Office of the Inspector General

U.S. Department of Justice

Review of the Federal Bureau of
Prisons’ Untimely Releases of Inmates

Evaluation and Inspections Division 16-03

May 2016

EXECUTIVE SUMMARY
Following news reports that the Federal Bureau of Prisons (BOP) had confined
an inmate for 13 months past his correct release date, the Department of Justice
(Department) Office of the Inspector General (OIG) initiated an examination of the
BOP’s process for ensuring federal inmates are released on their correct release
dates and the incidences of releases before or after the correct release date due to
staff error between 2009 and 2014. We found that of the 461,966 inmate releases
between 2009 and 2014, the BOP categorized 157 as untimely due to staff error.
We also learned that the BOP classifies a far greater number — 4,183 — as
untimely for other reasons.
According to the BOP, the vast majority of non-staff error “untimely” releases
were due to situations that are beyond its control, such as amended sentences that
result in shorter sentences than the time an inmate had already served. Also, data
and information we reviewed indicates that other entities inside and outside the
Department may sometimes contribute to untimely releases. Although BOP officials
told us that it was highly unlikely that staff error on the part of a Department entity
contributed to any of the 4,183 cases, they could not rule out the possibility and we
found that the BOP does not always have complete information about the
circumstances of untimely releases to which other entities contribute. We therefore
concluded that the Department should work with all relevant entities, both within
and outside the Department, to review the full range of possible reasons for
untimely releases and how to address those that are in any way preventable.
With regard to the 157 untimely releases that BOP categorized as due to staff
error, 152 were late releases and 5 were early releases. We found that three of the
late releases and three of the early releases involved an error resulting in more
than 1 year of over- or under-served time by the inmate. Table 1 displays the
number of days of over- or under-served time for the 157 untimely releases.
Table 1
Number of Untimely Releases from Prison Caused by Staff Errors,
By Days Over- or Under-served,
2009 – 2014
Days Over or Under Served 2009–2014 Percent of Total
LATE RELEASE

EARLY RELEASE

0 to 30 Days

91

60%

31 Days to 1 Year

58

38%

1+ Years

3

2%

0 to 30 Days

1

20%

1

20%

31 Days to 1 Year

j

7

l

1+ Years
3
60%
Source: OIG analysis of untimely releases identified by BOP as due to staff error,
calendar years 2009 – 2014

While the 157 untimely releases due to staff error was rare compared to the
461,966 releases by BOP during the 6-year period of our review (an error rate of

i

0.03 percent), the consequences of an untimely release can be extraordinarily
serious. Late releases from prison deprive inmates of their liberty, while early
releases can put communities at risk if the inmates are dangerous. Early releases
also can harm an inmate and the inmate’s family, particularly if the inmate’s efforts
to gain employment and reestablish ties with the community are interrupted by a
re-arrest for the purpose of completing the sentence.
Late releases also are costly: in addition to BOP costs associated with the
unauthorized period of incarceration, there is the potential for significant
compensatory judgments to those inmates who suffered an unconstitutional
deprivation of their liberty. For the 152 late releases, we estimated the total cost
to the BOP, exclusive of litigation and settlement costs, to be approximately
$669,814. In addition, between 2009 and 2015, the Department settled four
lawsuits by inmates alleging untimely release, one for $90,000; another for
$120,000; another for $295,000; and the fourth for $175,000. This does not
include additional costs the Department incurred as a result of these cases, such as
salary costs expended to handle the lawsuits.
Additionally, untimely releases, whether early or late, contravene judicial
sentencing orders, yet the BOP does not have in place a process to consider
whether to notify the sentencing court of an untimely release. We found that, for
late releases, the BOP notifies the relevant U.S. Probation office but does not
separately notify the sentencing court or the U.S. Attorney’s Office that handled the
case. For early releases, we found that BOP policy requires Wardens to notify “the
appropriate Judicial Official(s)” when an inmate who is deemed to be a threat to the
community is released early and when any inmate is released more than 30 days
early. Despite this policy, we found no such notifications occurred for the four
untimely early releases of 30 days or more that we reviewed.
We found that 127 of the 157 untimely releases due to staff error were the
result of errors made at BOP’s Designation and Sentence Computation Center
(DSCC). (The DSCC performs the vast majority of sentence computations the BOP
uses to determine an inmate’s release date.) The other 30 untimely releases were
the result of staff errors at non-DSCC entities, such as BOP institutions, BOP
Residential Reentry Management field offices, Residential Reentry Centers
(previously known as Community Corrections Centers), and private contract
prisons.
Prior to 2005, when the BOP consolidated sentence computation functions in
the DSCC, individual BOP facilities across the country had performed sentence
computation. The BOP transitioned sentence computations to the DSCC in order to
consolidate BOP sentence computation functions, reduce costs, and ensure
consistent application of laws and BOP policies. This change also helped reduce the
number of untimely releases caused by staff error. Based on BOP data, we
determined that during the 6-year period from 1999 and 2004 there were
approximately 344 untimely releases due to staff error (around 0.1 percent of all
releases) compared to 157 (around 0.03 percent of all releases) between 2009 and
2014.

ii

We concluded that the most common sentence computation errors resulted
from incorrect application of jail credit, incorrect determinations of primary
jurisdiction between federal and state custody, and errors relating to concurrent
versus consecutive sentences for defendants with multiple unexpired incarceration
sentences. We also found that poor communication with outside entities —
including local jails, courthouses, state departments of corrections, Native American
reservations, the U.S. Marshals Service, and others — resulted in DSCC staff not
obtaining complete and accurate sentencing information or interpreting sentencing
information incorrectly, leading to untimely releases.
The BOP has several processes in place to prevent untimely release. One is
to conduct a final release audit 12 months prior to an inmate’s scheduled release
date to discover and correct any errors and to have the ability to account for any
reduced sentence time. However, we found that in some cases 12 months is not
early enough to discover and correct the error and prevent the untimely release.
Specifically, 14 of the 19 errors discovered by final release audits might not have
resulted in untimely releases had the audits occurred 18 months before the
inmate’s scheduled release date. Had the final release audit occurred 24 months
prior to the release date, an additional 4 of the 19 errors might have been
discovered in time to prevent the untimely release. Therefore, we recommend that
BOP explore and implement additional sentence calculation processing or auditing
strategies, taking into account that conducting final release audits only 12 months
before an inmate’s projected release date has led to preventable untimely releases.
We also noted that the DSCC holds education events that bring together
officials from the DSCC as well as outside the BOP (including U.S. District Court
Judges, Federal Public Defenders, and U.S. Probation Office staff) to inform them
about topics relating to DSCC operations, such as sentence computations. In our
judgment, these events could help prevent untimely releases if the BOP included
factors that can affect its ability to release inmates on time, including showing
attendees how conflicting information can lead to incorrect sentence computations.
This report contains three case studies profiling inmate releases that were
more than 1 year late. The report also makes seven recommendations for the BOP
to help reduce untimely releases due to staff error.

iii

TABLE OF CONTENTS
INTRODUCTION ............................................................................................ 1
Background ........................................................................................ 1
The BOP’s Processes for Computing and Verifying Inmate Sentences........... 1
Documentation of Untimely Releases ...................................................... 3
Reviews of the DSCC............................................................................ 4
Scope of the Review............................................................................. 4
RESULTS OF THE REVIEW .............................................................................. 5
The BOP Has Limited Information about the Causes of Releases
Considered Untimely That Are Not Attributed to BOP Staff Error ................. 5
Untimely Releases Due to Staff Error, While Rare in the Context
of All Inmate Releases, Occurred Over 150 Times in a Recent
6-Year Period, Resulting in Significant Harm ............................................ 8
Early Releases Result in Costs to the Department and the Taxpayer,
as well as Potential Harm to Inmates, Their Families, and
Their Communities............................................................................. 13
The Causes of Untimely Releases Vary .................................................. 13
Case Studies of Releases Over 1 Year Late ............................................ 22
CONCLUSION AND RECOMMENDATIONS ........................................................ 25
Conclusion ........................................................................................ 25
Recommendations ............................................................................. 25
APPENDIX 1: METHODOLOGY OF THE OIG REVIEW ........................................ 27
APPENDIX 2: BOP REVIEWS OF THE DESIGNATION AND SENTENCE
COMPUTATION CENTER ........................................................... 28
APPENDIX 3: THE ODAG AND BOP’S RESPONSE TO THE DRAFT REPORT ........... 29
APPENDIX 4: OIG ANALYSIS OF THE ODAG AND BOP’s RESPONSE ................... 32

INTRODUCTION
Background
Following news reports in late 2014 about an inmate’s lawsuit against the
Federal Bureau of Prisons (BOP) for confining him 13 months beyond his correct
release date, the Department of Justice (Department) Office of the Inspector
General (OIG) initiated a review of the BOP’s processes for ensuring inmates are
released on their correct release date.1 We found that of the 461,966 inmate
releases between 2009 and 2014, the BOP categorized 157 as “untimely” due to
staff error and a far greater number — 4,183 — as untimely for other reasons. As
discussed in this report, according to the BOP, the vast majority of non-staff error
untimely releases were due to situations beyond its control, such as amended
sentences that result in shorter sentences than the time an inmate had already
served.
Because the incident giving rise to this review resulted from BOP staff error,
we focused on analyzing untimely releases that the BOP attributed to staff error.
The BOP identified 157 such cases, including 148 cases attributable entirely to the
BOP or others under its direct control, such as employees of private contract
prisons. The remaining nine cases were due to a combination of factors including
BOP staff error, and we discuss these later in the report.
The BOP’s Processes for Computing and Verifying Inmate Sentences
Each year, the BOP conducts approximately 153,309 initial computations and
91,475 updates to existing computations to determine inmate release dates, and
the Designation and Sentence Computation Center (DSCC), located in Grand
Prairie, Texas, performs the vast majority of them. The DSCC, which consists of
approximately 330 employees, was created in 2005 to consolidate all BOP sentence
computation functions into one office to reduce costs and to ensure consistent
application of laws and BOP policies. Prior to 2005, individual BOP facilities across
the country had performed sentence computations. Between 2005 and 2010, the
BOP phased out sentence computation functions at BOP-operated facilities and
transferred these functions to the DSCC.2
1
Paul McEnroe, “Lost in Federal Prison – and Held 13 Months Too Long,” Star Tribune,
October 11, 2014, http://www.startribune.com/lost-in-federal-prison-and-held-13-months-toolong/278868951/ (accessed March 9, 2016). KTTC staff, “Man Kept Imprisoned 13 Months Past His
Release Seeks Lawsuit,” KTTC, October 13, 2014,
http://www.kttc.com/story/26776178/2014/10/13/man-kept-imprisoned-13-months-past-his-releaseseeks-lawsuit (accessed March 9, 2016). Associated Press, “Man Held 13 Months Too Long Sues
Prison System,” Washington Times, October 11, 2014,
http://www.washingtontimes.com/news/2014/oct/11/man-held-13-months-too-long-sues-prisonsystem/ (accessed March 9, 2016). The BOP settled the inmate’s civil suit for $175,000.
2

Due to existing contracts, several private contract institutions continued to perform
sentence computation functions after 2010. Currently, there are four private contract institutions that
continue to perform sentence computations: Big Spring Correctional Center, Big Spring, Texas
(contract ends March 2017); Reeves County Detention Center I/II, Pecos, Texas (contract ends

Continued

1

The DSCC divides sentence computation responsibilities among teams that
compute the sentences for federal inmates sentenced in 1 or more of the 94 federal
judicial districts. Within each team, Classification and Computation Technicians and
Specialists (CC staff) perform the day-to-day sentence computation functions of the
DSCC, including calculating inmate release dates.3 As shown in the Figure, these
functions are divided into three main steps (two for sentences less than 13 months)
and take place at different phases of an inmate’s incarceration. In addition to these
steps, sentence computation updates are performed when a change to an inmate’s
sentence occurs after the initial computation and certification audit.
Figure
Sentence Computation Steps
Sentences Less
Than 13 Months

New Information
Affecting an
Inmate’s Sentence

Inmate
Requiring a
Sentence
Computation

Initial
Computation
(upon being sentenced)

Sentences Greater
Than 13 Months

Updates to
Computation

Certification Audit*
(Immediately after initial computation)

Final Release Audit*
(12 months before release)

Inmate Released

* The certification and final release audits follow the same procedure except that they ordinarily occur
at two different stages of the inmate’s incarceration.
Source: OIG analysis of sentence computations functions at the DSCC

Initial Sentence Computation
The CC staff calculates an inmate’s release date at the beginning of the
inmate’s incarceration. During initial computations, the CC staff reviews a variety
of source documents and SENTRY, the BOP’s primary database for storing inmate
information, to calculate the inmate’s projected release date.4 If any necessary
January 2017); Reeves County Detention Center III, Pecos, Texas (contract ends December 2016);
and Taft Correctional Institution, Taft, California (contract ends August 2017). These institutions will
transfer their sentence computation responsibilities to the DSCC when their current contracts end, the
last of which will be in 2017.
3

Classification and Computation Specialists and Technicians perform many of the same
functions, but Specialists have more experience than Technicians, often answer the Technicians’
questions regarding sentence computation, and can conduct the full range of sentence computation
tasks. The DSCC provides CC staff basic and advanced training, on-the-job training, and quarterly
training.
4
Source documents include: a Judgment and Commitment Order describing the term of
imprisonment and any other punishment the inmate faces; a Prisoner Custody, Detention, and
Disposition Record (USM-129) from the U.S. Marshals Service (USMS), which notes where the USMS
has moved the inmate after sentencing and before the BOP assigns the inmate to a facility and
includes the inmate’s time in custody during this period; and a pre-sentence report from the U.S.
Probation Office in the sentencing jurisdiction, which usually includes the date the inmate was arrested
and what dates the inmate was held in custody prior to sentencing. Other source documents that may

Continued

2

source documents are missing or incomplete, the responsible staff member will try
to obtain them by contacting the appropriate state, local, federal, or other entity.
Certification Audit
The CC staff performs a certification audit after an initial computation and
after computation updates to review the work of the staff member who conducted
the computation, analyzing the same source documents. As an almost immediate
check on the work of the staff member who conducted the initial computation or
update, the staff member who performs a certification audit must be different from
the staff member who performed the initial sentence computation. If information is
missing, the staff member conducting the audit will typically direct the staff
member who conducted the initial computation (or his or her replacement) to
obtain the information.
Final Release Audit
Performed 12 months before an inmate’s release date, the final release audit
is designed to ensure that the sentence computation is correct, for example, that all
updates have been incorporated, so that the inmate is released on his or her
correct release date.5 Because CC Specialists are more experienced than
Technicians, only CC Specialists conduct final release audits. This audit is a final
check on the initial computation and the certification audit, as well as any
computation updates. The staff member conducting this audit will review the same
source documents as those reviewed for the initial computation and certification
audit. Similar to the certification audit, if necessary documentation is missing, the
staff member conducting the final release audit will typically direct the staff
member who conducted the initial computation to obtain the documentation.
Documentation of Untimely Releases
The BOP uses two forms to document untimely releases: the Untimely
Release Notification (BP 558) and the Late Release Notice for United States
Probation Service (BP 623). When an untimely release occurs, the BOP requires an
employee from the office where the error occurred to complete a BP 558 form to
document the untimely release, including an explanation of the reasons it happened
and what has been done to make sure the mistake is not made again. If an inmate
is subject to supervised release following a sentence from which the inmate was
released late and accumulated over-served time, DSCC staff is also required to
complete a BP 623 form and send it to the U.S. Probation Office. In this way, the

be used are the sentencing court’s statement of reasons for the sentence and the inmate’s National
Crime Information Center report.
5

Final release audits occur 12 months before an inmate’s release to allow any decrease in the
inmate’s sentence that results from an error found by the audit to more likely be absorbed by the
12 months than to result in an untimely release.

3

BOP notifies the U.S. Probation Office that a court may consider the inmate’s overserved time in calculating the prison time for a term of supervised release.6
Reviews of the DSCC
In addition to auditing individual sentence computations, the BOP’s Program
Review Division (PRD) conducts reviews of the DSCC and the DSCC conducts
internal reviews of itself. The PRD conducts program reviews of all BOP programs
in regular cycles where the cycle length depends on the rating the program
received from the previous review. The PRD evaluates several areas of the DSCC,
including sentence computations. The DSCC also conducts internal reviews,
including reviews to assess actions taken to address areas of concern found during
the PRD reviews.
Last, the DSCC conducts monthly audits, which it refers to as “perpetual”
audits, for each sentence computation team. The perpetual audits review
170 (10 per team) randomly selected sentence computations that have undergone
certification audits. The results of these perpetual audits provide information for
quarterly reports and are the basis for in-house training. See Appendix 2 for
additional details about the PRD and DSCC program review process.
Scope of the Review
In this review, we examined untimely releases of BOP inmates between
calendar years 2009 and 2014 that the BOP classified as resulting from staff errors.
Our report focuses on assessing the reasons for the untimely releases due to BOP
staff error, the impact on the inmates, and their associated costs. We interviewed
BOP officials, reviewed public and internal BOP documents, and analyzed data.
During our assessment we also discovered releases that the BOP classified as
untimely, due to a variety of reasons other than staff error, which can involve
entities outside of the BOP. Although an in-depth analysis of the untimely releases
not due to staff error was beyond the scope of this review, we describe the
information we received regarding these as well. We offer recommendations at the
conclusion of this report to assist the BOP and the Department in preventing and
appropriately addressing untimely releases. Appendix 1 provides additional details
about our methodology.

6

Supervised release is a type of post-confinement monitoring overseen by federal district
courts with the assistance of federal Probation Officers. A sentencing court is authorized (in some
cases, required) to impose a supervised release in addition to imprisonment. While on supervised
release after reentry into the community, an offender must abide by certain terms, some mandated by
statute and others imposed at the court’s discretion. Supervised release may be revoked if an
offender violates a term. Upon finding such a violation, the court typically has discretion as to
whether to require the offender to serve all or part of the supervised release term in prison. If an
inmate who was released late and is subject to supervised release violates the terms of supervised
release, the BOP will apply any time the inmate over-served toward the supervised release violation
sentence. See 18 U.S.C. § 3585(b).

4

RESULTS OF THE REVIEW
The BOP Has Limited Information about the Causes of Releases Considered
Untimely That Are Not Attributed to BOP Staff Error
We found that the BOP categorized 157 releases that occurred between 2009
and 2014 as untimely due to staff error, and a far greater number — 4,183 — as
untimely for other reasons (a total of 4,340). BOP officials stated that most, if not
all, of the untimely releases not attributed to staff error are beyond the BOP’s
control, either because an entity other than the BOP was responsible for or
contributed to making the release untimely or because the untimely release was
likely not preventable. For example, a BOP official told us that in many of these
cases an inmate was released following a retroactive action that reduced his or her
sentence to less time than he or she already had served. Under the BOP’s current
definition, any time served beyond the revised sentence is counted as “over-served
time” and any release for which there is over-served time is counted as “untimely.”
We also observed that non-BOP entities can contribute to an untimely release even
in instances where there is also BOP staff error. It was beyond the scope of this
review to fully examine or assess untimely releases not attributed to staff error;
however, we are concerned that because the BOP does not always have complete
information about the circumstances of untimely releases that other entities
contribute to, steps that could be taken to prevent them won’t be identified.
We discussed with BOP officials the circumstances that might result in
untimely releases that are not due to staff error and some of the actions the BOP
takes to address these untimely releases. We found that while the Designation and
Sentence Computation Center (DSCC) reviews every untimely release to determine
whether there was staff error and classifies it into a general category by cause, the
DSCC focuses on identifying actions that could prevent future BOP staff errors,
rather than on understanding all possible contributing factors, including those
outside of the BOP’s direct control.7 For this reason, the BOP does not
systematically obtain the information it would need to determine with certainty the
role other entities may have played. BOP managers told us that collecting this
information would be laborious and that it may not be helpful in preventing
untimely releases because, they believe, the vast majority of untimely releases are
due to some form of technicality, such as the example of a retroactively reduced
sentence provided above. Additionally, BOP managers stated that without
thoroughly re-reviewing the 4,183 untimely releases not due to staff error, they
could not be certain that any given untimely release could not have been prevented
by one of the agencies involved.
Table 2 summarizes the BOP’s classification and explanation of the untimely
releases that occurred between fiscal year (FY) 2009 and 2014 and that it considers
not due to staff error.
7
One BOP manager indicated that any trend, regardless of whether it was due to staff error,
would also be identified during the DSCC’s review process.

5

Table 2
Number of Untimely Releases by Cause, Not Including Staff Error
FYs 2009 – 2014
BOP
Classification

Federal Court
Order/Amended
Judgment

Federal
Ineffectual
Sentence
U.S. Parole
Commission
Decisions
“E-Des” Cases
Submitted Late
by the U.S.
Marshals
Service or U.S.
Probation Office
Untimely
Releases with
No Department
Involvement

Definition/Explanation
An order that was revisited and modified for either a legal or a
judicial reason. Occasionally, these orders will be modified to
include reducing the sentence. If the offender has already
served more time than the new sentence imposes, the inmate
becomes an immediate release and may have over-served time.
An ineffectual sentence occurs when the statutory release date
is earlier than the date the sentence was imposed. This happens
when a court sentences an offender based on Sentencing
Guidelines and imposes terms that do not account for the period
of time that the offender was held in official detention, instead of
imposing a “time served” order.
A U.S. Parole Commission decision may cause an untimely
release when the statutory release date is earlier than the date
of the Commission’s decision.

Number
of
Untimely
Releases

1,874

596

165

This is a relatively new category of untimely release that the
BOP began tracking in the fourth quarter of 2014. It occurs
when the statutory release date is later than the date the
sentence was imposed but earlier than the date the BOP
received all the documents necessary to designate the inmate.
Any outside local or state agency that provides pre-sentence jail
time and new or updated sentencing information after the
inmate is in BOP custody may contribute to the inmate receiving
credit the BOP did not previously provide. On occasion, a state
entity might have its sentence overturned and vacated after the
federal sentence has been computed, and the BOP then has to
adjust the inmate’s creditable time.

TOTAL

49

1,499

4,183

* The statutory release date is the date that a person must be released at the expiration of his
or her term of sentence, less any time deducted for good conduct.
** Designation is the process of determining which BOP institution is to house an inmate. The
BOP is solely responsible for determining where a federal offender will serve his or her sentence.
To make a designation, the DSCC must first receive all sentencing material regarding the
offender, including documents from the sentencing court, the U.S. Probation Office, and the U.S.
Marshals Service (USMS). “E-Designate” is a system whereby the materials needed to designate
inmates are shared between agencies electronically.
*** This table reflects the BOP’s classification of untimely releases not due to staff error. We
did not review the documentation associated with them or attempt to confirm the BOP’s
classification. The USMS noted that without reviewing the particulars of each of the 49 E-des
cases, it could not verify that they were due to late submission.
Source: Untimely releases the BOP identified as not caused by staff error, FYs 2009 – 2014

The BOP explained that at the department level, the U.S. Marshals Service
(USMS) and the U.S. Parole Commission, respectively, can contribute to causing
untimely releases through either having custody of or granting parole for BOP
inmates. Non-Department entities such as local and state agencies that had

6

custody of an inmate prior to the BOP’s custody can cause or contribute to untimely
releases because the length of an inmate’s sentence may be affected by overturned
or vacated state sentences. To correctly update an inmate’s release date, the BOP
must receive timely and accurate sentencing information.
As shown in Table 2, the BOP attributed as many as 3,969 of the untimely
releases during our review period to court orders and sentencing technicalities at
the federal, state, or local level.8 In addition to the situation described above about
a release that was technically untimely even though no staff error contributed,
DSCC officials presented examples to us of ways that sentencing orders can be
written that leave room for the kind of confusion in computing sentences that can
result in an untimely release. For example, on occasion sentencing orders are
written in a manner that could appear to contravene governing law, such as
sentences that appear to credit time spent in service of another sentence when the
statute governing sentence calculation does not allow it.9 This can create confusion
about the point at which the beginning of the sentence should be calculated,
thereby potentially leading to an untimely release.
One DSCC official told us that, when a sentencing judge’s order is
ambiguous, BOP Classification and Computation staff (CC staff) will try to reach the
court for clarification. However, if the CC staff is unsuccessful in obtaining the
information necessary to fully understand the order so that it can complete the
computation within the DSCC’s established time frame, the staff must interpret the
order according to what it thinks the judge’s intention is and move on to other
cases. Another example a DSCC official described is that sometimes a judge may
include instruction in a sentencing order about how to account for jail credit in a
specific sentence. When this occurs, it can create a difference between what the
inmate’s sentence would be if the calculation was based on the judge’s order and
what the sentence would be if the BOP followed BOP policies, which, according to
the BOP, are based on federal statute and case law.10 We note, however, that
addressing such issues only on a case-by-case basis rather than through systemwide changes precludes the implementation of changes that could prevent this type
of error.
In the OIG’s judgment, the fact that the BOP classified 4,340 releases (nearly
1 percent of all releases) as untimely during the period of our review, but only
8

The 3,969 includes Federal Court Order/Amended Judgment, Federal Ineffectual Sentence,
and Untimely Releases with No Department Involvement.
9

See 18 U.S.C. § 3585(b).

10
According to the BOP’s Legal Resource Guide to the Federal Bureau of Prisons (2014), 12,
the BOP is solely responsible for calculating federal terms of imprisonment. This guide lists the BOP’s
policies for calculating the terms of imprisonment as Program Statements 5880.28, Sentence
Computation Manual (CCCA of 1984); 5880.30, Sentence Computation Manual ("Old Law" Pre-CCCA1984); 5880.33, District of Columbia Sentence Computation Manual; and 5110.16, Administration of
Sentence for Military Inmates. The guide notes that “occasionally, a ‘Judgment and Commitment’ may
direct the defendant’s term of imprisonment to be calculated in a manner contrary to law.” In 1992,
the U.S. Supreme Court held that a district court cannot award jail credit at sentencing (see United
States v. Wilson, 503 U.S. 329 (1992)).

7

comprehensively assessed the 157 that it attributed to BOP staff error,
demonstrates a need for further review of all untimely releases. We believe the
BOP should be a central participant in such a review, but we also believe the review
must include the many entities within and outside the Department that have roles
in writing, interpreting, and executing judicial orders and other sentencing
documents. Coordinated action is needed to identify and implement changes that
would broadly address the significant problem of untimely release.
For such a review to provide a better understanding of the causes of
untimely releases, information about the circumstances surrounding each untimely
release, including, where applicable, the role that non-BOP entities played, must be
available. However, in reviewing untimely release documentation, we found that
the details of the circumstances that led to untimely release are not always
complete enough to determine the cause of each case. This is especially true when
a non-BOP entity may have contributed to the untimely release. The specific details
needed to fully assess the untimely release are known only to the non-BOP entity
that was directly involved, and the BOP does not uniformly have or include this
information in its untimely release documentation. Improving the documentation of
untimely releases, potentially by having all the agencies that contributed to a given
untimely release provide explanatory information, may be a necessary first step
toward a review.
Untimely Releases Due to Staff Error, While Rare in the Context of All
Inmate Releases, Occurred Over 150 Times in a Recent 6-Year Period,
Resulting in Significant Harm
Based on BOP data, we determined that between 2009 and 2014 there were
157 total untimely releases due to staff error out of the 461,966 total inmates
released, representing 0.03 percent of all releases during that 6-year period.11 We
found that of the 157 untimely releases, 152 were late releases and 5 were early
releases. Of the five early releases, three occurred in 2009 and were early by
955 days, 730 days, and 368 days, respectively; one took place in 2014 and was
31 days early; and one took place in 2013 and was 7 days early.
As shown in Table 3, of the 152 late releases during this period, 60 percent
resulted in over-served time of less than 30 days and 40 percent resulted in overserved time of more than 30 days. Three of the inmates over-served by more than
1 additional year: these inmates over-served 928, 541, and 406 days,
11

We found that 148 of these 157 untimely releases were due entirely to BOP staff
error. The remaining 9 fell into the following categories: 2 resulted from a combination of BOP staff
error and over-served time due to a sentence reduction or an ineffectual sentence (we excluded overserved days due to sentence reductions and ineffectual sentences from our analysis); 4 resulted from
a combination of BOP staff error and USMS actions that we could not determine as staff error;
2 resulted from a combination of BOP staff error and either U.S. Probation Office or sentencing court
actions that we could not verify were staff error; and 1 case in which we could not determine whether
the BOP, the USMS, or both were responsible for the untimely release. The information we received
from each agency about these untimely releases was not always detailed or conclusive enough to
allow us to definitively determine the role of each agency or whether staff error was a contributing
factor.

8

respectively. In addition, Table 3 shows that the percentage of releases that were
late was relatively constant between calendar years (CY) 2009 and 2014.
Table 3
Late Releases Caused by Staff Error, by Year and Number of Days Late
CYs 2009 – 2014
DAYS LATE
0-30

2009

2010

2011

2012

2013 2014

TOTAL

16

19

13

16

15

12

91

31-90

8

6

4

5

8

6

37

91-365

5

2

5

3

2

4

21

365+

0

1

0

0

2

0

3

29

28

22

24

27

22

152

Total Untimely Releases

Percentage of All Releases 0.04% 0.04% 0.03% 0.03% 0.03% 0.03%
0.03%
Source: OIG analysis of untimely releases identified by the BOP as caused by staff
error, CYs 2009 – 2014

Untimely Late Releases Result in the Deprivation of Freedom and Contravene the
Sentencing Court’s Order
Untimely late releases are unjust. The length of a prison term is intended to
be commensurate with the crime and other relevant facts of the case, as
determined by the sentencing judge. When inmates are released later than their
correct release date they are deprived of their freedom without just cause during
the days they are over-serving. In the case that was the impetus for this review, a
former BOP inmate filed suit and alleged he was falsely imprisoned because he was
incarcerated beyond his release date.12 The BOP settled with the inmate, and his
case was dismissed. When untimely late releases occur, the harm to the inmate
can be significant and irreparable.
Additionally, untimely releases, whether early or late, result in incarceration
for less or more time than contemplated by the sentencing court. However, we
found that while the BOP program statement contains a notification policy, it is not
comprehensive and the BOP does not have in place a process to consider whether
the sentencing court, as opposed to other offices within the court system, should be
notified of an untimely release.
The BOP program statement setting forth the notification policy and
requirements states:
In any case involving an inmate’s untimely release (including court
ordered releases that are not acted upon in a timely manner) of
30 days or more and/or has been identified as a threat to the
community [sic], the Warden must forward a personal letter to the

12

Hickman v. United States, 14-cv-03041 (D. Minn.), filed July 30, 2014.

9

appropriate Judicial Official(s) providing details of the untimely
release.13
To satisfy this policy for late releases, a BOP official told us that it completes
and forwards a “BP-623: Late Release Notice for United States Probation Service”
form to the appropriate U.S. Probation Office. The BOP sends these notices for all
late releases, including those of less than 30 days. The BOP does not separately
notify the sentencing court, and it does not notify the U.S. Attorney’s Office that
handled the case. One BOP official told us that the BOP does not notify the
sentencing judge for untimely late releases because the term of imprisonment has
been satisfied. With regard to notification to the inmates themselves, BOP Program
Statement 5800.15 requires that a roster of sentence computation changes be
given to unit teams each day, and that a copy of the inmate’s sentence
computation be sent to the inmate’s file and the inmate. The Program Statement
does not indicate if inmates are to be made aware that they have been held beyond
their correct release date for any reason, and it was beyond the scope of this
review to determine if that is done as a matter of practice.
For early releases, BOP officials told us that the BOP notifies the sentencing
court specifically in cases in which an inmate has been released more than 30 days
early, because of the threat the incorrectly released inmate may pose to the public
safety of the community. However, we found that no such notifications occurred
for the four early releases of 30 days or more that we reviewed. For three of these
four cases, a BOP official told us that they did not notify the sentencing court
because the inmates were returned to custody prior to the notification letter being
drafted.14 In the fourth case, a BOP official cited as the basis for not notifying the
sentencing court the fact that the error resulting in the inmate’s early release was
not identified until after the inmate’s correct release date had passed.
Given the sentencing court’s inherent interest in ensuring that its orders are
obeyed, in our judgment the Department should ensure that its components
consider whether, in cases of early or late release due to staff error or any other
reason, notification to the sentencing court would be appropriate. Such an effort
would require that the BOP consult closely with other Department components,
particularly the Executive Office for United States Attorneys, about the
circumstances under which it might be appropriate to notify the courts and/or the
specific U.S. Attorney’s Office that handled a case that led to an untimely release.

13

BOP Program Statement 5800.15, Correctional Systems Manual, January 1, 2009,
Chapter 9, section 906.
14

As discussed later in the report, of the five inmates mistakenly released early, one was out
of custody for 14 days prior to being returned to custody, one for 5 days, and one for less than a full
day. For the other two inmates, the correct release date passed while they were out of custody — one
after 6 days and the other after 30 days.

10

Untimely Releases Result in Abrupt Reentry of Inmates into Society
When the BOP discovers that an inmate is erroneously over-serving time on
an existing sentence, the inmate is immediately released to minimize the number of
days over-served. Unfortunately, this may result in the release of inmates who are
less prepared to reenter society than would have been the case if no error had been
made. The BOP has a variety of programs geared toward preparing inmates for
successful reentry; some are planned around an inmate’s release date. For
example, according to the BOP’s Program Statement, participation in the Release
Preparation Program begins 30 months prior to release and covers programming in
six core areas: (1) Health and Nutrition, (2) Employment, (3) Personal Finance and
Consumer Skills, (4) Information and Community Resources, (5) Release
Requirements and Procedures, and (6) Personal Growth and Development.15 With
some exceptions, all inmates in the BOP’s custody are required to participate in the
Release Preparation Program.16
In addition to an untimely released inmate’s inability to participate in the
Release Preparation Program, an official from the Administrative Office of the U.S.
Courts told us that an abrupt release makes the inmate’s reentry, as well as the
jobs of those who assist with the inmate’s post-release supervision process, much
more difficult. One DSCC official we spoke with suggested that abrupt reentries
into society would be especially difficult for inmates who had been in prison for a
long time.
Untimely Late Releases Result in Significant Monetary Costs to the Taxpayer
There are a number of significant monetary costs that result from untimely
releases due to staff error, including the daily cost of incarcerating inmates beyond
their correct release dates as well as any litigation and settlement costs.17 We
found that between 2009 and 2014, inmates over-served a total of 8,917 days due
to staff error, at an estimated total cost of nearly $669,814 (see Table 4), exclusive
of litigation and settlement costs. In aggregate, inmates over-serving time due to
staff error in medium security facilities cost the BOP the most, at approximately
$254,313; inmates over-serving time in administrative facilities cost the BOP the
least, at approximately $6,364.18
15

BOP Program Statement P5325.07, Release Preparation Program, December 31, 2007,
paragraph 10.b. The BOP’s Release Preparation Program is the subject of an ongoing OIG review.
16

Inmates not required to participate include those committed for psychological study and
observation (to assess their need for psychological treatment or services), those serving a sentence of
6 months or less, those with a death sentence, those confined in an administrative maximum security
institution, those with a “Will Deport Order,” state inmates, deportable aliens, and inmates arriving at
their initially designated institution with less than 6 months remaining on their sentence.
17
It was beyond the scope of our review to analyze the monetary costs resulting from
untimely releases not due to staff error.
18
The BOP’s daily cost to house inmates depends on the security level: the higher the
security level of the inmate, the greater the cost. Administrative facilities are institutions with special
missions that house inmates at all security levels. Inmates are assigned to administrative facilities for

Continued

11

Table 4
Estimated Incarceration Costs of Late Releases Due to
Staff Error, by Security Level
CYs 2009 – 2014

Number of Inmates Total Days Over‐Served
484
Minimum
18
3,474
Low
64
3,487
Medium
45
1,390
21
High
77
Administrative
2
5
Unknown
2
8,917
Total
152

Total Cost to BOP
$
28,365
$
249,648
$
254,313
$
130,810
$
6,364
$
314
$
669,814

Source: OIG analysis of untimely releases identified by BOP as caused by staff
error, CYs 2009 – 2014

In addition to the cost of incarceration, under the Federal Tort Claims Act of
1946, inmates may also seek monetary damages from the federal government for
time over-served.19 An inmate’s first step in seeking compensation for being held
beyond the correct release date is to file an administrative claim with the BOP. If
the BOP denies the claim, or the inmate is dissatisfied with the BOP’s settlement
offer, the inmate may file suit in federal court.
Between 2009 and 2015, the federal government settled four lawsuits filed
by inmates alleging untimely release.20 These settlements were for $90,000;
$120,000; $295,000; and $175,000.21 According to the BOP, these were the only
administrative or litigation settlements for untimely releases that the federal
government paid during this period. This does not include additional costs the
Department incurred as a result of these cases, such as salary costs expended to
handle the lawsuits.

reasons other than security level, such as to be near a court where they are required to be present.
To calculate the cost of an untimely release, we used the average daily cost to house an inmate at the
inmate’s security level during the year the inmate was released. There were two inmates who overserved a total of 5 days for whom the BOP could not identify the security level. We based the average
daily cost on the fiscal year average of the total daily cost per inmate per day by security level, as
provided by the BOP. We determined the number of days by the total number of over-served days for
all late releases, by security level.
19
Title VI of the Legislative Reorganization Act of 1946, codified at 28 U.S.C. § 1346(b) and
28 U.S.C. § 2671-2680.
20
One additional lawsuit, filed between 2009 and 2014, was dismissed. In general, the
relevant U.S. Attorney’s Office handles litigation for untimely release lawsuits.
21

These settlements were paid out of the U.S. Treasury Department’s Judgment Fund.

12

Early Releases Result in Costs to the Department and the Taxpayer, as well
as Potential Harm to Inmates, Their Families, and Their Communities
After an untimely early release occurs, the USMS locates and apprehends the
inmate and then transports the inmate to a BOP facility to undergo intake
processing. These actions cost the BOP and the USMS time and resources.
The potential harm can also be significant for the inmate, his or her family,
and the community. For example, they may suffer harm if the inmate is rearrested
after he or she has gained employment and reestablished ties with family and
community. Also, the community into which the inmate is released early may be
placed in danger if the inmate pursues criminal activity upon release. Though these
potential harms exist, we determined that for the five early releases between 2009
and 2014, none of the inmates was rearrested, charged with, or convicted of any
crimes during the time he or she was in the community before the correct release
date. Additionally, none of the inmates mistakenly released was free for more than
30 days: three were out of custody for no more than 14 days; for the other two,
the correct release date passed while they were out of custody — one after 6 days
and the other after 30 days.
The Causes of Untimely Releases Vary
We found that most untimely releases due to staff error between 2009 and
2014 were the result of mistakes at the DSCC. However, other BOP entities and
possibly the USMS also made mistakes that led to untimely releases. Of the
157 untimely releases between 2009 and 2014 that were within the scope of our
review, the BOP classified 127 as the result of errors made at the DSCC and 30 as
the result of errors at non-DSCC entities.
Untimely Releases as a Result of DSCC Staff Error
Even among untimely releases due to staff error, it was not possible to
classify each untimely release by a single, mutually exclusive cause because some
untimely releases were due to a combination of factors. Rather than comparing
types of errors, we sought to highlight the mistakes that, if addressed, could make
the most difference for reducing the number and severity of untimely releases.
Below we explain the most significant types of errors and how they occurred.
In our review of each of the 127 untimely releases that the BOP classified as
being due to DSCC staff error, we identified certain mistakes as most significant
either because they occurred the most frequently or because they tended to result
in a significant number of days over served. Specifically, misapplication of jail
credit resulted in 59 of the 127 untimely releases — the most of any type of error
we identified in our analysis. Our analysis also found that errors in identifying
primary jurisdiction and consecutive versus concurrent sentences tended to result
in a significant number of days over served. Specifically, the average number of
days over served in untimely releases due to errors in identifying primary
jurisdiction was 169 and those due to errors in determining consecutive versus
concurrent sentences was 102 — the highest of any error type that we identified in

13

our analysis. Below, we elaborate on the errors we believe are most significant
based on our analysis of the 127 errors that occurred at the DSCC.


Misapplication of
Jail Credit.
Example of Misapplication of Jail Credit Resulting in
Correctly computing
Untimely Release
an inmate’s release
One untimely release was the result of the CC staff not
date depends on
following up on documentation that should have led the staff
accurately
member to contact a local jail. A CC staff member reviewed a
document that indicated the inmate may have been arrested
determining the
for a probation violation, but the arrest was not verified in the
amount of jail credit
document. This should have led the staff member to contact
the inmate has that
the jail where the inmate would have been held to ask
should be applied to
whether the arrest had actually taken place and, if so, how
many days the inmate had spent in jail. If the staff member
the sentence. Jail
had contacted the jail, he would have discovered
credit, also referred
approximately 1 month of additional jail credit. After the
to as prior-custody
inmate made an informal request to review his release date,
credit, is typically the
the DSCC discovered and applied the additional jail credit.
amount of time the
Ultimately, the error resulted in a late release of 13 days.
inmate spent in local,
state, or federal custody after the date of offense and before commencement
of the federal sentence. Once applied, this credit reduces the amount of time
inmates must serve in the BOP. When it is not applied — or is applied
incorrectly — it can result in an incorrect release date computation. We
found that incorrect jail credit calculation by DSCC staff caused, in whole or
in part, 59 of the 157 untimely releases within our scope between 2009 and
2014, making it by far the most frequent staff error-related cause of
untimely releases during this period.
Most of the DSCC staff we interviewed said that determining jail credit is the
most difficult part of sentence computation because accurate calculation
requires obtaining and analyzing information from other entities that held the
inmate prior to BOP custody. Local jails, courthouses, state departments of
corrections, and Native American reservations are some of the entities that
DSCC staff may need to contact to obtain this information.
We found three common causes of jail credit error attributable to BOP staff:
(1) not following up to obtain or verify jail credit information; (2) failing to
account for relevant information in the documents; and (3) not following up
on relevant documentation in a timely fashion, the consequences of which
are most pronounced for inmates with shorter sentences. See the text box
above for an example of an untimely release caused by misapplication of jail
credit.



Incorrect Interpretation of Source Documents. As stated above, the CC
staff analyzes documents to determine an inmate’s date of release.
Misinterpretation of one or more of these documents can lead to an incorrect
release date computation. We identified two categories of errors stemming
from a misinterpretation of source documents that were more frequent than
others, which resulted in, on average, longer over-served time than other
14

errors: (1) misidentification of primary jurisdiction and (2) misidentification
of concurrent versus consecutive sentences.
o

Primary Jurisdiction. Several complications can arise when computing the
sentence of an inmate subject to the authority of multiple jurisdictions.
While complicated, especially when an inmate has multiple state and
federal sentences and has been arrested and released by multiple state
and federal authorities, correctly determining which jurisdiction has
“primary jurisdiction” is essential for accurately computing an inmate’s
sentence. Generally the authority, whether federal or state, that first
arrested the inmate has primary jurisdiction over the case. For example,
if an inmate with federal and state charges pending is arrested by federal
agents, the inmate is under primary federal jurisdiction. However, if the
inmate posts federal bond and is released to state authorities, the primary
jurisdiction can change to state jurisdiction. Primary jurisdiction can also
be transferred if an inmate had been released on bail or parole, state
charges are dismissed, or state and federal authorities make an
agreement regarding custody.22
Incorrect determinations of primary jurisdiction often result in inaccurate
application of jail credit or an incorrect sentence start date, which can
lead to an untimely release. We found that 9 of the 157 untimely
releases within the scope of our review were the result, in whole or in
part, from an incorrect determination of primary jurisdiction. These
untimely releases resulted in an average of 169 days of over-served time,
which is much higher than the average of 59 days of over-served time for
all other untimely releases between 2009 and 2014. We describe a
primary jurisdiction error below in Case 2 under Case Studies of Releases
over 1 Year Late.

o

Concurrent versus Consecutive Sentences. If an inmate has been
convicted of more than one crime, the two or more sentences can run
concurrently (at the same time) or consecutively (one sentence begins
after the other ends). We found that nine of the untimely releases within
the scope of our review were, in whole or in part, the result of an
incorrect determination of concurrent versus consecutive sentences.
These errors resulted in an average of 102 days of over-served time, as
compared to the average of 63 days of over-served time for all other
untimely releases between 2009 and 2014. We also found instances
where the DSCC staff mistakenly used the incorrect sentence type even
when the correct sentence type was indicated in the documentation.

22

In the BOP’s comments to this report, the BOP stated that “DSCC staff determinations of
sentence computations may be further complicated when federal sentencing courts expect the federal
government to take a defendant into custody immediately upon sentencing, which is contrary to the
training DSCC staff receive and the computation system utilized at the DSCC.”

15

DSCC Communication with External Entities to Obtain Jail Credit Information
All of the CC staff members we interviewed, as well as many of the other
DSCC staff members with whom we spoke, told us that communication with entities
outside the DSCC to obtain jail credit information can be challenging and that poor
communication with such entities could lead to incorrect release date computations.
As mentioned above, these entities include local jails, courthouses, state
departments of corrections, Native American reservations, the USMS, and others.
We identified several factors that can make communication with external
entities difficult. First, it is not always apparent who has the information the DSCC
staff member needs. CC staff members told us that, while phone numbers for most
entities are available online, whom to talk to and what communication method they
prefer or require is not always apparent. Second, CC staff members told us that
the information external entities provide the DSCC may simply be incorrect or
contrary to information they previously provided, requiring additional clarification.
Third, CC staff members told us that entities vary in their use of terminology. For
example, some entities, particularly local jails, use the term “time served” to
indicate that the inmate was released after serving all of his or her time. In
contrast, the BOP commonly uses this term to indicate all the time the inmate has
served up until that point. DSCC officials told us that such terminology differences
could lead to a misunderstanding and misapplication of jail credit.
Finally, CC staff members told us that difficulty with establishing contact with
some external entities could lead to an incorrect sentence computation and
untimely release. Such a scenario is of particular concern for inmates with short
sentences. For example, a CC staff member who was computing a sentence for an
inmate sentenced to serve 4 months failed to timely follow up with a jail, leading to
an untimely release. The staff member was not able to contact the jail on
January 9 — the day sentence computation was started — and did not attempt to
contact the jail again or complete the sentence computation within the 30-day time
frame for short sentences. On March 3 of the same year, the DSCC discovered that
the sentence computation had not been completed. Once the DSCC successfully
contacted the jail, added the additional jail credit, and completed the computation,
the inmate had already over-served 2 days. A senior DSCC official told us that,
especially because of the short sentence, the DSCC staff member should have
continued to try to contact the jail in the days after January 9.
CC and other DSCC staff members told us that experience computing
sentences at the DSCC plays the most important part in being able to successfully
communicate with external entities regarding jail credit. Over time, CC staff
members develop a rapport with the officials with whom they regularly
communicate. They also obtain direct phone numbers, e-mail addresses, preferred
communication methods, and other helpful information not always available online.
We learned that experienced CC staff members keep this information using their
individual rolodexes, electronic documents, or other methods, and share this
information with their teammates only on an ad hoc basis.

16

However, several DSCC staff members told the OIG that it would be useful to
have a master document that combines all of the knowledge gained from the
team’s experience communicating with external entities in the jurisdictions the
team covers. We also believe that such a document could be useful if DSCC team
members had immediate access to edit its content to keep it up to date. In
addition to serving as a more complete list of contacts than each staff member
would have separately, it would also assist with retaining knowledge that may
otherwise disappear as a result of staff turnover.
Establishing professional relationships is crucial for CC staff members who
must obtain accurate and timely information from staffs working at a wide variety
of federal, state, and local entities. We found that the CC staff does not receive
training focused on establishing and maintaining professional relationships.23 We
believe such training could improve interactions between the DSCC staff and staffs
at external entities and could lead to improved access to jail credit information.
Training may also improve information sharing between DSCC teammates, with a
common goal of ensuring that inmate release dates are calculated correctly.
Though the DSCC does not have training focused on establishing and maintaining
professional relationships, DSCC managers do counsel and train staff members
responsible for errors that led to an untimely release on how the mistakes occurred
and how to avoid them in the future. However, we found that the DSCC does not
track such counseling or training in an employee’s personnel file, which would allow
a later reassessment if the same staff member had made a similar error.
Untimely Releases as a Result of Non-DSCC Staff Error
The BOP attributed 30 late untimely releases (out of the 157 in the scope of
our review) between 2009 and 2014 to entities other than the DSCC. Our review of
the untimely release forms (BP 558 and 623) for these cases indicated that of the
entities that were in whole or in part responsible for these untimely releases, BOP
institutions were responsible for 14; BOP Residential Reentry Management field
offices for 6; Residential Reentry Centers, previously known as Community
Corrections Centers, for 3; private contract prisons for 3; and a BOP regional office
for 1.24

23

Training in interpersonal communication styles could help alleviate computation mistakes.
One CC staff member described the importance of building and maintaining a good rapport with
representatives from external entities by telling us that a point of contact with whom he had
developed an effective working relationship refused to take calls directly from the DSCC staff after
having a negative interaction with another DSCC employee. This example shows how poor
interpersonal communication among CC staff members can impede timely and accurate sentence
computations. A DSCC manager noted that the teams often work with the same jurisdictions
regularly; therefore, building rapport can assist teams in obtaining the information necessary to make
accurate calculations.
24

In one of the six untimely releases we attributed to errors of Residential Reentry
Management field offices, the error occurred at the local jail to which the field office had assigned the
BOP inmate. For 3 of the 30 untimely releases, the BOP attributed error to non-DSCC staffs; but we
could not determine whether the USMS, a BOP institution, or both were responsible. BOP
documentation indicated that the USMS was at least partially responsible, but the USMS told us that

Continued

17

The forms for each untimely release, known as Untimely Release Notification
(BP 558) and Late Release Notice for United States Probation Service (BP 623),
described the reason for the untimely release. Based on our assessment of each
narrative, we determined that the four most common descriptions of untimely
releases due to non-DSCC staff error were:
1. Notification or Information Failures. This includes untimely notifications
as well as failures to update, send, or use the correct information. For
example, in one situation an inmate incorrectly lost good time credit after a
staff member credited it to the wrong inmate. By the time the error was
discovered and the good time reapplied, the inmate had over-served 3 days.
2. Releases the Day After the Correct Release Date, Sometimes without
Explanation. In some of these cases, the BOP was unable to provide us
additional information regarding the staff errors that caused the late release
or to determine with certainty whether the untimely release was in fact the
result of staff error. For example, in one of these cases, after the DSCC
updated a sentence computation and notified the facility that the inmate
should be released that day, the institution did not release the inmate until
the next day and the untimely release report offered no explanation for the
delay. According to the BOP, generally when an institution cannot release an
inmate on the day he or she is supposed to be released, it can be due to
transportation issues, medication requirements, lack of release plans, or
incomplete victim and witness notification or Adam Walsh Child Protection
and Safety Act of 2006 review requirements.
3. Sentence Computation Errors, Primarily by BOP Institutions That
Computed Inmate Sentences Prior to the DSCC’s Existence, and One
Computed by a Contract Facility That Had Yet to Transition Its
Sentence Computation Functions to the DSCC. For example, a private
contract facility staff member incorrectly computed an inmate’s sentence by
misidentifying and misapplying jail credit, resulting in 281 days over-served,
the most significant late release due to non-DSCC error.
4. Failures to Properly Monitor Inmate Release Dates. In one situation,
institution staff failed to run an inmate release roster containing the
scheduled release dates of individual inmates, resulting in one inmate overserving 1 day.
A BOP official told us that the causes of untimely releases are not always
identifiable because official documentation sometimes provides inadequate
descriptions. We believe that not having thorough and accurate descriptions of the

its records did not indicate that the USMS was responsible for the over-served time. This lack of
clarity underscores the need for an assessment of untimely releases at the Department level that
would allow for a better understanding of how multiple agencies can contribute to untimely releases so
that effective solutions for preventing them can be developed.

18

events leading to untimely release hinders the BOP in identifying and addressing
their underlying causes.
The Transition of Sentence Computation Responsibility to the DSCC Appears to
Have Reduced the Number of Untimely Releases
The BOP transitioned sentence computations to the DSCC in 2005 in order to
consolidate all BOP sentence computation functions, reduce costs, and ensure
consistent application of laws and BOP policies. According to the BOP’s information,
the transfer also appears to have reduced the number of untimely releases due to
staff error. As shown in Table 5, between 1999 and 2004 — the 6 years before the
DSCC was created in 2005 — there were approximately 344 untimely releases due
to staff error (approximately 0.1 percent of the total number of inmates released
during the period), compared to 157 during the 6 years between 2009 and 2014
that are the subject of this review (approximately 0.03 percent of the total number
of inmates released during this period).25
Table 5
Untimely Releases Due to Staff Error before and after
Creation of the DSCC
FYs 1999 – 2004 and CYs 2009 – 2014

Year
1999
2000
2001
Pre‐DSCC
2002
2003
2004
2009
2010
2011
Post‐DSCC
2012
2013
2014

1st
18
22
22
16
17
19
8
10
9
3
5
8

Quarter
2nd 3rd
9 11
17 6
19 22
14 12
10 16
16 14
11 11
5 10
3 4
6 8
4 10
3 5

4th
15
22
16
11
2
3
6
7
9
7

Total
53
45
85
58
54
49
32
28
22
24
28
23

Grand Total % of All Releases

344

0.10%

157

0.03%

Sources: BOP-provided information for FYs 1999 – 2004 and OIG analysis of
untimely releases due to staff error as identified by the BOP, CYs 2009 – 2014

25

The BOP told us that it was unable to provide untimely release information for the fourth
quarters of 2000 and 2004 because it had retained some untimely release records for analysis and
destroyed others as part of the Records Information Disposal Schedule (they were older than 6 years).

19

BOP Processes to Prevent Untimely Releases
The DSCC staff members we interviewed, including the Chief of the DSCC,
told us that because people perform sentence computations, mistakes are
inevitable. There is the possibility of human error at every step of the sentence
computation process, from initial computation to final release audit. We identified
several crucial measures the BOP has in place that help prevent untimely release,
including certification and final release audits that double check the initial sentence
computation, program reviews of DSCC operations that can result in operational
improvements, and processes through which inmates can request a review of their
release date. We found that while these measures help to reduce untimely releases
due to staff error, they did not eliminate them. As described below, we also
identified one additional measure the DSCC could take to help prevent untimely
releases due to staff error.
Certification and Final Release Audits
Between 2009 and 2014, 29 of the 157 staff errors that resulted in untimely
releases were discovered by certification audits and 19 were discovered by final
release audits. We found that final release audits are likely to prevent an untimely
release, but only if they occur early enough before an inmate’s release date that
any reduced sentence time can be accounted for. Currently, the DSCC performs a
final release audit 12 months before an inmate’s projected release date. Yet for 14
of the 19 errors discovered by final release audits, the untimely release may not
have resulted if the final release audit had occurred 18 months before the inmate’s
release date instead of 12 months.26 Similarly, 18 of the 19 errors may not have
resulted in untimely release if the final release audit had occurred 24 months prior
to the release date.27
Program Reviews
The BOP’s Program Review Division (PRD) and the DSCC routinely conduct
program reviews of the DSCC to evaluate trends and identify deficiencies in its
operations. Program reviews have resulted in changes to DSCC operations that can
help prevent untimely releases. Two PRD reviews, conducted in August of 2009
and September of 2011, rated the DSCC as “acceptable.” However, a September
2013 review rated the DSCC as “deficient.” The PRD’s report stated that the
“deficient” rating reflected in part that “initial sentence computations are not always
computed and audited within established time frames” and not all source
documents “support the sentence computation.” To address the former finding,
DSCC documentation indicated that the DSCC filled at least four vacant positions
26

We based this determination on the number of over-served days for each inmate. Each of
the 14 inmates had less than 6 months of over-served time when the error was discovered during the
final release audit, meaning that conducting the final release audits 6 months earlier may have
allowed the BOP to avoid the untimely releases.
27
These four inmates had between 6 and 12 months of over-served time at the time the final
release audit discovered the error.

20

following a period of low staffing levels that the BOP attributed to sequestration and
a hiring freeze, and DSCC Section Chiefs began having bi-weekly meetings with
managers to ensure the timely completion of certification audits. The DSCC also
identified specific teams that had difficulty meeting time frame requirements and
placed special emphasis on assisting those teams in complying with the established
timeframes. To address the latter finding, according to documentation the BOP
provided, the DSCC established a tracking system to identify sentence computation
errors and trends, including the tracking of missing source documents; devised a
method to assist in naming and saving of source documents; and used bi-weekly
meetings to discuss the findings of the tracking system’s error report. In
September 2014, the PRD closed the 2013 review, stating in a memorandum that
the DSCC had resolved all of the deficiencies to the PRD’s satisfaction and had put
in place administrative controls to prevent recurrence. We did not independently
assess the actions the DSCC took that resulted in the PRD closing the review.
Perpetual Audits
“Perpetual audits” are another process the DSCC uses to improve operations
and sentence computation accuracy, whereby every month the DSCC staff checks
170 sentence computations (10 for each DSCC team). A DSCC official said that, as
a matter of practice, the DSCC uses perpetual audits to identify common mistakes
and then conducts training to help DSCC staff avoid making them. In this review,
we assessed the monthly perpetual audits that the DSCC conducted between
October 2013 and November 2014. Our assessment identified an overall sentence
computation error rate of greater than 5 percent for 12 out of 13 months examined.
A DSCC official confirmed that the DSCC considers error rates greater than
5 percent to be high. One month’s overall error rate was 12 percent. The DSCC
told us that to address the high error rates it expanded training for DSCC
employees throughout the year by including in the training more complex
computation situations and additional information about verifying state information
and by conducting training more frequently. The BOP told us that these actions
resulted in an error rate of 2.4 percent and an overall rating of “good” on the PRD’s
January 2015 program review.
Educational Events
We also found that the DSCC conducts educational events related to DSCC
operations, such as sentence computation, that may have unmet potential for
preventing untimely release. Some of these events bring together officials from the
DSCC and non-BOP officials including U.S. District Court Judges, federal Public
Defenders, and U.S. Probation Office staff. At these events, the non-DSCC
attendees request the topics that are discussed, such as sentence computation, the
interaction of federal and non-federal sentences, judicial recommendations, and a
general understanding of the role of the DSCC. The DSCC conducted 13 of these
educational events between 2009 and 2014.
Historically, the DSCC has not used these events explicitly for the purpose of
educating the participants on the means to prevent untimely release, even though
they may sometimes offer information that could be helpful in preventing them.
21

During this review, a DSCC manager expressed the opinion that these education
events have the potential to help reduce the number of untimely releases. For
example, showing criminal justice officials how conflicting information can lead to
an incorrect sentence computation may prompt them to adjust the way they record
information. The official also told us that the DSCC has no formal participant
feedback process for these events.
We agree that these events could provide a useful forum to discuss topics
and share information that could help avoid untimely releases, including topics and
information identified by DSCC staff (as opposed to non-DSCC attendees) as
pertinent to the effort to prevent untimely releases. The DSCC could further
capitalize on these events if it were to identify and invite participation from nonBOP officials whose actions contribute to ensuring timely inmate releases, but who
have not previously taken part in these events. We also note that as valuable as
these events may already be, the DSCC is missing an opportunity to further
enhance their impact and effectiveness by not collecting feedback from participants,
which it could use to tailor content to participant needs and interests.
Inquiries by Inmates, Their Families, and Advocates
Inmates, their families, and other advocates have used formal and informal
inquiry processes to help bring to light errors made during the sentence
computation process. Inmates can, in writing or orally, informally request that the
BOP review their release date. Between 2009 and 2014, 54 of the 157 errors that
resulted in untimely releases were discovered through inmates’ informal requests
for review.
Inmates may also at any time file a formal administrative remedy request
relating to any aspect of their confinement, including their release date.
Administrative remedy requests that challenge a release date have three increasing
levels of review: (1) the inmate’s institution, (2) the regional office, and (3) the
Central Office. Between 2009 and 2014, 6 of the 157 errors that resulted in
untimely releases were discovered through the administrative remedy process,
3 were discovered as a result of an attorney’s inquiry, and 1 was discovered as a
result of a family’s inquiry.
Case Studies of Releases Over 1 Year Late
Below we describe the three untimely releases that resulted in inmates
spending particularly significant amounts of over-served time in prison. We
describe why these late releases occurred, what was done to prevent the mistake
from happening again, and the cost to the BOP. In each case, the BOP released the
inmate as soon as the error was discovered and the sentence correctly computed.
Case 1: Misapplication of Jail Credit Resulting in 928 Days of Over-served Time
When the DSCC performed the initial sentence computation for this inmate, it
did not obtain the relevant state sentencing documents that would have led it to
apply jail credit to the inmate’s federal sentence. The DSCC did not discover the

22

sentence computation error during the certification audit. Instead, the inmate’s
inquiry about his release date brought the error to the attention of a BOP regional
office. A total of 3 days passed from the date the inmate made the inquiry until the
inmate was released. DSCC officials we spoke to said they did not know why the
staff did not obtain state sentencing information during the initial sentence
computation. The inmate was improperly confined in prison for 928 days, or nearly
3 years, past the correct release date. The result was a serious deprivation of the
inmate’s liberty and a violation of the court’s sentencing order. There also was an
additional cost to taxpayers of $66,732 for incarcerating the inmate in a medium
security facility for this additional period. The BOP offered no compensation to the
inmate, and the inmate did not seek monetary relief. DSCC officials told us that
they took no additional action as a result of this error because the staff members
responsible for it were no longer employed by the BOP when the error was
discovered.
Case 2: Incorrect Primary Jurisdiction Determination Resulting in 541 Days of
Over-served Time
When the DSCC performed the initial sentence computation for this inmate, it
was unaware of a key document, specifically an amendment to a state judge’s
previous order that resulted in the release of the inmate from state prison. It was
only when the inmate challenged the sentence computation that the DSCC staff
discovered the error by finding a reference to the amendment in the online Public
Access to Court Electronic Records (PACER) system and, as a result, released the
inmate.28 DSCC officials told us that the CC staff was not aware of the missing
document because staff members do not search PACER when they perform
sentence computations. DSCC managers told us that the DSCC staff uses PACER
only for verifying court orders and reviewing dockets when responding to inquiries
and not for sentence computations because doing so would be too cost and labor
intensive.29 As a result of not being aware of the amendment, the DSCC staff
incorrectly computed the inmate’s sentence based on primary state jurisdiction
instead of primary federal jurisdiction. This resulted in jail credit not being applied
to the inmate’s sentence and the inmate being improperly confined for 541 days, or
nearly 1.5 years, past the correct release date, a serious deprivation of the inmate’s
liberty and a violation of the court’s sentencing order. The error also resulted in an
additional cost to taxpayers of approximately $40,428 for incarcerating the inmate
in a medium security facility for this additional period. The BOP offered no
compensation to the inmate, and the inmate did not seek monetary relief. DSCC
officials told us that as a result of this untimely release they provided training on
the late discovery of information and primary jurisdiction to the team that worked
on this inmate’s sentence computation.
28

PACER is an electronic public access service that allows users to obtain case and docket
information online from federal courts.
29

The BOP stated that the DSCC staff annually conducts approximately 153,309 initial
computations and 91,475 updates to existing computations. Because PACER currently charges
10 cents for every page viewed, viewing one page for each inmate would cost approximately $24,478.
This underestimates the total cost because dockets are often longer than one page.

23

Case 3: Misinterpretation of a Judge’s Sentencing Order Resulting in 406 Days of
Over-served Time
This inmate was initially convicted on two counts that were to run
consecutively. A judge subsequently reduced both counts to a combined
76 months. However, in updating the sentence, the DSCC staff misinterpreted the
judge’s sentencing order to mean that only the longer of the two counts was to be
reduced to 76 months. Both the DSCC staff member performing the computation
update and the staff member performing the certification audit to the update made
this mistake. As a result, the DSCC scheduled the inmate’s release for a
significantly later date than it should have. The DSCC discovered the error during
the final release audit, approximately 12 months prior to the inmate’s incorrect
release date. Ultimately, the inmate was improperly confined for 406 days, or over
1 year, past the correct release date. This resulted in a serious deprivation of the
inmate’s liberty and a violation of the court’s sentencing order. It also cost
taxpayers approximately $31,619 to incarcerate the inmate in a low security facility
for this additional period. The inmate filed a lawsuit against the BOP that ended in
a settlement of $175,000. The DSCC Chief’s note on the BP 558 form indicated
that staff members responsible for the error were reminded to clarify with their
supervisor any sentencing order that may be ambiguous.

24

CONCLUSION AND RECOMMENDATIONS
Conclusion
We found that of the 461,966 inmate releases between 2009 and 2014, the
BOP categorized 157 as “untimely” due to staff error and a far greater number —
4,183 — as untimely for other reasons. According to the BOP, the vast majority of
untimely releases not due to staff error were due to situations beyond its control,
such as amended sentences that resulted in shorter sentences than the time an
inmate had already served, compared to the 157 it classified as due to staff error.
Data and information we reviewed indicates that other entities may sometimes
contribute to untimely releases. Additionally, we found that the BOP does not
always have complete information about the circumstances of untimely releases
that other entities contributed to. We believe that a review of untimely releases
with participation from all relevant entities, both within and outside of the
Department, is needed to assess the full range of causes of untimely releases and
how best to address those that are in any way preventable.
Our review focused on analyzing the 157 untimely releases that the BOP
attributed to staff error. We found that untimely releases due to staff error are rare
as a percentage of all inmate releases but, like all untimely releases, can cause
very significant harms when they do occur. Late untimely releases inappropriately
deprive inmates of their freedom, while early releases result in several potential
harms to inmates, their families, and their communities. Both early and late
untimely releases contravene court sentencing orders. Moreover, 6 of the
157 releases we examined were untimely by more than a year: we identified early
releases of 955, 730, and 368 days and late releases of 928, 541, and 406 days.
While the causes of untimely releases due to staff error are varied, we found
that most are the result of an incorrectly computed release date. These
computation errors often occur due to the Designation and Sentence Computation
Center (DSCC) staff either not getting complete and accurate information from nonBOP entities or not interpreting the information they receive correctly. Although
the BOP has several processes in place to prevent untimely releases, we found that
there were several steps the BOP could take to further reduce the possibility of
untimely release. These include improving, through training and otherwise, the
communication tools and skills of the DSCC staff members who must obtain and
interpret information from wide variety of federal, state, and local entities outside
the DSCC, and implementing additional sentence calculation processing or auditing
strategies taking into account that conducting final release audits only 12 months
before an inmate’s projected release date has led to preventable untimely releases.
Recommendations
To enhance understanding of the circumstances and causes of untimely
releases so that the utility of cross-agency solutions for preventing them can be
considered, we recommend that the Office of the Deputy Attorney General (ODAG):

25

1.

Work with all relevant entities, both within and outside of the Department, to
review the full range of possible reasons for all untimely releases, including
those that may not be due to staff error but may still be preventable, and to
identify opportunities to reduce the likelihood that they occur.

We make the following recommendations for the BOP to ensure that
appropriate notifications are made to the court in cases of untimely release and
improve the DSCC’s ability to prevent and appropriately handle untimely releases:
2.

Work with the ODAG to establish a policy to identify under what
circumstances it is appropriate to notify the relevant U.S. Attorney’s Office,
court official(s), and the affected inmate of an untimely release for any
reason, and the reason therefor, and develop a process to ensure that timely
notification takes place as appropriate.

3.

Ensure that BOP staff members thoroughly and consistently describe and
identify the cause of each untimely release in official documentation (such as
BP 558 and 623 forms) and use this documentation to support the DSCC’s
trend analysis for improving training and operations.

4.

Implement methods and training to address communication challenges in
obtaining accurate and timely jail credit information from judicial as well as
other necessary sources.

5.

To address the concerns identified in this report that conducting final release
audits 12 months before an inmate’s projected release date has led to
untimely releases due to BOP staff error, explore and implement sentence
calculation processing or auditing strategies designed to decrease the
likelihood of sentence calculation errors by BOP staff.

6.

Incorporate specific untimely release issues and possible solutions into
education events, and establish a process for obtaining formal feedback from
participants to ensure methods for preventing untimely releases are fully
explored.

7.

Identify and invite participation from additional non-BOP officials who do not
typically participate in education events, but whose actions could help to
ensure timely releases, including U.S. District Court Judges, federal Public
Defenders, and U.S. Probation Office staff.

26

APPENDIX 1
METHODOLOGY OF THE OIG REVIEW
The OIG reviewed untimely releases between calendar years (CY) 2009 and
2014. Our review entailed data analysis of BOP untimely release information and
interviews with BOP officials knowledgeable about the BOP’s process for computing
sentences. Of the 4,340 untimely releases that occurred during the scope of our
review, we focused primarily on the 157 untimely releases that the BOP identified
as having been the result of staff error and determined which entities contributed to
the errors. We also reviewed information pertaining to the 4,183 untimely releases
not due to staff error, such as the categories of causes for them, but we did not
review individual cases.
We obtained untimely release documentation, in the form of Untimely
Release Notification (BP 558) and Late Release Notice for United States Probation
Service (BP 623) for untimely releases due to staff error from CY 2009 through CY
2014, for Designation and Sentence Computation Center (DSCC) and non-DSCC
untimely releases. We compiled this data into a database and analyzed the
descriptive statistics. The sample dataset included all untimely release cases the
BOP classified as due to staff error during the period of our review; we excluded
none from the statistical analysis. Our analysis incorporated a count of the number
of cases, the number of total days inmates were released either before or after the
correct release date, the average (mean) days under-served or over-served, and its
associated standard deviation, among other common descriptive statistical
methods. As part of this analysis, we used the narrative portion of the untimely
release documentation forms to ascertain information about each untimely release,
including the security level of the inmate, how the BOP discovered the error, and
the primary factor that caused the error. We also obtained from the BOP the
average per-inmate daily cost by security level and year and used it to estimate the
costs of the untimely releases due to staff error within our sample. We also
reviewed information obtained from e-mail correspondence and phone
conversations with non-BOP entities such as the U.S. Marshals Service, the
Executive Office for United States Attorneys, and the Administrative Office of the
U.S. Courts.
We conducted 28 interviews with BOP staff to learn about the BOP’s sentence
computation process and perform a qualitative assessment of untimely releases.
During these interviews, we inquired about a variety of topics, including explaining
or providing context to the data analysis described above. These interviews
included non-DSCC staff as well as DSCC staff at various levels within the
organization and included Classification and Computation Technicians, Classification
and Computation Specialists, management, and the DSCC Chief.

27

APPENDIX 2
BOP REVIEWS OF THE DESIGNATION AND SENTENCE
COMPUTATION CENTER
Beyond audits of individual sentence computations, the BOP, as well as the
Designation and Sentence Computation Center (DSCC), conducts program reviews.
The BOP’s Program Review Division (PRD), tasked with providing oversight of BOP
program performance and compliance, reviews the DSCC to evaluate several areas,
sentence computations being one of these areas. The PRD conducts program
reviews on regular cycles: every 3 years when the DSCC has received an
operational rating of “good or superior,” 2 years for a rating of “acceptable,” and
1 to 1.5 years for a rating of “deficient.” When the PRD reviews the DSCC, it issues
a report with any specific areas of deficiency and an overall operational rating. The
DSCC must respond to the report within 30 days of issuance. Then, within 120 to
150 days of the issuance of the report, the DSCC must conduct its own follow-up
review and send a report to the PRD. The follow-up review shows the preliminary
results of any actions taken in response to the program review’s findings. If no
deficiencies are found in the PRD’s program review, then no response or follow-up
review is required.
During years without program reviews, the DSCC is required to conduct
internal reviews of its programs, called operational reviews, in a process similar to
that of the PRD’s program reviews. Operational reviews are also reported to the
PRD and have a similar response and follow-up review process based upon the
findings of the operational review.

28

APPENDIX 3
THE ODAG AND BOP’S RESPONSE TO THE DRAFT REPORT
U.S. Department of Jnstice
Office of the Deputy Attorney General

Washington, D.C. 205}0

Associate Deputy Attorney General

May 18.2016

MEMORANDUM FOR NINA S. PELLETIER
ASSISTANT INSPECTOR GENERAL
EVALVA TION AND INSPECTIONS DIVISION
FROM:

Carlos F. Uriarte
CrV..
Associate Deputy Attorney General
Office of the Deputy Attorney General

/£/tf

Raphael A. Prober
Associate Deputy Attorney General
Office of the Deputy Attorney General
Rae Woods {,Fl,,( ~,,. Rw
Senior Counsel to the Deputy Attorney General
Office of the Deputy Attorney General
Thomas R. Kane, Acting Director
Federal Bureau of Prisons
SUBJECT:

/2i)f p, J (< }(_

Response to Draft Report of the Office of the Inspector General: Review a/the
Federal Bureau o,l Prisons' Untimely Release of Inmates

We appreciate the review undertaken by the Department of Justice (Department) Office of the
Inspector General (OIG) entitled Review of the Federal Bureau al Prisons' Untimely /le/ease of
Inmates. The OIG report contains seven recommendations; the Department concurs with all
seven, subject to the comments provided below outlining the Department's continuing concerns
(which are largely definitional in nature). As certain of the recommendations arc directed to or
otherwise involve the Office of the Deputy Attorney General (ODAG), and as the balance arc
directed to Federal Bureau of Prisons (BOP), this memorandum is being submitted jointly by
ODAG and BOP. With regard to the O!G · s forn1al drati report provided on April 20, 2016,
while neither ODAG nor BOP contest the underlying data or conclusions contained in the report,
the report's presentation and imprecise use of the term ·'untimely release" continues to be
significantly misleading and blurs the critical distinction between miscalculated prison sentences
due to BOP staff error ( 157). versus "untimely releases" for other. non-erroneous reasons
(4J 83).

29

SUBJECT: Response to Draft Report oflhe Office of the Inspector General:
Review of the Federal Bureau of Prisons' Untimely Release of Inmates

Page 2

13y failing to clearly and consistently define terms and distinguish throughout the report between
miscalculated sentences due to BOP staff error. and ''untimely releases·' for non-error reasons.
the report fosters a misleading perception that the 4,183 cases reviewed by BOP as "untimely
releases" are equivalent in nature with the 157 cases which actually resulted from BOP staff
error, which is absolutely not the case. In fact, many of those 4,183 cases reviewed by BOP as
"'untimely releases" are entirely appropriate releases, as described in further detail below. Stated
affirmatively, OIG's review revealed 80P's accuracy rate for sentence calculations during the
timeframe reviewed to be 99. 97%.
Recommendation 1: To the Office o.fthe Deputy Allorney General, to work with all relevant
entities, both within and oulside ofJhe Department, to review thefi,ll range ojjJossible reasons
for all untimely releases, including those that may not he due to stqff error but may still be
preventable, and to identify opportunities Jo reduce the likelihood that they occur.
Response: ODAG concurs with Recommendation 1, and agrees to look at the full range of
untimely releases, but will focus its efforts on trying to understand those untimely releases that
are both preventable and improper. As discussed in the report, the vast majority (4,183) of
"'untimely releases" are the result of court initiated actions (i.e., original, amended, and/or
vacated court orders), and are not errors of any type. It is, therefore, presumed that the courts'
orders, and their consequences, were intentional. Insofar as calculation, and recalculation, of
inmates' sentences consistent with valid court orders is in the ordinary course of business, there
is no demonstrated need for ODAG to work to ''reduce the likelihood'" of preventing such
occurrences.
Recommendation 2: To BOP, to work with ODAG to establish a policy to ident/fy under what
circumstances it is appropriate Jo notify the relevant U.S. Attorney's Ofjice and/or court
officia/(s), and the affected inmate t?{ an untimely release for any reason, and the reason
therefore. of an untimely release for a,~y reason including but not limited to those caused by staff
error.
Response: BOP agrees with this recommendation, and will work with ODAG to establish, or
revise, its policy to identify under what circumstances it is appropriate to notify the relevant U.S.
Attorney's offices anrl/or cou11 officials(s)-, nnrl the i:ltl'ected inmate, anrl develop a process to
ensure that timely notification occurs.
Recommendation 3: To BOP, to ensure that BOP staff members thoroughf:v and consistently
describe and idenl(fy the cause of each untimely release in l(/jicial documentalion (rnch as
BP 558 and 623 .forms) and use this documentation to support the DSCC 's trend anulysisfor
improving training and operations.
Response: BOP agrees with this recommendation and will explore methods frlr consistently
describing and identifying the cause of untimely releases due to staff error in official
documentation, for use in support of DSCC trend analysis for improving training and operations.
Recommendation 4: To BOP, 10 implement methods and Lraining lo address communication
challenges in obtaining accurate and timelyjail credit informationfi-omjudicial as ·well as other

30

SUBJECT: Response to Draft Report of the Ot1ice of the Inspector General:
Review of the Federal Bureau of Prisons' Untimely Release of Inmates

Page 3

necessary sources.
Response: BOP agrees with this recommendation and will explore enhanced methods and
training to obtain accurate and timely jail credit infonnation from judicial, as well as other
necessary sources.
Recommendation 5: To BOP, to address the concerns identfjied in this report that conducting
final rdease audits 12 months before an inmate's projected release date has led to untimely
releases due to BOP staff error, explore and implement sentence calculation processinx or
auditing strategies designed to decrease the likelihood of sentence calculation errors by BOP
staff
Response: BOP agrees with this recommendation, and will explore and implement sentence
calculation processing or auditing strategies designed to decrease the likelihood of sentence
calculation errors by BOP staff.
Recommendation 6: To BOP, incorporate specific untimely release issues and possible
solutions into education events, and establish a process for obtaining formal feedback from
participants to ensure methods for preventing untimely release are fully explored.

Response: BOP agrees with this recommendation and will incorporate untimely release issues
related to staff error into education events and request feedback from participants to ensure
methods of preventing such.

Recommendation 7: To BOP, identify and invite participation from additional non-BOP
officials who do not typically participate in education events, but whose actions could help to
ensure timely releases, including US. District Court Judges, federal Public Defenders, and US.
Probation Office staff
Response: BOP agrees with this recommendation and will explore identifying and inviting nonBOP representatives from the referenced offices to participate in BOP educational events.
Additionally, BOP"s response to this recommendation will include BOP/DSCC staff
pm1icipation in educational events coordim1ted by those non-BOP offices"

31

APPENDIX 4

OIG ANALYSIS OF THE ODAG AND BOP’s RESPONSE
The OIG provided a draft of this report to the Office of the Deputy Attorney
General (ODAG) and the BOP for their comments. The ODAG and BOP’s response is
included in Appendix 3 of this report. Below, we discuss the OIG’s analysis of the
ODAG and BOP’s response and the actions necessary to close the
recommendations.
Recommendation 1: Work with all relevant entities, both within and
outside of the Department, to review the full range of possible reasons for all
untimely releases, including those that may not be due to staff error but may still
be preventable, and to identify opportunities to reduce the likelihood that they
occur.
Status: Resolved.
ODAG and BOP Response: The ODAG concurred with the recommendation
and stated that it agrees to look at the full range of untimely releases while
focusing on trying to understand those untimely releases that are both preventable
and improper. The ODAG referred to information in the report showing that the
vast majority of releases that the BOP classifies as “untimely” are the result of court
initiated actions rather than error, noting that there is no demonstrated need for
the ODAG to reduce the likelihood of preventing non-erroneous occurrences.
OIG Analysis: The ODAG’s planned actions are responsive to the
recommendation, as long as its review includes a complete assessment of what
constitutes a “preventable and improper” untimely release.
In their comments, the ODAG and BOP referred to continuing definitional
concerns with our report pertaining to distinguishing between untimely releases due
to miscalculated prison sentences resulting from BOP staff error on the one hand,
and those due to other non-erroneous reasons on the other hand. In this report,
we use the term “untimely release” the way the BOP currently uses it, as a broad
description of all instances in which an inmate was not released on his or her
correct date, including when this occurs for non-erroneous reasons. The term
therefore includes instances we refer to as technicalities, such as when the “correct
date” is established retroactively because of a court’s sentencing adjustments.
Because the BOP’s definition of “untimely release” blurs the distinction
between proper and improper, and between preventable and unpreventable, we
believe that examining, and possibly revising, the definition of untimely release
could be a useful aspect of the ODAG’s assessment. But, at a minimum, the
ODAG’s assessment should avoid artificially narrowing the universe of untimely
releases at issue, since the intent of our recommendation is for the ODAG, BOP,
and all other stakeholders to explore, collaboratively, all the ways untimely releases
occur so that they can identify and implement methods to prevent them and avoid
as many untimely releases as possible.
32

Additionally, we noted in our report that it was beyond the scope of our
review to assess the circumstances of untimely releases not due to staff error, of
which the BOP’s records indicate there were more than 4,000. To the extent that
the ODAG and BOP’s joint response is intended to imply that any particular portion,
or all, of those more than 4,000 untimely releases were not the result of an error,
or not preventable, we emphasize our finding that “the BOP does not always have
complete information about the circumstances of untimely releases to which other
entities contribute.” Therefore, we do not believe that the ODAG should define the
scope of the multiagency review at the outset in a manner that would categorically
exclude cases about which the BOP lacks complete information. While we agree
that there is no need — and no way — to prevent releases that are “untimely” only
because of a retroactive sentence adjustment, we believe a full review is needed
because of the wide range of factors that may contribute to untimely releases,
some of which may currently be unknown to the ODAG, the BOP, or other
stakeholders, and may be preventable.
By August 31, 2016, please provide the ODAG’s plan for looking into the full
range of untimely releases, including its framework for defining and identifying
which untimely releases are preventable and improper, as well as a list of the
agencies it plans to include in its review.
Recommendation 2: Work with the ODAG to establish a policy to identify
under what circumstances it is appropriate to notify the relevant U.S. Attorney’s
Office, court official(s), and the affected inmate of an untimely release for any
reason, and the reason therefor, and develop a process to ensure that timely
notification takes place as appropriate.
Status: Resolved.
ODAG and BOP Response: The BOP stated that it agrees with this
recommendation and that it will work with the ODAG to establish, or revise, its
policy to identify the circumstances under which it is appropriate to notify the
relevant U.S. Attorney’s Office and/or court officials(s), and the affected inmate,
and will develop a process to ensure that timely notification occurs.
OIG Analysis: The BOP’s planned actions are responsive to the
recommendation. By August 31, 2016, please provide the policy and/or policy
changes that identify the circumstances of untimely release under which it is
appropriate to notify the relevant U.S. Attorney’s Office, court officials(s), and the
affected inmate, including the reason for the untimely release, as well as
documentation of the process that will ensure that timely notification occurs.
Recommendation 3: Ensure that BOP staff members thoroughly and
consistently describe and identify the cause of each untimely release in official
documentation (such as BP 558 and 623 forms) and use this documentation to
support the Designation and Sentence Computation Center’s (DSCC) trend analysis
for improving training and operations.

33

Status: Resolved.
ODAG and BOP Response: The BOP stated that it agrees with this
recommendation and will explore methods for consistently describing and
identifying the cause of untimely releases due to staff error in official
documentation.
OIG Analysis: The BOP’s response is partially responsive to the
recommendation. To be fully responsive to this recommendation, the BOP must
also explore methods for consistently describing and identifying in official
documentation the causes of untimely releases not due to staff error. We note that
it will not always be possible to completely document untimely releases beyond the
BOP’s control. However, it is still important to enhance such documentation where
possible because this information can also be useful in trend analysis for improving
training and operations. By August 31, 2016, please provide documentation of the
methods developed for consistently describing and identifying the cause of all
untimely releases.
Recommendation 4: Implement methods and training to address
communication challenges in obtaining accurate and timely jail credit information
from judicial as well as other necessary sources.
Status: Resolved.
ODAG and BOP Response: The BOP stated that it agrees with this
recommendation and will explore enhanced methods and training to obtain accurate
and timely jail credit information from judicial as well as other necessary sources.
OIG Analysis: The BOP’s planned actions are responsive to the
recommendation. By August 31, 2016, please provide documentation of the
enhanced methods and training explored — indicating which of them the BOP plans
to implement — to obtain accurate and timely jail credit information from judicial as
well as other necessary sources.
Recommendation 5: To address the concerns identified in this report that
conducting final release audits 12 months before an inmate’s projected release date
has led to untimely releases due to BOP staff error, explore and implement
sentence calculation processing or auditing strategies designed to decrease the
likelihood of sentence calculation errors by BOP staff.
Status: Resolved.
ODAG and BOP Response: The BOP stated that it agrees with the
recommendation and that it will explore and implement sentence calculation
processing or auditing strategies designed to decrease the likelihood of sentence
calculation errors by BOP staff.

34

OIG Analysis: The BOP’s planned actions are responsive to the
recommendation. By August 31, 2016, please provide documentation of sentence
calculation processing or auditing strategies that have been implemented and are
designed to decrease the likelihood of sentence calculation errors by BOP staff.
Recommendation 6: Incorporate specific untimely release issues and
possible solutions into education events, and establish a process for obtaining
formal feedback from participants to ensure methods for preventing untimely
releases are fully explored.
Status: Resolved.
ODAG and BOP Response: The BOP stated that it agrees with the
recommendation and that it will incorporate untimely release issues related to staff
error into education events and will request feedback from participants to ensure
methods of preventing such.
OIG Analysis: The BOP’s response is partially responsive to the
recommendation. To be fully responsive to this recommendation, the BOP must
also incorporate into education events any untimely release issues identified in the
ODAG’s multiagency review or by the BOP that are related to non-staff errors. In
fact, because the purpose of these events is to educate representatives from
criminal justice agencies outside of the BOP, these events are specifically an
opportunity to inform representatives from other agencies about what they can do
that would reduce the likelihood of untimely releases. By August 31, 2016, please
provide a list of education events conducted or planned since the issuance of this
report, a list of the untimely release issues incorporated into each event, and
documentation of the process for requesting feedback.
Recommendation 7: Identify and invite participation from additional nonBOP officials who do not typically participate in education events, but whose actions
could help to ensure timely releases, including U.S. District Court Judges, federal
Public Defenders, and U.S. Probation Office staff.
Status: Resolved.
ODAG and BOP Response: The BOP stated that it agrees with this
recommendation and that it will explore identifying and inviting non-BOP
representatives from referenced offices to participate in BOP educational events and
that it will include BOP/DSCC staff in educational events coordinated by those nonBOP offices.
OIG Analysis: The BOP’s response is responsive to the recommendation.
By August 31, 2016, please provide documentation of which non-BOP
representatives the BOP has invited or has identified to invite to BOP educational
events, a list of planned events through 2017, and an indication of which events
coordinated by non-BOP offices the BOP has identified for BOP/DSCC staff to
participate in.
35

The Department of Justice Office of the Inspector General
(DOJ OIG) is a statutorily created independent entity
whose mission is to detect and deter waste, fraud,
abuse, and misconduct in the Department of Justice, and
to promote economy and efficiency in the Department’s
operations. Information may be reported to the DOJ
OIG’s hotline at www.justice.gov/oig/hotline or
(800) 869-4499.

Office of the Inspector General
U.S. Department of Justice
www.justice.gov/oig