Review of the United States Marshals Service's
Apprehension of Violent Fugitives
Evaluation and Inspections Report I-2005-008
Office of the Inspector General
We examined the performance of Deputy Marshals working individually or as members of warrant squads, district fugitive task forces, or Regional Task Forces in apprehending federal and state violent fugitives from Fiscal Year 2001 to FY 2004, as well as the reasons for any differences in performance. We did not separately analyze the performance of the proposed RTFs listed in the Department’s October 2004 Fugitive Apprehension Report because the Assistant Director for the Investigative Services Division told us that he would eventually like all 94 districts to be included in RTFs. Therefore, we considered all districts currently not part of an RTF as a single group.
We conducted interviews with officials in United States Marshals Service headquarters – the Assistant Director and other personnel within the Investigative Services Division, as well as personnel within the Analytical Support Unit and the Management and Budget Division. To examine the differences in the various approaches to fugitive apprehension in the district offices, we visited 12 of the 15 districts that are part of the 5 RTFs, as well as 5 districts that are not part of an RTF. In each district, we interviewed the Marshal and Chief Deputy Marshal (when available), warrant squad supervisor (when available), Task Force Supervisory Inspector and Inspectors (if applicable), Deputy Marshals involved with the Electronic Surveillance Unit and the Financial Surveillance Unit, the WIN data administrator, and Deputy Marshals involved in fugitive investigations. We also interviewed representatives from other federal, state, and local agencies who worked with the districts and RTFs to investigate and apprehend fugitives. The districts and RTF offices we visited were:
New York/New Jersey RTF:
Pacific Southwest RTF:
Great Lakes RTF:
Capital Area RTF:
We also conducted site visits in the following five districts:
Win Analysis. To analyze the USMS fugitive apprehension program from FY 2001 through FY 2004, we examined data from the Warrant Information Network database. We examined all warrants that were closed on or after September 30, 2000, and that were received on or before October 1, 2004. According to the Department’s FY 2004 Performance Report, the USMS compares information contained in WIN to a random sample of the Federal Bureau of Investigation’s National Crime Information Center records to verify the accuracy of the information. USMS headquarters also coordinates with districts to verify that warrants are validated against the signed paper records. In addition to the internal check of the WIN data, the OIG conducted an audit of WIN in November 2002 in accordance with the Government Information Security Reform Act.25 This audit assessed the management, technical, and operational controls that protected WIN data from unauthorized use, loss, or modification. The OIG found significant vulnerabilities and made 28 recommendations to improve the security of the WIN database. Currently, the one recommendation that is still open is that the USMS enforce "Department-wide identification and authentication policies and ensure that only authorized personnel can login to the system."
Staff Year Analysis. To analyze the time spent by the USMS on fugitive investigations, we reviewed time and attendance data recorded by USMS personnel on the standard USM-7 form used by all USMS personnel to record the time spent on USMS duties, such as fugitive investigation, court operations, and prisoner transportation. Personnel assign their time to specific project codes. For example, if a Deputy Marshal investigated a federal felony fugitive for 3 hours, the time would be recorded under the appropriate federal felony project code.
We requested data from the USM-7 database for all fugitive investigation time for FY 2001 through FY 2004. However, the USMS could provide data only from FY 2002 through FY 2004. We also received a USM 7 project code guide, which contained a list of project code definitions. For the purposes of our review, we used data entered under project codes related to federal felony warrants, federal non-felony warrants, state warrants, RTF assignments, and other administrative functions. We eliminated project codes that were not relevant to our review or that were not related to fugitive investigations (such as project codes related to court security services). Besides time recorded by personnel in the 94 districts, the USM-7 database captures time recorded by USMS headquarters personnel as spent on fugitive investigations. We allocated this headquarters time to each of the districts based on the proportion of the district’s fugitive apprehension efforts to the overall USMS fugitive apprehension efforts.
We calculated staff years by dividing the number of hours worked in a particular district or grouping of districts by 2,087 hours (the Office of Personnel Management equivalent of one "full time equivalent" work year). We did not calculate staff years allocated for state or federal fugitive investigations separately because the USM-7 data does not distinguish whether the time spent by Deputy Marshals on an RTF was for a federal or state fugitive investigation.
We did not test the validity of the USMS time and attendance records. However, the USMS internally verifies USM-7 data each year, using a USMS budget model to identify anything out of the ordinary reported by the districts. The Management and Budget Division relies on district supervisors to verify the accuracy of the USM-7 forms before forwarding them to headquarters, and the Office of Inspections also conducts random compliance reviews to ensure that personnel comply with USM-7 policy, such as obtaining the proper supervisory signatures. No external audits have been conducted on USM-7 data.
Electronic Surveillance Unit Database Analysis. To analyze the involvement of Electronic Surveillance Unit personnel in fugitive investigations, we reviewed the database that tracks the unit’s assistance with fugitive investigations and types of surveillance operations. We examined the number of fugitive investigations in which Electronic Surveillance Unit personnel assisted, as well as the number of arrests that resulted from this assistance. We were unable to determine from the database whether records pertained to federal or state fugitive investigations, or whether the investigations originated from a district or a task force. In October 2004, the Electronic Surveillance Unit added new variables to its database to track who requested assistance and whether the investigation involved a federal or state fugitive.
Electronic Surveillance Unit personnel stated that they conducted an annual check of WIN to verify the Electronic Surveillance Unit database numbers and to validate the number of arrests and arrest dates. We did not test the validity of the database.
District Survey. To learn more about each district’s fugitive operations, we surveyed each of the 94 USMS district offices. We received responses from 88 of the 94 (93.6 percent) district offices.
The survey contained 19 questions regarding the following:
We used the responses from the survey to further understand the fugitive apprehension process, to verify information gathered during our interviews, and to increase our understanding about districts that we did not visit.
In WIN, fugitive warrants are divided into general categories, as shown in Table 4. In our analysis, one warrant equals one fugitive, even if one fugitive had multiple warrants.
Table 4: Category and Types of Federal Warrants
Federal Felony Fugitives. We defined federal felony fugitives as those with Class I or Class II felony warrants.
Federal Non-Felony Fugitives. We defined federal non-felony fugitives as those with Class II non-felony warrants.
Violent Federal Fugitives. We defined violent federal fugitives as those fugitives who were designated in WIN as a 15 Most Wanted, Major Case, or Category 1 fugitive. (A Category 1 fugitive is one whose underlying or current offense is homicide, kidnapping, assault, aggravated assault, robbery, arson, an "over-the-wall" escape, or involves weapons or explosives.) Violent federal fugitives can be either Class I or Class II.
State Fugitives. We defined state fugitives as those with a state or local warrant. Some local jurisdictions, most notably the cities of Chicago and New York, conduct state fugitive investigations based on probable cause rather than a warrant. The USMS Deputy Marshals sometimes participated in these investigations, but we included these fugitive investigations only if the USMS entered a subsequently obtained warrant in WIN.
Violent State Fugitives. We defined violent state fugitives as those fugitives whose underlying or current offense was one used to designate a federal fugitive as a Category 1 fugitive. The offenses include homicide, kidnapping, assault, aggravated assault, robbery, arson, an over-the-wall escape, or an offense involving weapons or explosives.
Fugitives Apprehended. We defined fugitives apprehended as ones who were physically arrested by the USMS, who were arrested by another agency under the direction of the USMS (directed arrest), or who surrendered to the USMS or another agency.
Fugitives Cleared. We defined fugitives cleared as ones whose warrants were closed by means other than apprehension by the USMS, as defined above. This includes warrants cleared through arrests by other agencies, lodged detainers, dismissals, the return of the warrants to the originating agency, or the purging of the warrants.
Clearances or Apprehensions by RTFs. Although WIN contains a field that captures whether a district task force or RTF apprehended a fugitive, we learned that USMS personnel did not consistently enter data in this field when a task force was involved. Therefore, we analyzed RTF performance based on the districts that constituted the RTFs as a whole (for example, the Pacific Southwest RTF comprises the Southern and Central districts of California). Any analysis regarding fugitives cleared or apprehended is based on the performance of the districts participating in the RTFs because we could not analyze individual RTFs’ performance separately from that of the districts in which they operated.
Fugitives at Large. We defined fugitives at large in a particular fiscal year as those fugitives in WIN whose warrants were open at the end of the fiscal year, regardless of when the warrant was opened.
Apprehensions per Staff Year. We divided the number of apprehensions of fugitives by the staff years applied to fugitive apprehension by a given group of districts. For example, for federal fugitives, we divided federal apprehensions in a group of districts in 1 year by total staff years applied to fugitive apprehension in that same group of districts. If a group of districts apprehended 100 federal fugitives and devoted 10 staff years to fugitive apprehension in 1 year, then that group of districts apprehended 10 federal fugitives per staff year. This measure captures the number of fugitives apprehended per staff year.
Average Number of Days for Fugitive Arrest. Prior to FY 2002, the Department measured the performance of the USMS fugitive apprehension program by "the average number of days for fugitive arrest" for Major Case fugitives and described the public benefit as follows: "The quicker a fugitive is captured; the less the public is exposed to further risk of crime." The Department discontinued this measure in FY 2002 because "a warrant can become a major case at any point in its investigation, [so] the age of the warrant can dramatically skew the overall average number of days for fugitive arrest in the major case category." The USMS never applied this now abandoned measure to all violent federal fugitive investigations.
To determine whether the average number of days for fugitive arrest was a useful measure at all, we calculated the average number of days between the issuance of a warrant and the closure of the warrant from FY 2001 to FY 2004 for violent federal fugitives. We found that the average number of days that violent federal fugitives remained at large before apprehension varied from 111 to 113 days, even though half were apprehended within 12 days. Marshals and Deputy Marshals we interviewed told us that they often apprehend the newest fugitives first, or as one RTF Supervisory Inspector stated, "while the trail is still warm." Even though half of the violent federal fugitives were apprehended within 12 days, some remained at large for years. These investigations that span years explain why the average number of days for fugitive arrest is so much larger than the median of 12 days. The average time violent federal fugitives remain at large will not change significantly until the backlog of long-term fugitives is reduced. We agree with the Department’s conclusion that the average number of days for a fugitive arrest is not currently a useful measure of the performance of the USMS fugitive apprehension program and did not include this analysis.