cox Committee, after its Chair, Judge Emmett R. Cox) to reevaluate the DPRC concept and report on proposals for containing death penalty litigation costs. The subcommittee is reviewing available data, soliciting the views of various sources, and planning to complete its work during the summer of 1995. AOUSC's Office of Audit continues to play a role in evaluating various aspects of the program. The Office, either directly or through its contractor accounting firm, conducts periodic financial audits of court operations, which include reviews of FPD activities and panel attorney payments. The Office's audit staff has identified some FPD office deficiencies regarding payroll, time and attendance, and disbursements, as well as problems in some DPRCs' reporting on time spent on federal representations. At its January 1995 meeting, the Judicial Conference's Committee on Defender Services directed AOUSC to contract centrally for audits of CDOs and DPRCs. AOUSC's comments on a draft of this report mentioned some additional initiatives (see app. III). AGENCY COMMENTS ----------------------------------------------------------- Letter :11 AOUSC's written comments incorporated the views of the Chair of the Judicial Conference's Committee on Defender Services. AOUSC said that the findings and issues identified in our report would assist the Judiciary in its ongoing efforts to increase the type, quality, and consistency of data being collected on the activities of FDOs, DPRCs, and panel attorneys. AOUSC identified a number of specific efforts, planned and underway, to improve data collection and analysis and contain Defender Services costs. A number of these efforts were initiated in response to our work. On the basis of our review of the information provided by AOUSC, we agree that successful completion of AOUSC's efforts--both those mentioned in its letter and those discussed in our report-- would help to rectify the issues identified in our report. The full text of AOUSC's comments are in appendix III. --------------------------------------------------------- Letter :11.1 We are sending copies of this report to the Subcommittees on Commerce, Justice, and State, the Judiciary, and Related Agencies of the House and Senate Appropriations Committees; the Chairmen and Ranking Minority Members of the House and Senate Committees on the Judiciary; the Chairman and Ranking Minority Member of the House Judiciary Subcommittee on the Judiciary and Intellectual Property; the Chairman and Ranking Minority Member of the Senate Judiciary Subcommittee on the Courts and Administrative Practice; the Director of the Administrative Office of the U.S. Courts; the Chairman of the Judicial Conference of the United States' Committee on Defender Services; and other interested parties. We will also make copies available to others upon request. Major contributors to this report are listed in appendix IV. If you have any questions about the contents of this report, please call me on (202) 512-8777. Sincerely yours, Norman J. Rabkin Director, Administration of Justice Issues OBJECTIVES, SCOPE, AND METHODOLOGY =========================================================== Appendix I The former Chairman, and now Ranking Minority Member, of the Subcommittee on the Departments of Commerce, Justice, and State, the Judiciary, and Related Agencies, Senate Committee on Appropriations, requested that we examine a number of issues regarding the Federal Defender Services program and the causes of the rapid growth in both program workload and costs. Specifically, we were asked to assess, for fiscal years 1990 to 1993 and the first half of 1994: the causes of the growth in overall Defender Services workload and costs, including the reasons that Defender Services workload has grown faster than district court criminal cases; the comparative costs of representations closed provided by FDOs and panel attorneys and, if the FDO costs were lower, what actions the federal judiciary was taking to increase the use of FDOs; the additional costs of paying standard higher hourly rates to panel attorneys in all or parts of 16 districts; and the causes of the increased workload and costs of the DPRCs, including whether the DPRCs had helped to reduce or contain the costs of panel attorney death penalty representations. However, because of the data limitations described below, and as agreed with the requestor, we changed our objectives to provide information on (1) reasons that Defender Services workload has grown faster than district court criminal cases; (2) data available to assess the causes of increased Defender Services workload and costs; (3) comparative costs of representations closed by federal defender organizations and by private attorneys directly assigned by federal judges; (4) additional costs of paying higher standard hourly rates to private, court-appointed attorneys in all or parts of 16 districts; (5) the comparative costs of panel attorney and DPRC death penalty representations; and (6) the potential causes of increased DPRC workload and costs. To meet these objectives, we reviewed prior studies of the Federal Defender Services program, including those of the Committee to Review the Criminal Justice Act (commonly referred to as the Prado Committee), the Judicial Conference of the United States, the Defender Services Division, and the Financial Analysis Office of the Administrative Office of the U.S. Courts (AOUSC). We also reviewed a number of private studies of death penalty litigation, including studies of DPRCs by the Spangenberg Group, a consulting firm. We met with federal court officials, including the Chairman of the Executive Committee of the Judicial Conference of the United States, the Chair of the Judicial Conference Committee on Defender Services, the Defender Services Advisory Committee, and the Chief Judge of the Second Circuit court of appeals. We also met with judges, federal and community defenders, and other officials in the Eastern and Southern Districts of New York, Eastern District of Michigan, District of New Jersey, District of Delaware, and Southern District of California. These districts were judgmentally selected. Delaware was chosen because it was served by the FPD for New Jersey. The districts of Southern and Eastern New York, Eastern Michigan, and New Jersey had large criminal caseloads and were near our regional office staff. We chose the Southern District of California because Defender Services officials told us this district had developed a method of tracking attorney hours that captured many of the elements, such as representations for defendants charged with offenses carrying a mandatory minimum sentence, not captured by the national data systems used in our analysis. We also met with AOUSC officials in the Office of Audit, Defender Services Division, Office of Finance and Budget, and Statistics Division to discuss a number of issues, including studies they had or were conducting on Defender Services workload and costs. DATA AVAILABLE ON DEFENDER SERVICES WORKLOAD AND COSTS We identified the workload and cost data available from AOUSC. FPD and CDO workload data are reported semiannually by each office on form JS-50. We obtained these semiannual workload reports for fiscal years 1990, 1991, 1992, 1993, and the first half of fiscal year 1994 and produced a dataset to analyze FPD and CDO workload by case type, disposition (where available), and type of representation--criminal, appeals, and all other representations (such as bail and probation/parole revocation hearings and habeas corpus proceedings). We obtained a copy of the automated database created from the vouchers submitted by panel attorneys for fiscal years 1990 through 1993, plus the first 6 months of fiscal year 1994. This database contained data on the amount of in-court and out-of- court compensation requested by panel attorneys and approved by the court, the approval date, the payment date, and the attorney's authorized hourly in-court and out-of-court billing rates. We did not verify the accuracy of the data entered on the forms or the database AOUSC provided. However, edit checks on the internal reliability of data in selected data fields in the database, such as attorney hourly rates, revealed some questions about the accuracy of the data in these fields. These problems are discussed in later sections of this appendix. AOUSC's master criminal file provided a wide variety of data on overall criminal workload in the federal district courts. We used this database, the annual statistical reports of the courts, and a Statistics Division analysis to identify trends in total district court criminal workload, such as the number of multiple defendant criminal cases filed in district courts, the number of trials with four or more defendants, and trial lengths in days. However, the master criminal file did not reliably differentiate between Defender Services cases and other cases, precluding a comparison of overall workload trends and Defender Services workload trends. In addition, data on some of the workload and cost variables we were asked to examine were unavailable or too incomplete to be analyzed. These included data on: the number of defendants who required court-appointed counsel because their assets had been seized or frozen (not available); the total number of cases in which defendants were charged with offenses carrying mandatory minimum sentences (not available nationally, although we found that the CDO for the Southern District of California maintained such data, beginning in fiscal year 1993); the number of panel attorney appointments resulting from multiple defendant cases in which more than one defendant required a court-appointed attorney (not available nationally);\1 the additional hours Defender Services attorneys expend on representations because of the implementation of the federal sentencing guidelines (not available); federal prosecutions of cases developed by state investigative agencies, such as weapons or drug cases prosecuted in federal courts to take advantage of higher mandatory minimum sentences imposed under the federal sentencing guidelines (not available from AOUSC).\2 the link between the increased number of assistant U.S. Attorneys and the number of criminal cases filed in U.S. district courts (only basic trends can be described from available data, such as the percent increase in resources for criminal prosecutions and criminal cases initiated in federal courts). Defender Services officials said that each of these factors had contributed to increased attorney hours and costs per representation closed. Early in our work we informed Subcommittee staff that these data limitations precluded an assessment of the direct impact of individual factors on Defender Services workload and costs. Therefore, as agreed with the requestor, we changed our objectives. Table I.1 summarizes the limitations of available Defender Services workload and cost data for criminal representations closed. Table I.1 Criminal Defendant Workload Data Available From Four Databases of the Federal Judiciary All Panel San criminal FDOs attorney Diego Type of data cases\a \b s\c CDO\d ------------------------ -------- ---- -------- -------- Workload data ------------------------------------------------------------ Number of new defendants Yes\e Yes No\f Yes who received appointed counsel Number of multidefendant Yes No No Yes cases Number of defendants Yes Yes\ Yes\g Yes whose cases were g completed (disposed) Number of in-court hours No\h Yes Yes Yes per representation closed Number of out-of-court No\h No\i Yes Yes hours per representation closed Disposition and offense data ------------------------------------------------------------ Total dispositions by Yes No No Yes type of disposition--cases\j Total dispositions by Yes Yes\ No Yes type of g disposition--defendants Disposition by type of Yes\k No No Yes offense Number of defendants charged with mandatory In No No Yes minimum offense part\l Cost data ------------------------------------------------------------ Cost per representation No Yes Yes Yes closed\m Cost per representation closed by district No\h Yes\ In Yes n part\o Cost by type of No\h No Yes\ No representation closed\p Cost by offense category No\h No Yes No Appeals data ------------------------------------------------------------ Number of appeals closed Yes\ Yes Yes Yes Appeal of sentence Yes No\q No\q Yes Appeal of conviction Yes Yes Yes Yes ------------------------------------------------------------ \a Criminal cases in district courts were recorded in the master criminal file; the appeals database contained data on criminal appeals; the trial database included data on the number and length of criminal trials. \b Data for these offices are reported semiannually on the JS-50 form. \c Panel attorney workload data were derived from the CJA-20 database which includes data from the forms (called CJA-20s) on which panel attorneys requested payment and on which approved payments were recorded. Attorney requests for payment for other expenses, such as expert witnesses and investigators, were reported on the CJA-21 form, which we did not review. \d The CDO for the Southern District of California in San Diego has developed its own automated system for tracking attorney hours and assignments that includes many data elements not recorded on the forms we reviewed. The San Diego system was not designed to track costs per representation. \e The master criminal file included a field for type of legal counsel, including whether the defendant had a court-appointed attorney. However, according to AOUSC, the data in this field were not reliable. Our own validity test confirmed this assessment. The data in this field could be entered twice--when the case was filed and when the case was closed. \f A representation appeared in the database only when an attorney submitted a request for partial or complete payment. This request may not have appeared in the file for the fiscal year in which the attorney was appointed. If the attorney submitted two vouchers for the same representation in two different fiscal years, the second voucher was counted as a separate representation. Thus, it was possible to double-count representations and defendants in such cases. For example, in fiscal years 1991 and 1992 we found that 1,591 (about 2 percent) of the panel attorney representations were double-counted. \g For panel attorneys, determining the number of unique defendants disposed requires matching individual vouchers with a number of data elements as discussed later in this appendix. The number of defendants disposed by each type of counsel cannot necessarily be added together to obtain the total number of separate defendants whom FDOs and panel attorneys have represented. It is possible for a single defendant to be represented by more than one type of attorney. For example, the defendant could be represented by a FPD or CDO attorney at the bail hearing, but by a panel attorney for the remainder of the case. In such instances, Defender Services data would report two representations--one for FDOs in the JS-50 reports (under "other" representations for the bail hearing), and one in the CJA-20 database (under criminal representations) for the panel attorney representation. \h Not available. The master criminal file and appeals databases were designed to track overall workload, not cost per case or attorney or judge hours per case. \i Out-of-court hours were to be reported beginning in fiscal year 1995. \j Types of dispositions included guilty pleas, dismissals, acquittals, and trials, including both nonjury and jury trials. The master criminal file defines nonjury trials as any contested proceeding before a judge at which evidence is introduced. \k The master criminal file contained data on the type of disposition by case and defendant, but not on the length of the trial, if there was a trial. The trial database contained data on the length of criminal trials but not by type of offense, such as drugs or weapons. \l The master criminal file included the most serious charge at indictment and the most serious charge of conviction as determined by the maximum penalty that could be imposed. If the mandatory minimum offense were not the most serious charge in either case, it would not be shown in the database. \m In its December 1992 study of Defender Services costs, AOUSC calculated the cost per representation closed for FPDs and CDOs by dividing their total obligations by the total number of representations closed. In the fiscal year 1995 budget submission, cost per representation closed was calculated as total obligations divided by total new appointments. The cost data for panel attorneys in both the December 1992 study and the fiscal year 1995 budget submission reflected representations for which at least partial payment had been made. Our analysis used representations closed for both types of counsel. For panel attorneys, we used the same definition of case closed as AOUSC. \n Some FDOs serve more than one district. There is a place on the form to identify the district of jurisdiction. We did not use these data or check their reliability. \o In the database the district of origin was not available for about 98 percent of appeals representations closed for fiscal years 1990 through 1993 and for about 85 percent of representations closed in the first 6 months of fiscal year 1994. The district of origin was missing for about 10 percent or less for all other types of representations. \p Types of representations included criminal, appeals, and other (which includes bail hearings, probation and parole revocations, and civil actions, such as habeas corpus proceedings and prisoner petitions). \q Data included total number of appeals, but not whether the appeal was an appeal of conviction, sentence, or both. Under the federal sentencing guidelines, a defendant may appeal the sentence without appealing the conviction. The government may also appeal the sentence. The types of data reported for panel attorneys and for FDOs differed in many respects, significantly limiting the comparisons that could be made between the work of panel attorneys and defender organizations. As shown in table I.1, the criminal caseload data available from the principal judiciary databases were varied and not always comparable. While the number of multiple defendant cases could be obtained from the master criminal file, for example, we could not determine how many defendants in these cases required court-appointed counsel. The national workload data available for FDOs and CDOs included the number of new representations, number of representations closed, and number of representations pending at the end of the fiscal year. However, the data available on panel attorneys included only the number of representations closed--representations for which panel attorneys had requested payment. If the representation were closed, but the panel attorney did not request payment during the same fiscal year, the representation would not appear in the database until the attorney had submitted a voucher for payment. COMPARISON OF DEFENDER SERVICES REPRESENTATIONS WITH DISTRICT COURT CRIMINAL WORKLOAD To determine why the rate of Defender Services representations closed has exceeded the increase in criminal filings in recent years, we tried to identify those Defender Services representations that were most analogous to criminal filings in district courts. Defender Services representations were reported in three broad categories--criminal, appeals, and other. After consulting with Defender Services and AOUSC's Statistics Division, we determined that the category "criminal representations" was most analogous to criminal filings as reported by the district courts and recorded in the master criminal file. Appeals are reported in AOUSC statistical reports as workload of the courts of appeals. Thus, such representations would not be reported in district court workload statistics. The category "other representations" included proceedings, such as prisoner petitions and habeas corpus petitions, that are civil actions and reported in district court civil workload statistics. Defendants disposed rather than commenced in district courts was the most appropriate comparison between Defender Services criminal representations and district court criminal workload statistics. This was because panel attorney data were not available on new appointments but were available only for representations closed (those for which payment had been requested). When this comparison was made, we found that Defender Services criminal representations did not exceed the number of criminal defendants disposed in district courts. COMPARING FDO AND PANEL ATTORNEY WORKLOAD AND COSTS As summarized in table I.1, comparable workload and cost data for FDOs and panel attorneys were limited. FPDs and CDOs reported aggregate data on case disposition (such as trial or guilty plea); panel attorneys did not. Panel attorneys reported both in-court and out-of-court hours; FPDs and CDOs reported only in- court hours. Because panel attorneys submitted vouchers for each type of representation they were assigned--criminal representations, appeals, other--it was possible to calculate the average cost per representation for different types of panel attorney representations. However, FPD and CDO cost data were not available by type of representation. Consequently, we could not determine FPD and CDO costs by type of representation using existing data. Therefore, we used aggregate FPD and CDO costs per representation when comparing panel attorney and FDO costs per representation. An AOUSC study found that CDOs had a lower average cost per representation closed than FPDs or panel attorneys. Defender Services officials suggested that a large number of relatively inexpensive immigration representations, particularly in the CDO for the Southern District of California in San Diego, may have largely accounted for the lower average CDO costs. The CDO for the Southern District of California accounted for more than half of all Defender Services immigration representations. Because of its impact on total CDO workload and costs, we examined CDO workload and costs with and without San Diego included in the analyses. The San Diego CDO has developed its own system for tracking attorney assignments and hours that captured a variety of data not available from the AOUSC forms, such as representations involving offenses carrying a mandatory minimum prison term, dispositions by type of offense, in-court and out-of-court hours by type of offense, and multiple defendant cases. The San Diego system was not designed to capture costs per representation closed but was principally designed to track attorney hours and aid in attorney assignments. The fiscal year 1993 San Diego data on attorney hours provided a case study of the impact of multiple defendant cases and representations involving mandatory minimum offenses on attorney hours. The San Diego data also illustrated how such attorney hour data could aid the Defender Services Division in analyzing its workload. -------------------- \1 In such cases, only one defendant was generally to be assigned to the local FPD or CDO; all remaining defendants requiring court-appointed counsel were to be assigned to panel attorneys. The rationale for limiting FPD and CDO appointments to a single defendant in multiple defendant cases is to avoid conflicts of interest that may arise from representing more than one defendant in the same case. \2 The U.S. Attorney database indicated that the U.S. Attorneys initiated a total of 35,000 criminal cases in federal district courts in fiscal year 1992, including 1,158 cases referred by state investigative agencies and 344 cases referred by District of Columbia authorities. About two-thirds of the cases referred by state or D.C. authorities were narcotics cases. HOW WE COUNTED FPD, CDO, AND PANEL ATTORNEY REPRESENTATIONS ------------------------------------------------------- Appendix I:0.1 To compare the workload and per representation costs for FPDs, CDOs, and panel attorneys, we developed a working definition of representation, and calculated costs and in-court hours per representation. As explained below, we used representations closed as our basic measure of Defender Services workload. Because FPDs and CDOs did not report out-of-court hours per representation closed for the years we reviewed, we could only compare in-court hours per representation closed for FPDs, CDOs, and panel attorneys. Prior to its fiscal year 1996 submission, the Judiciary's congressional budget submissions counted workload differently for FDOs and panel attorneys. FPDs and CDOs have been reported as cases opened, the basic equivalent of new court appointments. For panel attorneys, the submissions have used "cases against which payment is made in a given fiscal year." In its fiscal year 1996 submission, the Judiciary used "representations closed" for reporting FPD and CDO workload, adopting the method we used to count workload. Our analysis used FPD and CDO representations closed as reported on the AOUSC forms, because data for panel attorneys were available only for (1) closed representations; or (2) representations for which panel attorneys had requested partial payment, though the case may not yet have been completed.\3 We defined a panel attorney representation as the representation of a defendant charged with or convicted of a specific offense by a private, court-appointed attorney. Representations for convicted defendants included appeals of conviction and/or sentence or habeas corpus proceedings in which the defendant may be challenging a state conviction in federal court.\4 We took several steps to avoid double-counting panel attorney representations closed. When a panel attorney was appointed, the court assigned a unique voucher number that the panel attorney was to use when submitting request(s) for payment. Panel attorneys may submit more than one voucher for each representation using the voucher number they were assigned upon appointment. We considered all claims for payment under a single voucher number to be part of a single representation. If payments under the same voucher number were approved and paid in two successive fiscal years, AOUSC counted the request for payment in the new fiscal year as a separate representation. We found 691 voucher numbers that appeared in fiscal years 1990 and 1991, 900 numbers that appeared in fiscal years 1991 and 1992, and 1,202 numbers that appeared in fiscal years 1992 and 1993.\5 Thus, AOUSC would have counted as separate representations 691 voucher numbers in 1991, 900 in 1992, and 1,202 in 1993. We followed the AOUSC convention because we were tracking payments per representation closed in each fiscal year. Thus, like AOUSC, our counts of representations closed in each fiscal year were not precisely accurate. First, we overestimated representations actually closed in fiscal year 1990 by as much as 691 (2.4 percent of the 28,575 representations we attributed to 1990); in fiscal 1991 by as many as 900 (2.5 percent of 36,027 representations); and in fiscal year 1992 by as many as 1,202 representations (3.3 percent of 36,479 representations). Second, this method of counting did not accurately reflect panel attorney costs per representation closed. If each voucher number appearing in two fiscal years represented a single representation, then the true cost for each representation closed would be the total of all the payments made under each voucher number for both fiscal years. This approach affected hours per representation closed for the same reason. The total attorney hours appearing in both fiscal years would be the true hours per representation closed. Some defendants were represented by more than one panel attorney in the same case, but we counted such situations as a single representation. To determine a single panel attorney representation, we matched offense, court district, docket number, and defendant name. Multiple records and multiple voucher numbers that matched exactly on these data elements were considered part of a single representation. If they did not match on all data elements, we counted them as separate representations. -------------------- \3 While panel attorneys are generally expected to submit a single voucher after the representation is closed, they may and do submit interim vouchers for payment. \4 Habeas corpus proceedings are civil, not criminal, proceedings. \5 We did not determine how many numbers, if any, appeared in more than two successive fiscal years. HOW WE CALCULATED COSTS PER REPRESENTATION CLOSED ------------------------------------------------------- Appendix I:0.2 We calculated FPD and CDO costs per representation closed by dividing the total budget obligations for all FPDs or CDOs by the total number of FPD (or CDO) representations closed in that fiscal year, as reported on the AOUSC forms. Thus, if total FPD obligations were $100,000 and total representations closed were 1,000, the cost per FPD representation closed would be $100 ($100,000/1,000). This calculation was not very precise, but available data did not permit a more detailed estimate of costs per representation closed. This methodology may have overstated the FPD and CDO costs per representation closed because an undeterminable portion of their budgets was used for representations opened, but not closed, during the fiscal year. In addition, FDO attorneys provide training for panel attorneys, the costs of which are included in FDO budgets. The calculation of panel attorney costs per representation closed was somewhat more complicated. The AOUSC used "payment date"--the date on which the check was written--to calculate per representation costs. In fiscal years 1991 and 1992, payments to panel attorneys were deferred due to shortage of funds, with large numbers of payments made in the first month of the new fiscal year. Thus, payment date was not necessarily the best measure of when the representation was closed. Therefore, we used "certification date," the date on which the court approved payment, because this date was likely to be closer to the date on which the work was completed and payment requested than the payment date. We calculated costs per representation closed using a two-step process. First, to determine costs for closed representations having a single panel attorney, we aggregated all payments for each voucher number. That is, if there were five payments approved for voucher number 12345, we combined all five payments to obtain a single cost for that voucher number. Second, we aggregated all payments for each voucher number associated with the same representation. For example, if voucher numbers 54321 and 67890 were associated with the same representation, we combined them into a single total for that representation. As noted, our panel attorney costs per representation closed may be somewhat understated since in the database we used the costs for as many as 691 representations that were split between fiscal years 1990 and 1991, 900 between 1991 and 1992, and 1,202 between 1992 and 1993. Moreover, our costs per representation closed included attorney time only. The costs of experts, investigators, and other services, reported on another form, were not included in our costs because of the difficulty of matching these payments, recorded in a separate database, with the representations recorded in the AOUSC database for attorney hours. However, such costs were included in the costs for FPDs and CDOs, since these costs were included in the total budget obligations we used to calculate FPD and CDO costs. Thus, the FPD and CDO cost per representation closed used in our analysis included all costs incurred by these organizations, while our panel attorney cost per representation closed included only attorney time. Consequently, any FPD or CDO cost advantage we found in our analysis was probably somewhat understated. HOW WE CALCULATED HOURS PER REPRESENTATION CLOSED ------------------------------------------------------- Appendix I:0.3 We used the in-court hours reported on the JS-50 forms for representations closed to calculate FPD and CDO in-court hours per representation closed. FPDs and CDOs did not report out-of- court hours. Panel attorney vouchers included separate sections for in-court and out-of-court compensation requested and approved. In all but 16 districts panel attorneys were reimbursed at the standard rate of $60 per hour for in-court time and $40 per hour for out-of-court time. Thus, in these districts in-court hours are reimbursed at a rate 50 percent higher than out-of-court hours. The court may reduce the amount requested for either in-court compensation, out-of-court compensation, or both. However, where the court made a reduction in the amount approved for payment, the database did not usually indicate the number of attorney hours by which the request had been reduced. Rather, the database showed separately the amount requested and the amount approved for payment. Because attorney hours were not entered into the automated database, AOUSC calculated the actual number of in-court hours expended per panel attorney representations by dividing the total amount of compensation requested (for example, $2,400) for in-court time by the authorized hourly rate ($60 in most districts). The same calculation was made for out-of-court hours. The database supports no other reasonable method of calculation. Consequently, we used the same method to calculate attorney hours per representation closed. If the the authorized hourly rate is recorded erroneously on the CJA-20, this would affect the calculation of attorney hours. We found a wide variety of hourly rates reported in the database for both death penalty and nondeath penalty representations. Some appeared to be keying errors. For example, in more than 1,300 nondeath penalty panel attorney representations in fiscal year 1993 the out-of-court hourly rate was recorded as $752, perhaps reflecting a data entry error. The correct rate was probably $75--the higher standard hourly rate in those districts authorized to pay the higher rates. However, because of the method used to calculate panel attorney hours, this error would reduce by a factor of 10 the actual out-of-court hours reported as expended in such cases. Due to resource limitations, we did not validate the hourly rates entered in the database. Consequently, we could not determine the magnitude of the error that may be attributable to such erroneous hourly rates, and the data used in our analyses were based on the rates as reported in the database. FPDs and CDOs were not assigned death penalty cases during the period of our review. Therefore, when comparing FPD, CDO, and panel attorney average costs per representation closed, we used the average panel attorney cost per representation closed, excluding death penalty representations. We separately calculated the costs for panel attorney death penalty representations closed for the analysis on death penalty representations. HOW WE ESTIMATED THE IMPACT OF INCREASED REPRESENTATIONS CLOSED ON PROGRAM COSTS ------------------------------------------------------- Appendix I:0.4 To estimate the impact of increased workload (as measured by representations closed) on program costs, we estimated what fiscal year 1993 costs would have been if the average cost per representation closed had remained unchanged throughout the period at the fiscal year 1990 average cost per representation closed (unadjusted for inflation). For example, in fiscal year 1990, total panel attorney criminal representation costs were $47,212,797, with an average cost of $1,985 each. At $1,985 each, the 29,237 panel attorney criminal representations closed in fiscal year 1993 would have cost $58,035,445, or $10,822,648 more than the total 1990 costs. However, total 1993 panel attorney costs for criminal representations were actually $30,009,222 more than the 1990 total. Thus, we estimated that the increased workload accounted for about 36 percent of increased panel attorney costs for criminal representations between fiscal years 1990 and 1993.\6 We did not adjust the 1990 cost per representation closed for inflation because we wanted a comparable estimate for panel attorneys and FDOs. While FDO attorneys received cost-of-living adjustments during the period (though not always at the same time in the fiscal year), the basic panel attorney hourly rates in 78 of 94 districts remained unchanged at $40 (out-of-court) and $60 (in-court). We had already separately estimated the impact of paying higher standard hourly rates to panel attorneys in the 16 districts in which higher rates were paid in fiscal years 1991, 1992, and 1993. Although our use of 1990 average costs unadjusted for inflation may have underestimated the impact of increased workload on program costs, particularly for FDOs, it was the most straightforward method of developing reasonable, comparable estimates for panel attorneys and FDOs. HOW WE CALCULATED THE ADDITIONAL COST OF PAYING HIGHER HOURLY RATES IN 16 DISTRICTS In 78 districts, panel attorneys are paid a standard hourly rate of $40 for each out-of-court hour and $60 for each in-court hour. In all or parts of 16 districts, panel attorneys are paid a single, standard higher hourly rate of $60, $70, or $75 for both in-court and out-of-court hours. The applicable rate varies by district. The AOUSC database included the hourly rate and the separate amounts of compensation requested for in-court and out-of-court hours, but the database did not include the number of in-court or out-of-court hours on which the request was based. We excluded from our analysis representations compensated at rates other than the standard higher rates, as applicable, of $60, $70, or $75 per hour in these 16 districts. Representations compensated at other hourly rates would, by definition, have been exceptions to the standard higher hourly rates prevailing in these districts. In some of these districts, the higher rate applies only to specific location(s) within the district. To the extent that the higher standard rates were approved in specific cases outside these locations, they would be included in our analysis because we could identify representations only by district, not by specific locations within a district. We calculated the number of in-court hours and out-of-court hours by dividing the total requested amount by the approved hourly rate. We used these hours to recalculate the requested compensation at the rate of $40 for each out-of-court hour and $60 for each in-court hour. This provided a comparison between requested compensation at the higher standard rate in these districts and the amount that would have been requested using the lower standard hourly rates of $40 out-of-court and $60 in-court. However, the courts reduced some requests for compensation at the higher rates. Consequently, a comparison based solely on requested compensation would overstate the amount actually paid in the higher-rate districts. To determine how much had actually been approved and paid in each of the 16 higher-rate districts, we used the AOUSC database to identify the amount by which the requests for compensation had been reduced in each district. We then compared the amounts actually approved and paid at the higher rates with the amounts that would have been requested at the lower rates. We did not also reduce the calculation using the lower rates because the database provided no basis on which to make such an adjustment. The database did not explain the basis for the reductions in requested compensation, nor were there any data on the number of hours disallowed--only final dollar amounts approved and paid. Our method of calculation produced a conservative estimate of the additional costs of paying the higher rates in the 16 districts. Our estimates and cost comparisons were also conservative because we excluded from the analysis any voucher without a district identifier. For every type of representation--criminal, appeals, other--some representations did not include district identifiers, but only the appeals circuit in which the district was located. Each circuit, except the District of Columbia circuit, covers a number of states and districts. The higher-cost districts are in the second, third, seventh, ninth, and tenth circuits, but not all districts in each of these circuits were authorized to pay the higher panel attorney hourly rates. At least 98 percent of all appeals in the database for fiscal years 1990 through 1993, and 85 percent in the first half of 1994, were identified only by the circuit in which the appeal was filed. Consequently, our estimate of paying the higher rates excluded the costs of most appeals originating in the districts authorized to pay the higher rates. DEATH PENALTY RESOURCE CENTERS To assess the costs of representations by DPRCs we relied primarily on DPRC budget and workload data included in the DPRCs' quarterly reports and data on panel attorney costs per death penalty representation closed from the AOUSC database. Because there were no baseline cost data prior to the DPRCs' creation, we could not determine whether the services the DPRCs provided had helped to reduce the costs of death penalty litigation. Because of considerable variations in the type of work each DPRC does, including the mix of direct representations and assistance provided to private counsel, we could not determine whether the increased costs of the DPRCs were justified by the workload. About 37 percent of panel attorney death penalty representations closed in fiscal year 1991, 23 percent in fiscal year 1992, and 17 percent in fiscal year 1993 were not identified by district. This precluded a comparison of the costs of panel attorney death penalty representations closed in districts with and without DPRCs. To develop descriptive information on DPRCs, we reviewed the fiscal year 1991, 1992, and 1993 quarterly reports submitted by the DPRCs on their workload. California did not submit quarterly reports until fiscal year 1994 because it submitted monthly vouchers instead; fiscal year 1991 quarterly reports for the Georgia DPRC were also unavailable. We also reviewed studies by the Spangenberg Group, a consulting firm, on death penalty litigation and costs, and incorporated some of their data in our analysis, as appropriate. -------------------- \6 This is derived by dividing $10,822,648 (the cost of the increased workload at the 1990 average cost per representation closed) by $30,009,222 (the actual increased costs). SELECTED DATA ON DEFENDER SERVICES WORKLOAD AND COSTS ========================================================== Appendix II This appendix provides selected tables on Defender Services workload and costs for fiscal years 1990 through 1993. Where available, data for the first 6 months of fiscal year 1994 are included. The tables in this appendix supplement the tables and figures in the letter. Only overall costs per representation closed were available for FPDs and CDOs. However, per representation costs by type of representation and major offense category (within criminal representations) were available for panel attorneys and are shown in tables II.7 and II.8. Table II.1 Representations Closed by Federal Defender Organizations and Panel Attorneys, Fiscal Years 1990 Through First 6 Months of 1994 Type of representa tion FDOs PAs FDOs PAs FDOs PAs FDOs PAs FDOs PAs FDOs PAs ---------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- --------- ---------- ------------------ Criminal 23,743 23,782 24,660 28,119 24,634 28,555 25,670 29,237 12,115 13,485 8% 23% Appeals 1,391 1,825 1,746 3,699 1,818 3,225 2,149 3,977 1,077 2,000 54 118 Other 9,795 2,959 10,741 4,088 11,888 4,543 12,192 4,545 5,524 2,395 24 54 Death NA 9 NA 121 NA 156 NA 246 NA 171 NA 2,633 penalty ================================================================================================================================================================= Total 34,929 28,575 37,147 36,027 38,340 36,479 40,011 38,005 18,716 18,051 15% 33% ----------------------------------------------------------------------------------------------------------------------------------------------------------------- Legend: FDOs = Federal Defender Organizations PAs = Panel Attorneys NA = not applicable \a Percent change based upon fiscal year 1990-1993 data only. Source: GAO analysis of AOUSC data. Table II.2 Defender Services Representations Closed by Type of Attorney, Fiscal Years 1990 Through First 6 Months of 1994 Type of representa tion FPDs CDOs PAs FPDs CDOs PAs FPDs CDOs PAs FPDs CDOs PAs FPDs CDOs PAs ---------- ------ ------ ------ ------ ------ ------ -------- -------- -------- -------- -------- -------- -------- -------- -------- Criminal 15,326 8,417 23,782 17,068 7,592 28,119 17,749 6,885 28,555 19,220 6,450 29,237 8,891 3,224 13,485 Appeals 1,150 241 1,825 1,411 335 3,699 1,549 269 3,225 1,825 324 3,977 943 134 2,000 Other 6,747 3,048 2,959 7,163 3,578 4,088 8,007 3,881 4,543 8,667 3,525 4,545 3,932 1,592 2,395 Death NA NA 9 NA NA 121 NA NA 156 NA NA 246 NA NA 171 penalty ==================================================================================================================================================== Total 23,223 11,706 28,575 25,642 11,505 36,027 27,305 11,035 36,479 29,712 10,299 38,005 13,766 4,950 18,051 ---------------------------------------------------------------------------------------------------------------------------------------------------- Legend: FPDs = Federal Public Defenders CDOs = Community Defender Organizations PAs = Panel Attorneys NA = not applicable Source: GAO analysis of AOUSC data. Table II.3 Comparison of Average In-Court Hours by Type of Representation Closed for Federal Public Defender Organizations, Community Defender Organizations, and Panel Attorneys, Fiscal Years 1990 Through First 6 Months of 1994 1st Type of half 1st half 1st half representation 1990 1991 1992 1993 1994 1990 1991 1992 1993 1994 1990 1991 1992 1993 1994 ------------------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------------ ------ ------ ------ ------ ----------- Criminal 3.0 3.4 3.9 3.8 3.6 2.1 2.6 3.1 3.4 3.0 8.1 9.2 9.2 8.9 8.4 Appeals 0.5 0.5 0.5 0.8 0.7 0.8 0.7 0.7 0.7 0.6 0.9 0.7 0.8 0.7 0.7 Other 1.0 1.0 1.3 1.2 1.4 0.9 0.8 0.9 0.9 0.9 1.7 1.6 1.5 1.5 1.6 Overall average 2.3 2.6 2.9 2.8 2.8 1.7 2.0 2.3 2.4 2.3 7.0 7.5 7.6 7.2 6.8 ----------------------------------------------------------------------------------------------------------------------------------------------------- Source: GAO analysis of AOUSC data. Table II.4 Criminal Representations Closed by Federal Defender Organizations and Panel Attorneys, Fiscal Years 1990 Through First 6 Months of 1994 Ty pe of of fe ns e FDOs PAs FDOs PAs FDOs PAs FDOs PAs FDOs PAs FDOs PAs -- ---- ---- ---- ---- ---- ---- ---- ---- -------- ---- -------- ---- We 982 887 1,32 1,36 1,51 1,63 1,57 1,70 801 741 61% 92% a 1 5 7 1 7 2 p o n s Im 4,34 1,76 3,88 2,11 3,01 1,71 3,32 1,83 1,748 905 -23 4 m 2 5 9 0 9 0 9 4 i g r a t i o n Dr 5,82 9,44 6,30 11,8 7,22 12,5 7,61 13,6 3,296 6,20 31 45 u 8 3 1 83 4 89 1 59 1 g s Fr 2,85 1,78 2,94 2,35 2,87 2,56 3,07 2,81 1,590 1,45 8 57 a 5 9 0 4 9 3 9 4 9 u d Ot 9,73 9,89 10,2 10,4 9,99 10,0 10,0 9,22 4,680 4,17 3 -7 h 6 8 09 07 5 62 74 8 9 e r ================================================================================== To 23,7 23,7 24,6 28,1 24,6 28,5 25,6 29,2 12,115 13,4 8% 23% t 43 82 60 19 34 55 70 37 85 a l ---------------------------------------------------------------------------------- Legend: FDOs = Federal Defender Organizations PAs = Panel Attorneys \a Percent change based upon fiscal year 1990-1993 data only. Source: GAO analysis of AOUSC data. Table II.5 Comparison of Average In-Court Hours for Federal Public Defender Organizations, Community Defender Organizations, and Panel Attorneys by Major Offense Category in Criminal Representations Closed, Fiscal Years 1990 Through First 6 Months of 1994 Type of criminal representatio 1st half 1st half 1st half n 1990 1991 1992 1993 1994 1990 1991 1992 1993 1994 1990 1991 1992 1993 1994 ------------- ------ ------ ------ ------ ----------- ------ ------ ------ ------ ------------ ------ ------ ------ ------ ----------- Weapons 3.8 4.8 6.0 4.9 4.7 4.0 4.3 5.2 5.2 4.6 4.9 6.1 5.4 5.9 6.5 Immigration 1.7 2.0 2.5 2.5 2.5 0.5 0.7 0.9 0.9 0.9 3.1 2.8 3.4 3.2 2.6 Drugs 4.2 4.8 5.1 4.9 5.1 3.7 4.7 4.4 4.7 4.4 12.2 12.7 11.6 11.0 10.6 Fraud 3.2 3.6 4.6 4.6 4.2 2.6 3.3 3.1 2.7 2.7 7.0 9.8 11.7 11.0 10.4 Other\a 2.5 2.7 2.9 2.9 2.7 2.1 2.6 3.4 4.3 3.8 5.7 6.7 7.2 6.9 5.9 Average, all 3.0 3.4 3.9 3.8 3.6 2.1 2.6 3.1 3.4 3.0 8.1 9.2 9.2 8.9 8.4 offenses ----------------------------------------------------------------------------------------------------------------------------------------------------- \a Panel attorney data excludes hours for death penalty representations. Source: GAO analysis of AOUSC data. Table II.6 Average Cost per Representation Closed by Federal Public Defender Organizations, Community Defender Organizations, and Panel Attorneys, Fiscal Years 1990 Through First 6 Months of 1994 Average Number Average Number Average Number cost Number Average Number Average Number Average Type of attorney closed cost each closed cost each closed each closed cost each closed cost each closed cost each ------------------ ------ ---------- ------ ---------- -------- ------------ -------- ---------- -------- ---------- -------- ----------- FPDs 23,223 $1,647 25,642 $1,869 27,305 $2,380 29,712 $2,652 13,766 NA 27.9% 61.0% CDOs 11,706 $907 11,505 $1,107 11,035 $1,618 10,299 $2,123 4,950 NA -12.0 \134.1 Panel attorneys ----------------------------------------------------------------------------------------------------------------------------------------------------- Including death 28,575 $1,946 36,027 $2,383 36,479 $2,595 38,005 $2,746 18,051 $2,790 33.0% 41.1% penalty representations Excluding death 28,566 $1,936 35,906 $2,323 36,323 $2,458 37,759 $2,523 17,880 $2,638 32.2% 30.3% penalty representations ----------------------------------------------------------------------------------------------------------------------------------------------------- Legend: FDOs=Federal Defender Organizations CDOs=Community Defender Organizations NA=not available \a Percent change based upon fiscal year 1990-1993 data only. Source: GAO analysis of AOUSC data. Table II.7 Number of Panel Attorney Representations Closed and Average Cost per Representation Closed by Type of Representation, Fiscal Years 1990 Through First 6 Months of 1994 Type of Number Average Number Average Number Average Number Average Number Average Number Average representation closed cost each closed cost each closed cost each closed cost each closed cost each closed cost each ------------------ ------- --------- ------- --------- --------- ---------- --------- ---------- --------- ---------- -------- ---------- Criminal 23,782 $1,985 28,119 $2,385 28,555 $2,584 29,237 $2,641 13,485 $2,647 22.9% 33.0% Appeals 1,825 2,567 3,699 3,049 3,225 3,221 3,977 3,187 2,000 2,986 117.9 24.2 Other 2,959 1,148 4,088 1,245 4,543 1,125 4,545 1,183 2,395 1,334 53.6 3.0 Death penalty 9 33,254 121 20,206 156 34,607 246 36,878 171 32,222 2,633.3 10.9 ===================================================================================================================================================== Total 28,575 $1,946 36,027 $2,383 36,479 $2,595 38,005 $2,746 18,051 $2,790 33.0 41.1 ----------------------------------------------------------------------------------------------------------------------------------------------------- \a Percentages calculated using 1993 data because 1994 data include only the first 6 months of the fiscal year. Source: GAO analysis of AOUSC data. Table II.8 Number of Panel Attorney Criminal Representations Closed, Total Costs, and Average Cost per Representation Closed by Major Offense Category, Fiscal Years 1990 Through 1993 Total Average Total Averag Total Averag Total Average Major offense category number Total cost cost number Total cost e cost number Total cost e cost number Total cost cost ----------------------- -------- ------------- -------- ------ ------------ ------ ------ ------------ ------ ------ ------------ ------- Weapons 887 $1,339,719 $1,510 1,365 $2,726,530 $1,997 1,631 $3,138,966 $1,925 1,702 $3,507,405 $2,061 Immigration 1,765 1,435,092 813 2,110 1,736,657 823 1,710 1,774,858 1,038 1,834 1,862,166 1,015 Drugs 9,443 25,306,561 2,680 11,883 37,111,706 3,123 12,589 $40,763,518 3,238 13,659 43,052,951 3,152 Fraud 1,789 3,826,579 2,139 2,354 6,113,566 2,597 2,563 8,365,042 3,264 2,814 9,551,100 3,394 Other criminal 9,898 15,304,845 1,546 10,407 19,371,328 1,861 10,062 19,732,877 1,961 9,228 19,248,396 2,086 ===================================================================================================================================================== Total 23,782 $47,212,796 $1,985 28,119 $67,059,787 $2,385 28,555 $73,775,261 $2,584 29,237 $77,222,018 $2,641 ----------------------------------------------------------------------------------------------------------------------------------------------------- Source: GAO analysis of AOUSC data. Table II.9 Districts in Which Higher Standard Panel Attorney Rates Have Been Implemented as of January 1, 1990 Circuit/district $60 $70 $75 ------------------------------ -------- -------- -------- District of Columbia x Second Circuit New York, Southern x New York, Eastern x Third Circuit New Jersey x Sixth Circuit Michigan, Eastern (Detroit) x Ninth Circuit Alaska x Arizona (Phoenix and Tucson) x California, Northern x California, Central x California, Eastern x (Sacramento and Fresno) California, Southern x Hawaii x Nevada (Las Vegas and Reno) x Oregon\ x Washington, Western (Seattle) x Tenth Circuit New Mexico (Las Cruces) x ------------------------------------------------------------ \a The Judicial Conference has authorized all locations in these districts to pay the $75 per hour rate subject to the availability of funds. Source: AOUSC. Table II.10 Estimated Additional Cost of Paying Panel Attorneys Higher Standard Rates in Sixteen Districts (Based on Amounts Approved for Payment)\a Fiscal year Fiscal year Fiscal year First half Compensation rate 1991 1992 1993 1994 ----------------- -------------- -------------- -------------- ------------- In-court compensation -------------------------------------------------------------------------------- Higher rate\b $5,268,998 $6,327,422 $5,834,069 $2,611,307 Lower rate 4,457,413 5,334,113 5,046,650 2,249,896 ($60/hour)\c Difference 811,585 993,309 787,419 361,411 Out-of-court compensation -------------------------------------------------------------------------------- Higher rate\b 21,429,473 26,494,753 27,272,108 13,520,886 Lower rate 12,420,513 15,566,189 16,266,000 8,060,878 ($40/hour)\c Difference 9,008,960 10,928,564 11,006,108 5,460,008 ================================================================================ Total $9,820,545 $11,921,873 $11,793,527 $5,821,419 difference -------------------------------------------------------------------------------- \a Amounts in this table exclude most appeals representations. For about 98 percent of all appeals representations in fiscal years 1990 through 1993 and 85 percent in the first half of fiscal year 1994, the database does not contain an identifier for the district from which the case was appealed. \b Totals reflect any court reductions in the amount of compensation requested. \c Totals at lower rate assume compensation for total number of hours requested. Because the database did not reflect the number of hours by which requests were reduced, only dollar amounts of the reductions, we could not correspondingly reduce the amount that would have been approved at the lower rates. Source: GAO analysis of AOUSC data. Table II.11 Death Penalty Resource Centers, End of Fiscal Year 1993 Federal judicial Resource center Location districts served ------------------------------------------- ------------ --------------------- Alabama Capital Representation Resource Montgomery AL (N, M, S)\a Center Arizona Capital Representation Project Tempe AZ Arkansas Death Penalty Resource Center Little Rock AR (E, W) California Appellate Project San CA (N, E, C, S) Francisco Volunteer Lawyers' Resource Center, Inc. Tallahassee FL (N, M, S) Georgia Appellate Practice and Educational Atlanta GA (N, M, S) Resource Center, Inc. Capital Resource Center Evanston IL ( N, C, S) Kentucky Capital Litigation Resource Center Frankfort KY (E, W) Loyola Death Penalty Resource Center New Orleans LA (E, M, W) Mississippi Capital Defense Resource Center Jackson MS (N, S) Missouri Capital Punishment Resource Center Kansas City MO (E, W) Nevada Appellate and Post-Conviction Las Vegas NV Project, Inc. North Carolina Death Penalty Resource Raleigh NC (E, M, W) Center Ohio Death Penalty Resource Center Columbus OH (N, S) Capital Post-Conviction Project of the Norman OK (E, N, W) Oklahoma Appellate Public Defender System South Carolina Death Penalty Resource Columbia SC Center Capital Case Representation Resource Center Nashville TN (E, M, W) of Tennessee, Inc. Texas Appellate Practice and Educational Austin TX (N, E, W, S) Resource Center Virginia Post-Conviction Assistance Project Richmond VA (E, W) -------------------------------------------------------------------------------- \a Letters indicate districts as follows: E = eastern; W = western; S = southern; N = northern; M = middle; C = central. Source: AOUSC. (See figure in printed edition.)Appendix III COMMENTS FROM THE ADMINISTRATIVE OFFICE OF THE U.S. COURTS ========================================================== Appendix II (See figure in printed edition.) (See figure in printed edition.) MAJOR CONTRIBUTORS TO THIS REPORT ========================================================== Appendix IV GENERAL GOVERNMENT DIVISION, WASHINGTON, D.C. William O. Jenkins, Jr., Assistant Director, Administration of Justice Issues Doris Page, Senior Evaluator M. Grace Haskins, Senior Evaluator David Alexander, Senior Social Science Analyst BOSTON/NEW YORK FIELD OFFICE Rudy Plessing, Core Group Manager Amy Hutner, Senior Evaluator Lucine Moore-Willis, Evaluator DETROIT FIELD OFFICE Kelly Haggard, Senior Evaluator Keith Landrum, Evaluator William G. Sievert, Technical Assistance Group Manager OFFICE OF THE GENERAL COUNSEL Jan Montgomery, Assistant General Counsel