United States Government Accountability Office

Report to Congressional Requesters

September 2014

BUREAU OF
PRISONS
Information on Efforts
and Potential Options
to Save Costs

GAO-14-821

September 2014

BUREAU OF PRISONS
Information on Efforts and Potential Options to Save
Costs
Highlights of GAO-14-821, a report to
congressional requesters

Why GAO Did This Study

What GAO Found

BOP is responsible for the custody and
care of 216,000 federal inmates—an
almost 9-fold increase since 1980. At
the same time, BOP appropriations
increased more than 20-fold. DOJ
states that the costs of the growing
population are BOP’s greatest
challenge. BOP’s population size is
driven by several factors, such as law
enforcement policies and sentencing
laws.

Correctional services—which includes salaries and benefits for correctional
officers—is the Department of Justice’s (DOJ) Bureau of Prisons’ (BOP) largest
operational cost, and BOP has undertaken a number of initiatives to reduce
costs. Specifically, on the basis of GAO’s analysis of BOP’s fiscal year 2013
obligations of approximately $6.6 billion, BOP obligated the largest share—about
$3.9 billion, or 59 percent—for personnel compensation and benefits. Further,
BOP has undertaken a number of initiatives, such as renegotiated contracts, to
achieve cost savings of about $61 million over the last 3 years.

GAO was asked to review BOP’s
opportunities to save costs. This report
(1) describes BOP’s major costs and
actions to achieve savings, (2)
assesses the extent to which BOP has
mechanisms to identify additional
efficiencies, and (3) describes potential
changes within and outside of BOP’s
authority that might reduce costs.
GAO analyzed BOP financial data for
fiscal years 2009 through 2013,
reviewed but did not test its internal
control system and processes for
achieving efficiencies, and interviewed
BOP officials. On the basis of
sentencing reform options identified by
experts and actions by the Attorney
General, GAO developed a list of
policy options that could reduce BOP’s
population. GAO gathered views on
their potential effects from entities and
4 states selected for their criminal
justice expertise. The views are not
generalizable, but provide insights.

What GAO Recommends
GAO recommends that BOP establish
a mechanism to consistently monitor if
bureau-wide corrective actions address
repeated deficiencies and findings.
DOJ concurred.

View GAO-14-821. For more information,
contact David C. Maurer at (202) 512-9627 or
maurerd@gao.gov.

BOP has designed internal processes to identify opportunities for additional cost
efficiencies, but could improve the monitoring of corrective actions to achieve
them. For example, BOP focuses on cost efficiency and innovation in its strategic
plan and has developed mechanisms for staff to share information on best
practices and cost savings efforts. BOP also employs an internal control system
with processes, such as program reviews, that allows it to identify opportunities
for cost efficiencies. However, when program reviews repeatedly cited frequent
deficiencies or significant findings that could increase costs—such as insufficient
contract monitoring—responsible BOP Central Office divisions did not provide
specifics or documentation for how they always monitor the effectiveness of
corrective actions to prevent the same deficiency or issue from reoccurring.
Establishing a mechanism for relevant Central Office divisions to consistently
monitor the progress of bureau-wide corrective actions in the presence of
repeated frequent deficiencies or significant findings could help BOP better
ensure that it is resolving such deficiencies or issues promptly and, ultimately,
operating more efficiently.
BOP’s current authorities to reduce inmate sentence length result in limited cost
savings, but potential actions outside of its authority could have a greater impact
on costs. GAO has reported previously on BOP authorities to reduce inmate
sentences, and thus its costs, in detail, and found that inmate eligibility for certain
programs and lack of capacity affect BOP’s use of them. For example, greater
use of programs such as Compassionate Release for terminally ill inmates could
reduce sentences, but cost savings relative to BOP’s budget would be small—
about $651,000 in 2013. Additional opportunities outside of BOP’s authority,
including those requiring legislative or executive action, such as options to
reduce sentence length, could reduce BOP’s population, and thus potentially
significantly reduce its costs. For example, an option to reduce sentences of
incarcerated drug offenders by an average of 44 percent could save about $4.1
billion. Potential savings could be even higher if the changes sufficiently reduced
the inmate population to allow BOP to reduce its staff or close facilities. Expert
entities GAO consulted reported that all of the options GAO reviewed also have
advantages and disadvantages unrelated to costs that should be taken into
consideration, such as potential effects on public safety if released inmates
reoffend. GAO is not taking a position on any of these options, but presents
information on estimated cost savings and experts’ views of advantages and
disadvantages for such options to inform policymakers as they weigh whether
and how to address rising costs at BOP.

United States Government Accountability Office

Contents

Letter

1
Background
Correctional Services Is BOP’s Largest Operational Cost, and
BOP Has Initiated a Number of Efforts to Reduce Costs
BOP Has Designed Internal Processes to Identify Opportunities
for Additional Cost Efficiencies, but Could Improve the
Monitoring of Corrective Actions
Authorities within BOP’s Control to Reduce Costs Are Limited;
Options outside of BOP’s Authority Could Have a Larger Impact
on Costs
Conclusions
Recommendation for Executive Action
Agency Comments and Our Evaluation

8
13
19
28
48
49
49

Appendix I

Objectives, Scope, and Methodology

51

Appendix II

Letter Sent to Select States and Expert Entities

57

Instructions

58

Data on the Federal Bureau of Prisons’ (BOP) Largest
Cost Centers

65

Data on the Federal Bureau of Prisons’ (BOP) Largest
Subobject Class Expenses

66

Additional Federal Bureau of Prisons’ (BOP) Efforts to
Achieve Cost Savings and Efficiencies

68

Expert Entities Providing Input on Prison Population
Reduction Options

70

Appendix III

Appendix IV

Appendix V

Appendix VI

Page i

GAO-14-821 BOP Cost Efficiencies

Appendix VII

Appendix VIII

Analyses of Options Provided by the U.S. Sentencing
Commission (USSC)

71

GAO Contact and Staff Acknowledgments

72

Tables
Table 1: Examples of Cost-Savings Initiatives Adopted by the
Bureau of Prisons (BOP)
Table 2: Bureau of Prisons’ (BOP) Internal Control Processes
That Provide Opportunities to Identify Cost Efficiencies by
Internal Control Standard
Table 3: Key Statutory Provisions Available to the Bureau of
Prisons (BOP) to Reduce a Federal Prisoner’s Period of
Incarceration or Time in BOP Custody
Table 4: Options for Changes to Sentencing Processes, Number
of Eligible Inmates, Estimated Cost of Options, and
Assessment of Potential Advantages and Disadvantages
by Expert Entities and Four States with Sentencing
Change Experience
Table 5: Annual Obligations for Each of the Bureau of Prisons’
(BOP) 20 Largest Cost Centers, Fiscal Years 2009
through 2013
Table 6: Bureau of Prisons’ (BOP) Obligations by Its 20 Largest
Subobject Classes, Fiscal Years 2009 through 2013
Table 7: Additional Cost-Savings Efforts Under Way or Being
Pursued by the Bureau of Prisons (BOP)
Table 8: Expert Entities That GAO Consulted in Reviewing
Selected Policy Options

17
22
29

40
65
66
68
70

Figures
Figure 1: Bureau of Prisons’ (BOP) Fiscal Year 2013 Obligation
Levels by Cost Center
Figure 2: Descriptions of Bureau of Prisons’ (BOP) Largest Cost
Centers
Figure 3: Bureau of Prisons’ (BOP) Fiscal Year 2013 Obligation
Levels by Object Classification

Page ii

14
15
16

GAO-14-821 BOP Cost Efficiencies

Abbreviations:
ABA
American Bar Association
BOP
Bureau of Prisons
DOJ
Department of Justice
FSA
Fair Sentencing Act
GCT
good conduct time
HHS
Department of Health and Human Services
HRMD
Human Resource Management Division
HSD
Health Services Division
IG
Inspector General
IPTP
International Prisoner Transfer Program
JRI
Justice Reinvestment Initiative
OMB
Office of Management and Budget
PRD
Program Review Division
RDAP
Residential Drug Abuse Treatment Program
RRC
residential reentry center
RRM
Residential Reentry Management
RRMB
Residential Reentry Management Branch
RSD
Reentry Services Division
USSC
U.S. Sentencing Commission

This is a work of the U.S. government and is not subject to copyright protection in the
United States. The published product may be reproduced and distributed in its entirety
without further permission from GAO. However, because this work may contain
copyrighted images or other material, permission from the copyright holder may be
necessary if you wish to reproduce this material separately.

Page iii

GAO-14-821 BOP Cost Efficiencies

441 G St. N.W.
Washington, DC 20548

September 30, 2014
Congressional Requesters:
The inmate population at the Department of Justice’s (DOJ) Federal
Bureau of Prisons (BOP)—the primary cost driver of BOP’s budget—has
increased almost 9-fold since the early 1980s—from about 25,000 in
fiscal year 1980 to about 216,000 as of July 2014. Given these population
increases, and inflation, BOP’s appropriations have increased more than
20-fold during the same time period—from $330 million in fiscal year
1980, about 14 percent of DOJ’s total budget that year, to almost $7
billion requested for fiscal year 2015, or about one-quarter of DOJ’s
requested fiscal year 2015 budget. 1 According to the DOJ Inspector
General, the rising costs for BOP threaten the department’s ability to fulfill
its mission in other areas, including maintaining national security,
enforcing criminal law, and defending civil rights. Moreover, DOJ projects
that BOP costs will increase if the federal prison population grows through
2019 as forecast by DOJ in its Fiscal Year 2015 Performance Budget
Congressional Submission for the Federal Prison System. 2

1
BOP per capita annual cost per prisoner increased from about $13,400 in fiscal year
1980 to about $31,338 in fiscal year 2014 in nominal terms. However, after adjustment for
inflation, BOP spending per capita for salaries and expenses in 2014 is about 6.1 percent
less than in 1980. According to the chain-weighted gross domestic product price index, if
BOP spending per capita had kept even with inflation since 1980, annual cost per prisoner
would be about $33,251 for fiscal year 2014 (about 6.1 percent higher than $31,338).
Annual per capita cost per prisoner for fiscal year 2014 was calculated by dividing BOP’s
enacted fiscal year 2014 budget for Salaries and Expenses of $6.769 billion by total
prisoner population of about 216,000 as of July 24, 2014. BOP excludes appropriations for
Buildings and Facilities from its calculation of per capita costs because these may vary
substantially year to year. These costs are also excluded from our calculation of fiscal
year 2014 per capita costs. Prisoner population data are from BOP’s inmate tracking
database, the SENTRY Inmate Management System.
2

The DOJ Fiscal Year 2015 Performance Budget Congressional Submission projects
BOP’s population to increase by 2,500 (about 1 percent) for 2015 over the fiscal year
2014 level—which is forecast to be the same as the fiscal year 2013 level—and by about
8.5 percent by fiscal year 2019 over the July 2014 level. As a result of these projected
increases, system-wide overcrowding is estimated to increase from 36 percent in fiscal
year 2013 to 41 percent by fiscal year 2019. We have previously reported on these issues
in GAO, Bureau of Prisons: Growing Inmate Crowding Negatively Affects Inmates, Staff,
and Infrastructure, GAO-12-743 (Washington, D.C.: Sept. 12, 2012).

Page 1

GAO-14-821 BOP Cost Efficiencies

BOP’s mission is to confine federal offenders in safe and secure prisons
and community-based facilities, and to provide work and other selfimprovement opportunities to assist offenders in becoming law-abiding
citizens. Additionally, one of BOP’s strategic goals is to continually strive
toward improvements in its effective use of resources and its efficient
delivery of services. As of June 2014, BOP operates 120 prison
institutions nationwide, as well as paying for and monitoring 14 prisons
and, as of July 2014, 200 residential reentry centers (RRC) (halfway
houses)—operated by private contractors. According to officials, BOP’s
biggest challenge is managing the increasing federal inmate population,
and related responsibilities, within budgeted levels.
We have previously reported that the size of the federal prison population
is a function of many factors, including the nation’s crime levels, federal
sentencing laws, and law enforcement policies, all of which are beyond
the control of BOP. 3 In addition, the Sentencing Reform Act of 1984
abolished parole for federal offenders, and subsequent legislation
established mandatory minimum sentences for many federal offenses,
facts that limit the authority BOP has to affect the size of the prison
population or the length of prison sentences. 4 BOP has discretionary
authorities, or flexibilities, to reduce a federal prisoner’s period of
incarceration through programs such as the Residential Drug Abuse
Treatment Program (RDAP). 5 BOP also has statutory authority to award
“good conduct time” credit toward reducing an inmate’s sentence up to 54
days per year of sentence served, if the inmate has displayed exemplary

3

See GAO, Bureau of Prisons: Eligibility and Capacity Impact Use of Flexibilities to
Reduce Inmates’ Time in Prison, GAO-12-320 (Washington, D.C.: Feb. 7, 2012).

4

Pub. L. No. 98-473, 98 Stat. 1987. The act was effective for convictions on or after
November 1, 1987. Prior to the act, federal judges generally had broad discretion in
sentencing. Most criminal statutes provided only broad maximum terms of imprisonment.
Federal law outlined the maximum sentence, federal judges imposed a sentence within a
statutory range, and the federal parole official eventually determined the actual duration of
incarceration.
5

18 U.S.C. § 3621(e). We have previously reported on BOP’s use of these discretionary
authorities. See GAO-12-320.

Page 2

GAO-14-821 BOP Cost Efficiencies

compliance with institutional disciplinary regulations. 6 To help alleviate
costs associated with an increasing inmate population and inflation, as
noted in its fiscal year 2015 Performance Budget Congressional
submission, BOP continues to streamline operations and increase
efficiency in order to operate as efficiently and effectively as possible.
Additionally, actions involving policy changes outside of BOP’s authority
could reduce BOP’s inmate population and therefore its costs. These
include potential legislative actions by Congress, executive orders by the
President, policy changes implemented by the Attorney General, or
changes in federal sentencing guidelines by the U.S. Sentencing
Commission (USSC). 7
The Attorney General’s August 2013 Smart on Crime Initiative is an
example of a policy change that will affect the size of the BOP population.
The Attorney General issued a memorandum to the U.S. Attorneys and
Assistant Attorneys for the Criminal Division instructing prosecutors to
decline to charge the quantity necessary to trigger a mandatory minimum
sentence for certain defendants in cases involving the applicability of drug
law mandatory minimum sentences based on drug type and quantity. 8

6

Under 18 U.S.C. § 3624(b), a “prisoner who is serving a term of imprisonment of more
than 1 year other than a term of imprisonment for the duration of the prisoner’s life, may
receive credit toward the service of the prisoner’s sentence, beyond the time served, of up
to 54 days at the end of each year of the prisoner’s term of imprisonment, beginning at the
end of the first year of the term, subject to determination by the Bureau of Prisons that,
during that year, the prisoner has displayed exemplary compliance with institutional
disciplinary regulations.”
7
The USSC is an independent agency in the judicial branch of government. It was created
by the Sentencing Reform Act provisions of the Comprehensive Crime Control Act of
1984, Pub. L. No. 98-473, 98 Stat. 1976. The sentencing guidelines established by the
commission are designed to incorporate the purposes of sentencing (i.e., just punishment,
deterrence, incapacitation, and rehabilitation).
8

According to the memorandum, a defendant must be someone (1) whose relevant
conduct does not involve the use of violence, the credible threat of violence, the
possession of a weapon, the trafficking of drugs to or with minors, or the death or serious
bodily injury of any person; (2) who is not an organizer, leader, manager, or supervisor of
others within a criminal organization; (3) who does not have significant ties to large-scale
drug-trafficking organizations, gangs, or cartels; and (4) who does not have a significant
criminal history. The potential impact of this directive on BOP’s population is discussed
later in this report.

Page 3

GAO-14-821 BOP Cost Efficiencies

Growth in the federal inmate population and its effects on BOP’s budget
are a matter of continuing congressional interest. You asked us to review
the transparency of BOP’s budget justification and opportunities for BOP
to realize operational efficiencies. In December 2013 we reported that by
providing additional funding data to Congress in its budget justification,
BOP could clarify what it proposes to spend on specific categories and
subcategories. 9 We recommended that the Attorney General consult with
congressional decision makers on providing additional BOP funding detail
in future budget justifications, and in conjunction with BOP, take action as
appropriate. DOJ concurred with this recommendation and is taking
action to implement it. This report responds to the latter part of your
request by further describing BOP’s major costs and the actions it has
taken to identify and implement cost efficiency opportunities and reduce
costs. Specifically, this report addresses the following questions:
1. What are the major costs for BOP operations, and what actions has
BOP taken to implement cost savings?
2. To what extent does BOP have mechanisms to identify opportunities
for cost efficiencies and to take corrective actions that may improve
cost efficiency?
3. What potential changes, both within and outside of BOP’s authority,
could lead to cost reductions or improved efficiencies in BOP
operations, and what might be the potential impact of these changes?
To address the first question, we obtained historical obligation data from
BOP’s Financial Management Information System for fiscal years 20092013—a 5-year time frame to provide us with an adequate understanding
of trends in BOP obligation levels—to identify the major costs incurred by
BOP for its Salaries and Expenses account, which generally represents
98 percent of BOP’s budget. 10 We have assessed the reliability of these
data and determined them to be reliable for the purposes of this report.
This assessment included performing checks on the data received and
interviewing officials responsible for compiling and maintaining these
data. We also collected data (e.g., cost savings estimates prepared by

9
GAO, Bureau of Prisons: Opportunities Exist to Enhance the Transparency of Annual
Budget Justifications, GAO-14-121 (Washington, D.C.: Dec. 6, 2013).
10

BOP is appropriated funds through two accounts: Salaries and Expenses and Buildings
and Facilities. We focused our review on the Salaries and Expenses account as it
represents almost the entirety of BOP’s budget.

Page 4

GAO-14-821 BOP Cost Efficiencies

BOP) and documentation, such as memos and concept papers approved
by executive staff from fiscal years 2009 through 2013; through review of
these data and documents and interviews with relevant agency officials,
we determined that these data are also reliable for the purposes of this
report. In addition, we interviewed Central Office-, regional-, and
institutional-level officials to identify to the extent possible all existing cost
efficiency and savings initiatives adopted by BOP and their impact on its
overall budget.
To address the second question, we reviewed the processes and tools at
BOP during the same time period (fiscal years 2009-2013) that are used
to identify, implement, and promote cost-efficiency and savings initiatives
throughout its institutions, such as executive staff meetings and a catalog
compiling cost-savings initiatives. With respect to identifying additional
opportunities to realize cost efficiencies or reduce costs, using our
financial analysis as context, we analyzed elements of BOP’s internal
control system related to the control objective of achieving operational
efficiencies and interviewed relevant officials to assess whether BOP has
designed a management structure and processes to routinely assess its
administrative and operational activities for possible corrective action.
Specifically, we reviewed BOP’s mechanisms and processes leading to
its internal review of operational and administrative functions, including its
process for taking corrective action related to high-cost areas, and
compared these characteristics with those called for in Standards for
Internal Control in the Federal Government. 11
To address the third question, on potential policy changes both within and
outside BOP’s authority that could lead to BOP cost savings, we collected
and reviewed analysis and documentation from DOJ, BOP, the USSC,
and entities that we selected for their expertise in criminal justice issues,
and, in particular, potential changes to federal sentencing policies, and to
ensure that a wide range of views on the advantages and disadvantages

11
GAO, Standards for Internal Control in the Federal Government, GAO/AIMD-00.21.3.1
(Washington, D.C.: Nov. 1, 1999). Internal control is an integral component of an
organization’s management that provides reasonable assurance that the following
objectives are being achieved: effectiveness and efficiency of operations, reliability of
financial reporting, and compliance with applicable laws and regulations.

Page 5

GAO-14-821 BOP Cost Efficiencies

of these potential changes. 12 These entities may not be representative of
the universe of expert entities in the criminal justice field and therefore
may not represent all views on this topic; however, their views provide
insights. We obtained data from BOP for 2013 and 2014 on the impact on
BOP’s population and costs resulting from changes within BOP’s
authority. We assessed the reliability of these data by reviewing
documentation and interviewing knowledgeable officials, and determined
the data to be sufficiently reliable for the purposes of this report.
Further, on the basis of our assessment of proposals put forward by
several sources, including the Attorney General, expert entities, states
with experience with criminal justice reform, and advocacy entities with a
range of views on criminal justice reform, we identified eight options
outside of BOP’s authority that could reduce the size of the inmate
population. GAO is not taking a position on any of these options, but
presents information on estimated cost savings and experts’ views of
advantages and disadvantages for such options to inform policymakers
as they weigh whether and how to address rising costs at BOP.
For each of these eight options, we asked the USSC to estimate the
impact on the size of the BOP inmate population and, using the USSC
estimates, we then calculated cost savings or cost avoidance that could

12

The expert sources from which we requested or obtained analysis or information on
sentencing issues were: the American Bar Association (ABA); the Association of State
Correctional Administrators; the Brennan Center for Justice; the Council of State
Governments/Justice Center; Families Against Mandatory Minimums; the Fraternal Order
of Police; the Heritage Foundation; the National Association of Assistant U.S. Attorneys;
the National Association of Former U.S. Attorneys; the Pew Charitable Trusts; “Right on
Crime”, a project of the Texas Public Policy Foundation in cooperation with the Justice
Fellowship; the Sentencing Project; the Urban Institute; and the Vera Institute of Justice.
We identified these expert entities through several means, including (1) asking officials at
both the USSC and the Urban Institute who have previously worked on analyses and
reports relevant to prison population data and criminal justice to identify entities they
considered as expert in the field of sentencing reform and criminal justice, (2) conducting
a literature search to identify publications issued by some of the entities on sentencing
reform or corrections, and (3) identifying entities with known expertise and
authoritativeness (e.g., the ABA) in the criminal justice field through literature searches.

Page 6

GAO-14-821 BOP Cost Efficiencies

result from implementing a given option. 13 According to our audit
objectives, we determined that based on the expertise of the USSC, the
evidence obtained from the USSC provides a reasonable basis for our
findings and conclusions.
In addition, we asked the 14 expert entities to comment on the options
with regard to their view of the advantages or disadvantages of
implementation of the eight options. We also asked 4 states to comment
on the potential options for change. One of the 4 was Texas, which had
the largest number of incarcerated offenders in the nation (as of 2012)
and which was identified as a model for changes in state criminal justice
systems by the Justice Reinvestment Initiative (JRI), a public-private
partnership of DOJ and the Pew Charitable Trusts, to provide technical
assistance and financial support for system-wide criminal justice reform
efforts. 14 The other states were chosen on the basis of having had at least
1 year’s experience in implementing JRI programs; of these 8, we chose
the 3 that had the largest annual corrections budgets as of the year in
which JRI started in those states. We received acknowledgement of
receipt from 17 of the 18 entities (the 14 expert entities and 4 states) to
which we sent the letter. 15
Further details on our scope and methodology are contained in appendix
I. See appendix II for the letter sent to expert entities asking for their

13

The USSC calculates the impact of sentencing changes in terms of bed year savings; 1
bed year is defined by the USSC as 12 months of prison time (e.g., one offender in prison
for 1 year, or three offenders in prison for 4 months each, and so forth). Using the USSC
estimates of saved bed years, we calculated savings that would result from sentence
length reductions or from cost avoidance savings resulting from increases below what
would otherwise occur if an option was not implemented, by multiplying bed years by the
annual marginal cost for a prisoner. Marginal costs, calculated by BOP, cover such things
as food, medical care, clothing, and utilities. For fiscal year 2014, the average annual
marginal cost per offender is about $11,000. For some options that could affect the
sentences of about 78,000 or more offenders, equal to the BOP July 2014 level of systemwide overcrowding, we also calculated per capita savings, as such sentence reductions
could allow BOP to reduce its staff or close facilities.

14

Texas was identified as a model for criminal justice system reforms in a joint January
2014 DOJ-Urban Institute report assessing the JRI initiative. See: Urban InstituteDepartment of Justice Bureau of Justice Assistance, Justice Reinvestment Initiative State
Assessment Report (Washington, D.C.: January 2014).

15

Of the 17 entities that acknowledged receipt of our request for comments, 3 did not
provide substantive comments. Of these 3, 1 referred us to multiple articles published on
criminal justice issues.

Page 7

GAO-14-821 BOP Cost Efficiencies

assessment of options, which provides a list of the proposals that were
used to develop the selected options.
We conducted this performance audit from July 2013 to September 2014
in accordance with generally accepted government auditing standards.
Those standards require that we plan and perform the audit to obtain
sufficient, appropriate evidence to provide a reasonable basis for our
findings and conclusions based on our audit objectives. We believe that
the evidence obtained provides a reasonable basis for our findings and
conclusions based on our audit objectives.

Background
BOP’s Cost-Accounting
Methods

BOP uses a network of cost centers—individual accounting units in which
costs can be segregated—to obligate, account for, and manage various
costs. 16 Managers of each cost center are provided with information to
control and manage costs within their area of responsibility. Cost centers
can be structured along different dimensions, such as organizational
units, operating processes, and activities. 17 For example, BOP cost
centers include Correctional Services, Inside Medical Services, and Food
Services at each institution.

16

An obligation is a definite commitment that creates a legal liability of an agency for the
payment of goods or services received. An agency incurs an obligation, for example,
when it places an order, signs a contract, awards a grant, purchases services, or takes
other actions that require the government to make payments to the public or from one
government account to another. See GAO, A Glossary of Terms Used in the Federal
Budget Process, GAO-05-734SP (Washington, D.C.: September 2005). To report and
monitor obligated funds throughout the bureau, BOP employs cost center accounting and
designates managers for these cost centers. According to the Federal Accounting
Standards Advisory Board, components of an organization must accumulate costs for
each type of output they produce for various programs.
17

Statement of Federal Financial Accounting Standards 4: Managerial Cost Accounting
Standards and Concepts. Federal Accounting Standards Advisory Board Handbook,
Version 12 (June 2013).

Page 8

GAO-14-821 BOP Cost Efficiencies

BOP also tracks its obligations and expenditures by budget object and
subobject classification. 18 Object classes are categories used in budget
preparation to classify obligations by the items or services purchased by
the federal government (e.g., personnel compensation, contractual
services). Within these object classes, the Department of Justice has
assigned subobject class codes that further define the nature of services
or articles obligated and provide DOJ with a detail of expended funds. For
example, under the personnel compensation object class, BOP has
defined subobject class codes for items such as compensation for fulltime, permanent appointment employees; overtime; and holiday pay.
Object and subobject classes are used for direct tracking of specific items
on which funds are being spent (e.g., salaries, supplies, pharmaceuticals,
etc.) and are used across all cost centers.

BOP Organizational
Structure

To carry out its responsibility for the custody and care of federal
offenders, as of June 2014, BOP houses inmates across six geographic
regions in 120 federal institutions, 14 privately operated secure contract
facilities, home detention, and as of July 2014, 200 residential reentry
centers. 19 These federal institutions and other facilities are managed and
overseen through the following offices:
•

Central Office. The Central Office, which serves as BOP’s
headquarters, consists of nine divisions that provide oversight of
major BOP program areas and operations, such as correctional
programs and health services, as well as the National Institute of
Corrections. 20 BOP’s Program Review Division, which leads BOP’s
process for conducting internal reviews of each program or operation
at each BOP institution, is among these nine divisions.

18

In order to standardize the budgetary information received from federal agencies, the
Office of Management and Budget (OMB) assigns object classifications according to the
nature of the services or articles procured (e.g., Personnel Compensation; Supplies and
Materials; Rent, Communications, and Utilities; etc.).

19

According to BOP officials, privately operated secure contract facilities are low security
and primarily house non-U.S. citizens convicted of crimes while in this country legally or
illegally. Regarding the RRCs, BOP officials stated that they use an additional 49 RRCs
through intergovernmental agreements for work release purposes. Home detention
describes all circumstances under which an inmate is serving a portion of his or her
sentence while residing in his or her home.

20

National Institute of Corrections, a component of BOP, provides training, technical
assistance, information services, and policy/program development assistance to federal,
state, and local corrections agencies.

Page 9

GAO-14-821 BOP Cost Efficiencies

•

Grand Prairie Office Complex. This complex, located in Grand
Prairie, Texas, was created to consolidate various processes into one
location. It consists of three components: human resource services,
designation and sentence computation, and field acquisitions
exceeding $100,000.

•

Regional offices. BOP has six regional offices, each led by a
regional director, covering the Mid-Atlantic, North Central, Northeast,
South Central, Southeast, and Western regions of the United States.
These offices provide oversight, technical assistance, and training to
the BOP facilities located within each region.

•

Residential Reentry Management field offices. BOP has 25 field
offices that administer contracts for community-based programs and
serve as liaisons with various federal, state, and local groups within
their judicial districts. Staff at these offices also monitor residential
reentry centers (RRCs), which provide offenders with communitybased services to assist with their reentry needs.

BOP generally houses sentenced inmates in its long-term institutions.
Male long-term institutions include four security level designations––
minimum, low, medium, and high––and female institutions include three
security designations––minimum, low, and high. The security-level
designation of a facility depends on the level of security and staff
supervision that the institution is able to provide, such as the presence of
security towers; perimeter barriers; the type of inmate housing, including
dormitory, cubicle, or cell-type housing; and inmate-to-staff ratio.
Additionally, BOP designates some of its institutions as administrative
institutions, which house male and female inmates and specifically serve
inmates awaiting trial, or those with intensive medical or mental health
conditions, regardless of the level of supervision these inmates require.

The Sentencing Reform
Act of 1984 and Federal
Sentencing Guidelines

Prior to passage of the Sentencing Reform Act of 1984, federal judges
generally had broad discretion in sentencing. 21 Most criminal statutes
provided only broad maximum terms of imprisonment. Federal law
outlined the maximum sentence, federal judges imposed a sentence
within a statutory range, and the federal parole official eventually
determined the actual duration of incarceration. The Sentencing Reform

21

Pub. L. No. 98-473, 98 Stat. 1987.

Page 10

GAO-14-821 BOP Cost Efficiencies

Act of 1984 changed the federal sentencing structure. The act was
effective for offenses committed on or after November 1, 1987. The act
abolished parole for federal offenders sentenced after its effective date,
and subsequent legislation established mandatory minimum sentences
for many federal offenses.
The Sentencing Reform Act of 1984 also established the independent
USSC within the judicial branch and charged it with, among other things,
developing federal sentencing guidelines. 22 USSC establishes sentencing
policies and practices for the federal criminal justice system that provide
certainty and fairness in meeting the purposes of sentencing, avoiding
unwarranted sentencing disparities among defendants with similar
criminal records who have been found guilty of similar criminal conduct
while maintaining sufficient flexibility to permit individualized sentences
when warranted. 23 To foster this goal, the guidelines specify sentencing
guideline ranges—a range of time (in months) that offenders should serve
given the nature of their offense and other factors—but also permit
sentences to depart upward or downward from guideline ranges because
of aggravating or mitigating circumstances. In 2005, the Supreme Court
found the Sentencing Guidelines, which had previously been binding for
federal judges to follow in sentencing criminal defendants, to be advisory
in nature. 24 Regardless of the guidelines’ advisory nature, judges are still
required to calculate properly and consider the Sentencing Guidelines
and other sentencing goals. 25
In applying the USSC’s guidance, federal district judges are to determine
the appropriate sentencing guideline range for an offender based on
various factors related to (1) the offense and (2) the offender. The offense
is assigned an offense level; drug offenses are based on several factors,
such as the quantity and type of drug involved and whether the offense
involved violence.

22

Pub. L. No. 98-473, § 217, 98 Stat. 1987, 2017.

23

28 U.S.C. § 991(b)(1)(B).

24

See United States v. Booker, 543 U.S. 220 (2005).

25

See 18 U.S.C. § 3553(a).

Page 11

GAO-14-821 BOP Cost Efficiencies

The offender is also assigned a criminal history category based on the
number of criminal history points. USSC sentencing guidelines assign
criminal history points based on a defendant’s criminal record:
•

A single point is assigned for every other federal or state prior
sentence of conviction, subject to certain exceptions. 26

•

Two or more points are assigned for every prior sentence of
imprisonment or juvenile confinement of 60 days or more or for
offenses committed while the defendant was in prison; was an
escaped prisoner; or was on probation, parole, or supervised release;

•

Three points are assigned if an offender has been previously
sentenced or is being sentenced to a prison term of 13 months or
more.

Taken in combination, the offense level and criminal history category
correlate with a sentencing guideline range, expressed in months. For
certain types of offenses, including certain drug and weapons offenses,
statutorily specified mandatory minimum sentences supplant the lower
end of the otherwise applicable guidelines range. While federal law
generally requires a sentencing judge to impose a minimum sentence of
imprisonment following conviction for any of a number of federal offenses,
there are two statutorily provided exceptions. One—the substantial
assistance exception—is available upon motion of the government, and
allows the court to impose a sentence below the level established by
statute as a minimum to reflect a defendant’s substantial assistance in the
investigation or prosecution of another person who has committed an
offense. 27 The other, commonly referred to as the safety valve, was
created by the Violent Crime Control and Law Enforcement Act of 1994

26

Some types of sentences are not counted in calculating criminal history points. These
are: foreign sentences of imprisonment; sentences imposed by tribal courts; summary
court martial sentences; sentences imposed for expunged, reversed, vacated, or
invalidated convictions; sentences for certain petty offenses or minor misdemeanors. The
Sentencing Guidelines list two classes of these minor misdemeanor or petty offenses that
are not counted for criminal history purposes. One class consists of eight types of minor
offenses—such as hunting and fishing violations or juvenile truancy— that are not counted
regardless of the sentence imposed.

27

18 U.S.C. § 3553(e).

Page 12

GAO-14-821 BOP Cost Efficiencies

and is available for certain types of defendants for certain drug offenses
that carry minimum sentences. 28

Correctional Services
Is BOP’s Largest
Operational Cost, and
BOP Has Initiated a
Number of Efforts to
Reduce Costs
BOP’s Correctional
Services, Private Prison
Contracts, Medical Care,
and Food Services
Account for the Majority of
Its Costs, with the Largest
Share of Spending on
Personnel

During fiscal year 2013, BOP obligated approximately $6.6 billion for all
its programs and operations. According to our analysis of BOP’s fiscal
year 2013 obligations by cost center, correctional services, privately
operated secure contract facilities, medical care, and food services
represent the majority (approximately $3.56 billion, or 54 percent) of
BOP’s expenses, as illustrated in figure 1. See appendix III for a breakout
of cost trends in these and other key cost centers over the last 5 fiscal
years.

28
18 U.S.C. § 3553(f). For the safety valve exception, the court is to consider at
sentencing (1) the defendant’s criminal history; (2) whether the defendant used violence or
credible threats of violence or possessed a firearm or other dangerous weapon in
connection with the offense; (3) whether the offense resulted in death or serious bodily
injury to any person; (4) whether the defendant was an organizer, leader, manager, or
supervisor of others in the offense or was engaged in a continuing criminal enterprise; and
(5) the defendant has truthfully provided to the government all information and evidence
the defendant has concerning the offense or offenses that were part of the same course of
conduct or of a common scheme or plan.

Page 13

GAO-14-821 BOP Cost Efficiencies

Figure 1: Bureau of Prisons’ (BOP) Fiscal Year 2013 Obligation Levels by Cost
Center

Notes: The budget data for all cost centers include all staff salaries, operational supplies, contract
services, equipment, and other items related to that department. BOP organizes its cost centers
based on location within the organization; thus, there are cost centers at the institution level, regional
level, national training level, national program level, Central Office level, and contract confinement
level. Institution-level cost centers—which represent the majority of BOP obligations—are further
broken down by security level (e.g., high, medium, low, etc.). For the purposes of our review, we
combined the various institution security-level cost centers into common cost centers for all
institutions. This allows us to present data on, for example, correctional services, BOP medical care,
and so forth, for all institutions BOP-wide.
According to BOP, it considers some funds obligated within the “Other cost centers” category to be
part of the larger cost centers displayed in figure 1. Including BOP’s estimates of these funds in the
totals of the larger cost centers listed above would change totals approximately as follows:
•

Correctional services: An additional $100 million if special projects, urinalysis surveillance,
clothing allowances, and other items were included.

•

Medical services (both BOP and non-BOP): An additional $47 million if referral labs,
dialysis, emergency airlift, and other items were included.

•

Food service: An additional $10 million if farm and special projects were included.

•

Private operated secure facilities: An additional $11 million if BOP privatization staff
salaries were included.

•

Recategorizing these obligations would subsequently reduce the Other cost centers
category to approximately $2.8 billion, or 42% of totals for fiscal year 2013.

In addition, according to BOP, staff at institutions—regardless of the cost center functions that they
primarily serve—are trained as correctional officers and thus may be called upon to assist in
correctional services or other functions whenever necessary.
a
The BOP medical care cost center also contains approximately $103 million in funds transferred to
pay salaries and other costs associated with Public Health Services staff. According to BOP, these
are medical professionals employed by the Department of Health and Human Services (HHS) who

Page 14

GAO-14-821 BOP Cost Efficiencies

provide medical services to BOP institutions. While these Public Health Services staff are technically
paid by HHS, BOP transfers the necessary funds to HHS to pay for their services.
b
The “other cost centers” category contains a number of cost centers that together account for 46
percent of BOP’s total costs, as shown above; however, individually, each cost center included in the
“other” category accounts for 5 percent or less of BOP’s total costs

Figure 2 provides further information on the key cost centers represented
above.
Figure 2: Descriptions of Bureau of Prisons’ (BOP) Largest Cost Centers

Within its cost centers, the majority of costs across the bureau are
payment of staff salaries and benefits. Detailed data on these expenses
are tracked using budget object and subobject classifications. These
budget classifications are used consistently across all cost centers and
represent items or services actually purchased by BOP. Figure 3 shows

Page 15

GAO-14-821 BOP Cost Efficiencies

the budget object classifications that accounted for BOP’s total fiscal year
2013 obligations.
Figure 3: Bureau of Prisons’ (BOP) Fiscal Year 2013 Obligation Levels by Object
Classification

Notes: Totals do not add up to 100 percent because of rounding.
Included in these obligations are approximately $103 million in funds transferred to pay salaries and
other costs associated with Public Health Services staff. According to BOP, these are medical
professionals employed by the Department of Health and Human Services (HHS) who provide
medical services to BOP institutions as needed. While these Public Health Services staff are
technically paid by HHS, BOP transfers the necessary funds to HHS to pay for their services.
a

Personnel compensation and personnel benefits are actually two separate object classifications, but
were combined for the purposes of this figure.

b
”Other contractual services” covers obligations for contractual services for advisory and assistance
services (including contract staff), purchases of goods and services from government accounts;
operation and maintenance of facilities and equipment; payments for medical care, research and
development contracts; subsistence and support of persons; and other services not otherwise
classified.
c

”Supplies and materials” covers obligations for commodities whether acquired by formal contract or
other form of purchase that are: ordinarily consumed or expended within 1 year after they are put into
use, converted in the process of construction or manufacture, or used to form a minor part of
equipment or fixed property. Examples include food, pharmaceuticals, clothing, and office supplies.

d

Other object classifications include equipment, travel and transportation of persons, and other items.

As figure 3 shows, personnel compensation and benefits (retirement
contributions, health insurance benefits, etc.) account for approximately
59 percent of BOP’s overall obligations. Moreover, BOP’s personnel costs
are actually higher than 59 percent because compensation for other staff
employed on a contractual basis is included in the “other contractual
services” object class. This object class consists of privately operated

Page 16

GAO-14-821 BOP Cost Efficiencies

prison expenses, outside medical services, and other items—as well as
contract staff—and accounts for nearly 26 percent of fiscal year 2013
obligations. See appendix IV for more detailed data on object and
subobject class obligations from fiscal years 2009 through 2013.

BOP Has Identified and
Implemented Various Cost
Savings Efforts

BOP’s Director, in an April 2014 statement before Congress, highlighted
BOP’s commitment to stewardship of financial resources by seeking cost
avoidance and greater efficiencies throughout the bureau. 29 In
accordance with this focus, BOP has embarked on a number of initiatives
in recent years that have resulted in actual or projected cost savings. Key
initiatives for which BOP has estimated cost savings over the last 3 fiscal
years are listed in table 1.

Table 1: Examples of Cost-Savings Initiatives Adopted by the Bureau of Prisons (BOP)
Estimated cost savings

a

Cost savings initiative

Description

Renegotiated medical contracts

BOP’s institutions attempt to contract with
BOP reports that from April 2013 through April
medical service providers to set a fee
2014, it has saved nearly $33 million through
structure for necessary services that BOP
renegotiated contracts.
cannot provide in-house. These contracts
typically establish a rate of payment for
services at a certain premium over a
benchmark, such as standard Medicare rates.
Savings are generated by trying to reduce
premiums paid over the Medicare rate for
many BOP medical services contracts.

Industries, Education, and
Vocational Training efficiency
efforts

Industries, Education, and Vocational Training
Division recorded cost-saving items (most of
which are related to Federal Prison Industry
operations) completed during fiscal year 2013.
These items represent cost savings efforts
identified through the use of Lean Six Sigma
methodologies, such as changes to
b
warehouse lighting or materials processing.
BOP reports that it recently developed a
process to validate the actual cost savings
achieved through these efforts, and plans to
validate cost savings for fiscal year 2014
efforts.

BOP reports estimated savings of about $14
million in nonappropriated funds (i.e., the
costs of operating Federal Prison Industry
programs) during fiscal year 2013.

29

Statement of Charles E. Samuels, Jr., Director of the Federal Bureau of Prisons, before
the United States House Committee on Appropriations, Subcommittee on Commerce,
Justice, Science and Related Agencies, April 10, 2014.

Page 17

GAO-14-821 BOP Cost Efficiencies

a

Cost savings initiative

Description

Estimated cost savings

Administrative efficiencies

During fiscal year 2013, BOP conducted a
reorganization of its regional offices and
administrative operations, which resulted in a
reduction of 60 positions.

BOP reports salary savings of approximately
$7.2 million in fiscal year 2013, with additional
operational savings.

National Paving Project

Minimum security inmates from one BOP
institution have performed road-paving
services at 30 BOP institutions over the last
2 1/2 years, in lieu of hiring contractors to
perform such work.

BOP estimates that this project has saved a
minimum of about $2 million versus the cost of
employing a contractor, and officials believe
that actual savings have been higher when
taking into account factors such as markup
and overhead associated with the use of a
contractor.

Information technology savings

BOP is involved with a Department of Justice
(DOJ)–level working group to identify areas for
savings by combining the purchasing power of
DOJ components. For example,
BOP is using DOJ’s enterprise license
agreements for Microsoft, Oracle, and Adobe
products.

BOP reports that through use of another DOJ
component’s delivery order contract, it has
saved about $2.4 million in purchase and
maintenance agreements. BOP does not track
cost savings through the aforementioned
enterprise license agreements.

Food digester

A machine that converts food waste into
gray water, which is disposed of through the
institution’s sewer system, rather than at a
landfill, saving BOP on landfill costs. BOP
currently has 94 units in place at various
institutions, with another 10 awaiting
installation.

According to data provided by BOP, we
calculate that these units may save
approximately $1.8 million annually across
BOP.

Telemedicine

Use of videoconferencing to provide clinical
and consultative services in psychiatry and
dermatology. This enables BOP to have a
physician on contract at any location to
perform these services, rather than
contracting to bring in specialists for on-site
c
consultations.
BOP also contracts with a company to provide
remote interpretations of inmate X-rays, rather
than paying for the service on-site. This
approach has been in place since 2004.

BOP estimates that it saved approximately
$448,000 in calendar year 2013 through
videoconferencing for dermatology services. It
also states that it can make more costeffective decisions on medications and reduce
referral and treatment times for inmates.
BOP estimates that in fiscal year 2012 (the
last year for which data are available), it was
able to avoid approximately $387,000 in costs
at 86 institutions through this “teleradiology”
approach.

Source: GAO analysis of BOP information. | GAO-14-821
a

All cost estimates were provided by BOP. We assessed the reliability of these estimates and found
them sufficiently reliable to present as estimated savings.

b

Federal Prison Industries is a federal government corporation overseen by BOP that, among other
things, employs and provides job skills training to a number of federal inmates, and produces goods
and services for sale to the federal government. Lean Six Sigma is a data-driven approach to process
improvement based on the idea of eliminating defects and errors that contribute to losses of time,
money, opportunities, or business.

c
BOP states that it has a total of 14 psychiatrists (filling 20 allotted positions) providing virtual
psychiatric services to inmates at 90 BOP institutions. The dermatologist participating in this program
conducted 501 consultations with clinicians at 50 institutions during calendar year 2013.

Page 18

GAO-14-821 BOP Cost Efficiencies

These efforts reflect a focus on the part of both BOP management and
staff to identify and implement cost savings initiatives, on both a large and
a small scale. BOP-developed estimates of savings from these efforts
total about $61 million, which is just under 1 percent of total obligations
for fiscal year 2013. BOP officials stated that since such a large
component of BOP’s costs is related to paying salaries and benefits
necessary to manage its inmate population, their ability to recognize cost
savings beyond those efforts presented here is fairly limited.
In addition, BOP is in varying stages of taking additional actions such as
converting inmate case files from paper to electronic records that may
reduce costs—for which it has not determined or provided savings
estimates—that can be found in appendix V. According to BOP officials,
they may undertake such initiatives in order to operate more efficiently,
and these initiatives may also result in immediate cost savings. However,
the officials said that they do not always develop cost savings estimates
for initiatives whose primary purpose is to improve efficiency, rather than
achieve cost savings.

BOP Has Designed
Internal Processes to
Identify Opportunities
for Additional Cost
Efficiencies, but
Could Improve the
Monitoring of
Corrective Actions
BOP Has Designed
Mechanisms to Identify
Opportunities for Cost
Efficiencies

BOP has various mechanisms to identify opportunities for cost
efficiencies, including the following:
•

Strategic Plan. BOP’s Strategic Plan helps guide its efforts in fulfilling
its mission and achieving strategic goals. As one strategic goal, BOP
states that it will manage its operations and resources in an effective
manner that encourages innovation, and that it “continually strives

Page 19

GAO-14-821 BOP Cost Efficiencies

toward improvements in its effective use of resources and its efficient
delivery of services.” 30 Further, BOP has developed a strategic
objective focused on cost efficiency and innovation, stating that it will
continue to focus on reducing costs using efficient and cost-effective
methods to perform tasks. According to the Strategic Plan, it will
accomplish this through, in part, “continued emphasis on financial
planning, analyzing workload and staffing requirements, using
consolidated and shared services, increasing the use of technology,
and refining the processes of the BOP.” In support of this strategic
objective, BOP requires that institutions and regions submit to Central
Office established local best practices or cost savings measures that
can be shared and replicated at other institutions.
•

Cost Efficiencies and Innovations Catalog. This product is a
compendium of institution-level efficiency efforts and innovations that
serves as a reference for other institutions. The catalog is a result of
efforts to identify cost efficiencies as stated in BOP’s Strategic Plan.
To compile this catalog, BOP’s Central Office surveys institutions
twice a year regarding new or ongoing efforts that they consider to be
local best practices in achieving cost efficiencies. After compiling the
catalog, BOP distributes it bureau-wide, including to program
managers at all institutions. Some examples of cost-efficiency efforts
identified and shared throughout BOP via this catalog include efforts
in recycling material and food waste, water and energy conservation,
use of video conferencing, and inmate training and education.
Institutional officials we spoke with reported that they were aware of
and reviewed this catalog. BOP institution-level staff estimated that
the catalog issued in March 2013 contained items saving tens of
millions of dollars in expenses. 31

•

Best practices internal site. In April 2007, BOP implemented an
internal information-sharing site designed to facilitate bureau-wide
communication of efforts that have resulted in cost efficiencies to
bureau staff or inmates. This reference tool includes links to cost
reduction initiatives, energy conservation efforts, and other items that
have been reviewed and vetted by subject matter experts as having
demonstrated their value to BOP. The website includes institution-led

30

BOP, BOP Strategic Plan, (Washington, D.C.: February 2014).

31

The catalog contains specific dollar amounts for estimated savings by institution, but
these figures are not verified by BOP’s Central Office or other officials.

Page 20

GAO-14-821 BOP Cost Efficiencies

efforts such as battery-recycling operations and improved efficiencies
to water-heating systems.
•

General staff/program area meetings. Officials at BOP’s Central
Office, regional offices, and selected institutions with whom we spoke
all cited frequent communication between and among program
officials and executive staff regarding items such as general
operations, budget and cost levels, staffing levels and use of
overtime, health care topics, and other items. These officials
frequently cited the quarterly executive staff meeting, which includes
assistant directors from Central Office and the six regional directors,
as a key meeting for these discussions. At these meetings, among
other things, officials present and discuss position papers that lay out
information on initiatives that may have the potential for cost savings.
For instance, at one meeting, the participants discussed a paper
describing ongoing efforts to convert BOP’s Inmate Case Files from a
paper-based system to an electronic system, ultimately saving staff
resources.

•

Internal control. Internal control is an integral component of an
organization’s management that is to provide reasonable assurance
that objectives are achieved, including the efficiency of operations.
We focused our review on BOP’s internal control objective related to
achieving operational efficiencies. Given this, we reviewed the
elements of BOP’s internal control system that specifically provide
BOP with opportunities to identify cost efficiencies, and found that
BOP’s internal control processes align with Standards for Internal
Control in the Federal Government. Table 2 provides examples and
descriptions of BOP’s specific internal control processes, organized
by internal control standard, that allow it to identify opportunities for
cost efficiencies. We did not independently test BOP’s internal
controls to determine whether they mitigate all possible risks and are
operating as intended.

Page 21

GAO-14-821 BOP Cost Efficiencies

Table 2: Bureau of Prisons’ (BOP) Internal Control Processes That Provide Opportunities to Identify Cost Efficiencies by
Internal Control Standard
Internal control standard

BOP process

Environment
Management and employees should
establish and maintain an
environment throughout the
organization that sets a positive and
supportive attitude toward internal
control and conscientious
management. It provides discipline
and structure as well as the climate
which influences the quality of internal
control.

•

Central Office
oversight

BOP Central Office’s Program Review Division oversees compliance
with the Office of Management and Budget’s (OMB) Circular A-123
and conducts the internal review of administrative and operational
a
functions bureau-wide.
As noted above, BOP’s Strategic Plan also includes a strategic goal
to continually strive toward improvements in effective use of
resources and efficient delivery of services.
As noted above, BOP’s Director has stated BOP’s commitment to
stewardship of financial resources. For instance, in BOP’s fiscal year
2015 Budget Request, the Director stated that the agency will
continue to seek cost avoidance and find efficiencies throughout the
agency.

Risk assessment
Internal control should provide for an
assessment of the risks the agency
faces from both external and internal
sources.

•

Control Evaluation
Template
Management
Assessment

The Control Evaluation Template is a documented list of controls, and
the risk each control is intended to mitigate, in compliance with OMB
Circular A-123. For example, one identified risk is an error in
payment.
BOP’s Management Assessment process includes documenting risks
within BOP’s major program areas (e.g. Food Services). Assessing
risk informs the development of internal control and, consequently,
guides the internal control review process. For instance, one
identified risk is the probability of not conducting inventory properly,
which could affect the accuracy of data used when formulating and
submitting a budget. For each risk assessment tool, BOP assesses
the magnitude and likelihood of each risk.

Control activities
Control activities help ensure that
management’s directives are carried
out. The control activities should be
effective and efficient in
accomplishing the agency’s control
objectives, including achieving
efficiencies.

•

Control Evaluation
Template
Program
statements
(policy)
Program review
guidelines
(Internal control
testing plan)

The Control Evaluation Template documents internal controls,
designed to mitigate identified risks. The template also includes
control testing results, in compliance with OMB Circular A-123. For
example, certain BOP management review and approval chains are
designed to mitigate the risk of improper payment.
Program statements are internal policy documents that outline
program requirements, including key internal control activities. For
instance, program statements note that staff are to monitor the
performance of and payments to contractors.
Program review guidelines serve as a guide to conducting internal
reviews of program areas’ operations and administrative activities, to
ensure that key internal controls are strong and operated effectively
(i.e., internal control testing). Program reviews document the results
of this internal review process. For example, the guideline directs the
reviewer to assess whether and how well staff conduct inventory and
the program review documents the results of that assessment.

•

•

•

Page 22

BOP’s use in identifying opportunities for cost efficiencies

GAO-14-821 BOP Cost Efficiencies

Internal control standard

BOP process

Information and communication
Information should be recorded and
communicated to management and
others within the entity who need it
and in a form and within a time frame
that enables them to carry out their
internal control and other
responsibilities.

Program review results BOP documents and shares internal review results with the
organization that is reviewed, which are generally institutions.
Quarterly program
review summaries
Quarterly program review summaries compile numerous internal
review results and highlight frequent internal control deficiencies and
other significant findings. BOP distributes these summaries bureauwide. For example, if a significant finding related to an inventory
process is identified at one institution, the summary would cite the
issue anonymously for other institutions’ awareness.
As noted above, BOP also has two efforts—the Cost Efficiencies and
Innovations Catalog and the best practices intranet site—to
communicate efficiency and cost-saving efforts.

BOP’s use in identifying opportunities for cost efficiencies

Monitoring
Internal control monitoring should
assess the quality of performance
over time and ensure that the
findings of audits and other
reviews are promptly resolved.

Corrective action plan

Within BOP, the subject of an internal review is responsible for
devising and completing a corrective action plan to resolve any
internal control deficiency or other finding. For example, if an
institution did not conduct a quarterly inventory of property in a certain
program area, the institution is responsible for mitigating the risk of
inaccurate agency records.
BOP’s Program Review Division monitors whether corrective action
has been taken by the specific organization or program that is
reviewed. A follow-up internal review eventually verifies whether the
corrective action is adequate. The frequency with which a subsequent
review occurs depends on the likelihood of risk as determined during
the Management Assessment process. For example, BOP conducts
internal reviews of certain institutions’ operations, at a frequency
ranging from once a year to once every 3 years.

Source: GAO analysis of BOP information. | GAO-14-821
a

Office of Management and Budget, Circular A-123, Management’s Responsibility for Internal Control,
defines management’s responsibility for internal control in federal agencies. The circular states that
federal employees must ensure that resources are used efficiently and effectively to achieve desired
objectives. Internal controls are the tools (organization, policies, and procedures) to help managers
achieve results, including operational efficiency.

Page 23

GAO-14-821 BOP Cost Efficiencies

BOP’s Internal Control
System Helps Identify
Opportunities to Achieve
Efficiencies, but Lacks a
Mechanism to
Consistently Monitor
Bureau-wide Corrective
Actions

We found that BOP’s internal control system is designed to identify
additional opportunities for cost efficiencies, but BOP did not provide
specifics or documentation on a mechanism to consistently monitor
bureau-wide corrective actions. Program reviews—which document the
results of the Program Review Division’s (PRD) internal reviews of
programs and operations at each BOP institution—collectively cited
repeated frequent control deficiencies and significant findings over
multiple time periods that present inefficiencies or could increase costs. 32
However, for these repeated issues, BOP did not provide specifics or
documentation for how its corrective actions were always monitored
consistently to ensure that the actions addressed the issues bureauwide. 33 Specifically, we reviewed BOP’s quarterly program review
summaries—which compile each quarter’s internal review results and
highlight frequent deficiencies and significant findings—for the years
2009-2013 to identify significant findings as well as the most frequent
internal control deficiencies reported with respect to opportunities to
achieve cost efficiencies in the highest-cost areas identified above.
According to our review of these 20 quarterly summaries, frequent
deficiencies or significant findings were repeatedly reported in three highcost areas—time and attendance (14 of 20 quarters), food services (8 of
20 quarters), and RRC contract administration (6 of 20 quarters).
BOP has processes to identify and share information about these
frequent and significant deficiencies. In particular, PRD disseminates its
quarterly program review summaries to all BOP institutions and the other
Central Office divisions to make them aware of these deficiencies. For
instance, we found that officials from the Central Office Divisions with
oversight over the repeated frequent deficiencies and significant findings
32

BOP/PRD also reports frequent deficiencies and significant findings associated with
operational effectiveness and compliance-related issues, such as correctional services
and food safety. We determined these reported issues do not necessarily have a direct
impact on operational efficiency and therefore were not included in the scope of our
review.

33
A deficiency generally refers to an internal control deficiency (e.g., an internal control is
missing, needs improvement, or is not operated as designed). BOP officials stated that a
deficiency is considered “frequent” if it arises multiple times, within a quarterly time period,
in program reviews conducted at multiple institutions. A finding is an issue identified during
the course of a review. BOP officials stated that findings are considered significant if the
reported issue is associated with a high-cost program activity or an invaluable subject
area such as quality of care. A deficiency generally refers to the ineffectiveness of an
internal control activity. If the same finding or deficiency is reported across multiple
reporting periods, it is considered a repeat problem.

Page 24

GAO-14-821 BOP Cost Efficiencies

that we analyzed—the Human Resource Management Division (HRMD),
Health Services Division (HSD), and Reentry Services Division (RSD)—
were aware of them.
Moreover, one of the divisions, HRMD, formulated a corrective action
plan and has systematically taken steps to address and monitor progress
toward resolving reported deficiencies and findings related to time and
attendance that were identified at multiple institutions across multiple time
periods. For instance, BOP’s program reviews identified control issues
related to time and attendance in 14 of the 20 quarters from 2009 through
2013, such as inaccurate time and attendance records and the lack of
appropriate documentation to support payment. These reported issues
could lead to payment errors, such as incorrectly compensating staff for
overtime. HRMD officials reported that, as a result of these issues, they
implemented policy changes, such as developing mandatory training and
requiring that each human resource office audit a representative sample
of its time and attendance files annually. Further, BOP is currently
transitioning to a web-based time and attendance reporting system, which
includes information technology controls. Such information technology
controls establish segregation of duties (e.g., the staff approving reported
work hours is supervisory and different from the staff reporting the hours
worked) and the retention of electronic records in compliance with rules
and regulations, which officials report could help address the deficiencies.
To consistently monitor progress toward resolving these issues bureauwide, BOP’s Central Office has conducted a number of audits of time and
attendance records and reports that error rates in record accuracy have
declined from about 5 percent in 2010 to less than 2 percent in 2014.
The other two Central Office divisions with oversight over the repeated
frequent deficiencies and significant findings that we analyzed—HSD and
RSD—have taken actions at the national level to attempt to address
some of the deficiencies. However, they did not provide specifics or
documentation for how the divisions would monitor progress toward
correcting them. With respect to HSD, BOP’s internal reviews identified
findings related to the administration and management of resources in its
Food Services Program (constituting $408 million, or 6 percent, of BOP’s
obligations in fiscal year 2013) in 8 of the 20 quarters from 2009 through
2013, such as incomplete inventories; not properly reporting lost,
damaged, or stolen property; or failures to adjust inventory records
referenced when formulating and submitting an institution’s budget. At the
national level, BOP is providing training on the use of the Food Services
Program’s information system that includes a quarterly inventory control
function, and has formed support teams to provide assistance to

Page 25

GAO-14-821 BOP Cost Efficiencies

institutions identified as having deficient programs. For the staff
performing inventory control at the local level (e.g., warehouse staff), the
division is requesting that they attend national training or a similar training
program. However, HSD did not provide specifics or documentation for
establishing a mechanism that would allow it to monitor whether the
actions taken have produced needed improvements within Food Services
comprehensively across institutions.
With respect to RSD, BOP’s internal review identified findings related to
RRC contract oversight and monitoring (RRC contracts constituting $296
million, or 4 percent, of BOP’s obligations in fiscal year 2013) in 6 of 20
quarters from 2009 through 2013, such as monitoring reports not being
completed on a timely basis and inaccurate payments being made to
contractors. These reported issues could result in failing to address
contractor performance or overpayments. In July 2013, Residential
Reentry Management (RRM) offices nationwide were consolidated under
the direct management of the Residential Reentry Management Branch
(RRMB). Previously, the RRM field offices were under the operational
control of the six BOP regional offices. According to BOP, this
reorganization has allowed BOP to implement and monitor consistent
practices throughout the Residential Reentry offices. BOP also reported
that, as a result of these findings, management teams within the
Residential Reentry Management Branch provide guidance to individual
field offices that require corrective actions. Additionally, they have
scheduled a national-level training session in September 2014 to address
repeated frequent deficiencies, share best practices in contract
monitoring, and promote consistency and effective oversight of RRCs.
Similar to Food Services, however, RSD did not provide specifics or
documentation for establishing a mechanism that would allow it to monitor
whether the actions taken have produced needed improvements with the
Residential Reentry Program comprehensively across RRC contracts.
Internal control standards state that monitoring of internal controls should
include policies and procedures for ensuring that the findings of audits
and other reviews are promptly resolved. In particular, managers are to
promptly evaluate findings from audits and other reviews, including those
showing deficiencies and recommendations reported by auditors who
evaluate agencies’ operations; determine proper actions in response to
findings and recommendations from audits and reviews; and complete,
within established time frames, all actions that correct or otherwise
resolve the matters brought to management’s attention. According to the
standards, the resolution is completed only after action that corrects

Page 26

GAO-14-821 BOP Cost Efficiencies

identified deficiencies, produces improvements, or demonstrates the
findings and recommendations do not warrant management action. 34
For deficiencies and other findings associated with a particular institution,
BOP has a mechanism to monitor corrective actions at the individual
institution to help ensure that the deficiencies and findings do not reoccur
at that specific location. Specifically, according to BOP, the responsible
official of the entity reviewed (most often the warden of the institution) is
responsible for developing and implementing the corrective action, and
PRD monitors that specific institution’s progress in implementing the
corrective action. However, with respect to frequent deficiencies and
significant findings across a number of institutions repeatedly reported
over time, BOP did not provide specifics or documentation for
establishing a mechanism to consistently monitor bureau-wide corrective
actions to determine whether they are achieving desired results and
helping to prevent similar deficiencies from occurring at different
institutions over time. We found that in at least one instance in which
deficiencies were repeatedly found across a number of institutions—
HRMD’s process for resolving deficiencies associated with time and
attendance—a Central Office division did successfully take steps to
consistently monitor the deficiencies bureau-wide. Specifically, for the
repeated frequent deficiencies and significant findings associated with
time and attendance, we found that HRMD’s process for implementing
and monitoring its corrective actions has resulted in fewer errors and
stronger internal controls related to record retention and segregation of
duties, which could reduce the risk of erroneous payments, thereby
providing BOP with greater assurance that it is operating more efficiently
with regard to its staff costs across its many institutions.
However, despite the repeated nature and recency of findings within the
Food Services and Residential Reentry Programs, the relevant Central
Office divisions did not provide specifics or documentation for how they
have monitored their corrective actions to assess their progress toward
addressing these findings across BOP institutions. As a result, BOP does
not have reasonable assurance that actions taken are preventing certain
frequent and significant findings from reoccurring at other institutions,
thus hindering BOP’s ability to achieve cost efficiencies in these two highcost areas—accounting for spending of $408 million and $296.1 million in

34

GAO/AIMD-00-21.3.1.

Page 27

GAO-14-821 BOP Cost Efficiencies

fiscal year 2013, respectively. By establishing a mechanism for relevant
Central Office divisions to consistently monitor bureau-wide corrective
actions and assess their progress in the presence of repeated frequent
deficiencies or significant findings reported across multiple institutions,
BOP could be better assured that it is resolving such deficiencies
promptly and, ultimately, operating more efficiently. Further, although
repeated frequent deficiencies and significant findings related to
operational effectiveness and compliance-related issues were outside the
scope of our review, such a mechanism would also better position BOP to
resolve repeated deficiencies and significant findings in these areas.

Authorities within
BOP’s Control to
Reduce Costs Are
Limited; Options
outside of BOP’s
Authority Could Have
a Larger Impact on
Costs
BOP’s Current Authority to
Reduce Inmate Sentences
Results in Limited Cost
Savings Relative to Its
Overall Budget

BOP has certain discretionary authorities that can affect the period for
which an inmate is incarcerated or remains in BOP custody. We have
previously reported on these in detail, and found that inmate eligibility for
certain programs and lack of capacity affect BOP’s use of them. 35 Among
these discretionary authorities are those that permit BOP to
determine the eligibility of a foreign national offender for transfer to the
offender’s home country to serve out his or her sentence; 36

•

35

GAO-12-320.

36

See 18 U.S.C. § 4102.

Page 28

GAO-14-821 BOP Cost Efficiencies

•

reduce the length of the inmate’s sentence, based on inmate “good
conduct time (GCT);” 37

•

determine the eligibility of an offender for participation in a substance
abuse program (which may result in a sentence reduction of up to 1
year); 38 and

•

transfer an inmate out of prison to serve the remainder of his or her
sentence in a residential reentry center or in home detention. 39

Table 3 describes key statutory provisions in which BOP has a role in
reducing a federal prisoner’s period of incarceration in a BOP institution.
Table 3: Key Statutory Provisions Available to the Bureau of Prisons (BOP) to Reduce a Federal Prisoner’s Period of
Incarceration or Time in BOP Custody
Discretionary flexibilities

Description

International Prisoner Transfer Program
(IPTP), 18 U.S.C. § 4102.

The Department of Justice (DOJ) is authorized to transfer offenders under a sentence
of imprisonment, on parole, or on probation to the foreign countries of which they are
citizens or nationals, provided that there is a current treaty between the United States
and the other country that provides for such transfer, that the offender has voluntarily
requested the treaty transfer, and that the home country agrees to the transfer. BOP
determines whether an inmate meets specific minimum eligibility requirements; if these
are met, BOP forwards the request to DOJ’s Office of Enforcement Operations. DOJ is
responsible for determining whether an inmate is approved or denied for treaty transfer.

37
See GAO-12-320 for additional details on GCT. As authorized in statute, 18 U.S.C. §
3624(b), BOP may award “up to 54 days at the end of each year of the prisoner’s term of
imprisonment,” or 54 days per year of sentence served. As applied by BOP, however, this
actually results in 47 days earned per year of sentence imposed because inmates do not
earn GCT credit for years they do not ultimately serve because of being released early.
Calculation of the amount of time that BOP should award based on good conduct was
contested in court. In Barber v. Thomas, 130 S. Ct. 2499 (2010), the Supreme Court
upheld BOP’s methodology against a challenge brought by inmate petitioners. As a result,
BOP officials state that any change in the method of calculating GCT credits would require
a change to the law. Therefore, for our purposes in this report, GCT is both a discretionary
authority and one that is an option that could produce potential costs savings, provided the
calculation method is changed. We include GCT in the list of discretionary authorities and
as one of the sentencing guideline change options.
38

18 U.S.C. § 3621(e).

39

18 U.S.C. § 3624(c).

Page 29

GAO-14-821 BOP Cost Efficiencies

Discretionary flexibilities

Description

Reduction in Sentence through
Compassionate Release, 18 U.S.C. §
4205(g), 18 U.S.C. § 3582(c)(1)(A).

Upon motion of the Director of BOP, the court may reduce a term of imprisonment after
considering certain factors if it finds that either (1) extraordinary and compelling reasons
warrant such a reduction; or (2) the inmate is at least 70 years of age, has served at
least 30 years in prison for the offense or offenses for which the inmate is imprisoned,
and a determination has been made by the Director of BOP that the inmate is not a
danger to the safety of any person or the community; and that such a reduction is
consistent with applicable policy statements issued by the U.S. Sentencing Commission
(USSC).

Good conduct time (GCT)
18 U.S.C. § 3624(b).

BOP is authorized to award credit toward the service of an inmate’s sentence, beyond
the time served, of up to 54 days per year of sentence served if the inmate has
displayed exemplary compliance with institutional disciplinary regulations. To be
eligible, the inmate must be serving a sentence of more than 1 year other than a term of
imprisonment for life.

Residential Drug Abuse Treatment Program BOP is required to provide substance abuse treatment for each inmate it determines
(RDAP), 18 U.S.C. § 3621(e).
has a treatable condition of substance abuse. BOP must, subject to the availability of
appropriations, provide residential substance abuse treatment (and make arrangements
for appropriate aftercare) for all eligible inmates, with priority provided based on
proximity to release date. Through RDAP, BOP may reduce the sentence of an inmate
convicted of a nonviolent offense who successfully completes the treatment program for
a period of up to 1 year.
Residential Reentry Program
18 U.S.C. § 3624(c).

The Director of BOP must, to the extent practicable, ensure that an inmate spends a
portion of the final months of that inmate’s term (not to exceed 12 months), under
conditions that will afford the inmate a reasonable opportunity to adjust to and prepare
for reentry into the community. This may include a prisoner being placed in a residential
a
reentry center (RRC, or halfway house). As of April 2014, BOP supervised the
operations of 200 RRCs nationwide; RRCs are operated by private contractors that
b
have contracts with BOP.

Source: GAO analysis of federal statutes and BOP information. See GAO-14-821.
a

Home detention is an option for a limited number of offenders, and BOP requires that specific criteria
be met, both with regard to the offender’s eligibility and with having an acceptable home environment
for home detention. By statute, an eligible offender may be placed in home detention for the shorter of
10 percent of the term of imprisonment or 6 months.

b
BOP reported that it also has 49 intergovernmental agreements for “work release,” which it includes
under its count of RRCs (for a total of 249); of these agreements, 40 are with state and local
governments, and 9 are with the U.S. Marshals Service.

Of these authorities, two—IPTP and Compassionate Release—are
programs that can result in a direct reduction of BOP’s population and
thus cost savings. The number of inmates released through these two
programs annually is small, however, as are the estimated savings. For
example, the estimated 2013 cost savings from these programs were
about $2.6 million and $651,000, respectively, or about 0.04 and about
0.01 percent of BOP’s total fiscal year 2013 appropriations of $6.44
billion.

Page 30

GAO-14-821 BOP Cost Efficiencies

•

IPTP. As of June 28, 2014, BOP had about 54,000 foreign nationals in
its custody. However, eligibility criteria constrain the number who can
be transferred. Specifically, inmates must voluntarily request a
transfer to their home countries to serve out the remainder of their
sentences, the United States must have a treaty with the home
country permitting repatriation, eligibility for transfer must be
confirmed by BOP, the transfer request must be approved by DOJ’s
Office of Enforcement Operations, and the home country must agree
to receive the prisoner. Further, while the United States has treaties
with 94 foreign countries to permit IPTP transfers, the DOJ Inspector
General (IG) reported that, as of 2010, about 22 percent of the foreign
nationals in BOP come from countries with which there is no transfer
agreement. In addition, as of late June 2014, almost 37,400 (17.3
percent) of all BOP inmates were Mexican nationals (or about 70
percent of all BOP foreign nationals). However, since 2001, Mexico
has restricted eligibility for transfer from the United States, according
to a 2011 DOJ IG report. 40

In fiscal year 2013, DOJ approved 245 transfers out of BOP custody,
representing less than 0.5 percent of BOP’s foreign national population.
Through these transfers, BOP estimates that it avoided costs of
approximately $2.6 million in fiscal year 2013, or about 0.04 percent of
BOP’s fiscal year 2013 $6.44 billion appropriation. 41
•

Compassionate Release. We have previously reported on BOP’s
discretionary authority to request compassionate release for certain
offenders—termed “reduction in sentence” by BOP—and found that

40

DOJ, Office of the Inspector General: The Department of Justice’s International Prisoner
Transfer Program, I-2012-002 (Washington, D.C.: December 2011). The DOJ IG reported
that, in addition to eligibility criteria constraints on using the program, other factors within
BOP control also affected its use, such as incorrect determinations of inmate eligibility,
language barriers hindering inmate understanding of the program, or delays in processing
applications. The IG estimated that, for fiscal years 2005-2010, BOP delays in processing
applications resulted in about $8 million in additional costs, or $1.33 million annually, on
average, for each of those 6 years. The IG report stated that since 2001, Mexico’s
“significant restrictions” had resulted in few Mexican inmates being accepted for treaty
transfer consideration. Data on BOP inmates by nationality are from BOP’s SENTRY data
system. The IG made several recommendations for BOP to address these findings,
including ensuring that all staff involved in IPTP determinations are properly trained. BOP
concurred with the IG’s recommendations and stated that actions are under way to
address them.

41

Estimated cost avoidance based on a fiscal year 2013 marginal cost of $10,674 per
offender.

Page 31

GAO-14-821 BOP Cost Efficiencies

BOP used this authority infrequently because of, among other things,
BOP’s eligibility criteria. 42 Specifically, compassionate release of an
inmate historically has been limited by BOP to terminally ill patients
with an incurable disease and short life expectancy or to those who
are elderly with a profoundly debilitating medical condition. In April
2013, the DOJ IG reported that the existing BOP compassionate
release program had been poorly managed and implemented
inconsistently, likely resulting in eligible inmates not being considered
for release and in terminally ill inmates dying before their requests
were decided. 43 No cost savings estimate could be provided for earlier
releases because, according to the IG report, BOP did not maintain
cost data associated with the custody of inmates eligible for
consideration under the program, and BOP had not conducted any
analysis of cost savings achieved by releasing such inmates.
DOJ issued a revised program statement with less restrictive criteria for
compassionate release in August 2013. The revisions included making
terminally ill inmates with a life expectancy of up to 18 months eligible for
consideration instead of maintaining the previous 12-month limit. 44
According to BOP data, total releases under compassionate release
increased from 29 inmates in calendar year 2011 to 39 in calendar year
2012, and to 61 in calendar year 2013 (including 11 released under the
expanded criteria). An additional 2 had been released as of July 2014
under expanded nonmedical criteria. However, as a percentage of BOP’s
population and costs, the number of affected inmates and savings
remains small given the numerous factors involved in determining

42

GAO-12-320.

43
DOJ, Office of the Inspector General: The Federal Bureau of Prisons’ Compassionate
Release Program, I-2013-006 (Washington, D.C: April 2013). The IG made several
recommendations for BOP to address these findings, including establishing time frames
for processing requests at each step of the management review process. BOP generally
concurred with the IG’s recommendations and stated that actions are under way to
address them.
44

According to the revised program statement for Compassionate Release/Reduction in
Sentence, the following were added: criteria regarding requests based on medical
circumstances, nonmedical circumstances for elderly inmates, nonmedical circumstances
in which there has been the death or incapacitation of the family member caregiver of an
inmate’s child, nonmedical circumstances in which the spouse or registered partner of an
inmate has become incapacitated, and a list of factors that should be considered for all
requests.

Page 32

GAO-14-821 BOP Cost Efficiencies

eligibility. Inmates released under compassionate release in calendar
year 2013 represented about 0.03 percent of the total BOP inmate
population, and estimated savings from the releases amounted to about
$651,000, or about 0.01 percent of BOP’s fiscal year 2013 $6.44 billion
appropriation. 45
Funding and eligibility constraints also limit BOP’s use of RDAP and
RRCs to affect costs related to the number of inmates in BOP custody.
Some cost savings can result from inmates being able to reduce their
sentences as a result of successfully completing the RDAP program;
however, both RDAP and RRCs are types of rehabilitation programs
intended to reduce recidivism rather than serving as cost-saving
mechanisms. 46 Moreover, as RDAP and RRC are programs that require
professional staff and resources, increasing the number of offender
participants in either would require additional funding that could offset
potential cost savings. 47
•

RDAP. We have previously reported that, according to BOP, delays
resulting from system-wide demand for this treatment program have
prevented timely inmate entry into RDAP and thereby reduced the
number of eligible inmates receiving the maximum allowable sentence
reduction. This is because by the time some inmates complete RDAP,
they have fewer months remaining on their sentences than the
maximum allowable reduction of 12 months. 48 According to BOP, the

45
Data were provided by BOP for calendar year releases. Cost avoidance savings were
calculated by multiplying the BOP fiscal year 2013 marginal cost per offender ($10,674) by
the number of releases. Costs as a percentage of BOP appropriations were calculated by
dividing the cost savings by the cited BOP appropriations.
46

RDAP participation is limited to prisoners with a verifiable substance use disorder within
the 12 months prior to their arrest. They must also voluntarily apply for participation and
be approved by BOP for the program. RRCs are run by private contractors and cost more
than minimum security institutions, and about the same as low- or medium-security BOP
prisons on an annual basis per offender. We reported daily costs per prisoner previously
in GAO-12-320 as follows: $70.79 for RRCs, $57.55 for minimum security, $69.53 for low
security, and $71.91 for medium security.

47

We did not include in cost savings the potential long-term savings that could result from
reduced recidivism.

48

GAO-12-320. For fiscal years 2009-2011, 15,302 RDAP participants completed the
program and were eligible to receive a sentence reduction. However, of these participants,
19 percent received the maximum sentence reduction that corresponded to their sentence
length.

Page 33

GAO-14-821 BOP Cost Efficiencies

average participant reduced his or her sentence by about 10 months
in fiscal year 2013.
As we have also previously concluded, increasing RDAP participation
would cost more than the potential savings from reducing sentences by
relatively small amounts. For example, if every eligible RDAP participant
who completed the program in fiscal year 2011 had received his or her
maximum sentence reduction, BOP would have been responsible for
15,729 fewer months of inmate incarceration, yielding an estimated cost
savings of about $13.2 million. However, the fiscal year 2012 request for
an additional $15 million to permit more offenders to enter RDAP “on
time” and thereby to increase their allowable sentence reductions would
not have been enough to achieve the full 12-month potential reduction,
meaning that expanding RDAP would have cost at least $1.8 million more
than it would have saved. Similarly, according to DOJ, BOP received an
increase of $15 million in RDAP funding in fiscal year 2014 and this
funding will, by fiscal year 2016, increase the average sentence reduction
for RDAP participants from 10 months in fiscal year 2013 to an estimated
11.5 months. 49 BOP reported that reaching the full potential 12-month
sentence reduction for all participants in fiscal year 2013 would have
resulted in about $11 million in cost avoidance savings, but it did not
provide a calculation of the cost of providing additional spaces and staff to
permit the full sentence reduction for all participants.
These cost savings estimates do not take into account potential long-term
savings to society from reducing substance abuse and recidivism among
released inmates. While RDAP may contribute to achieving such savings,
expanding it above current levels requires an up-front cost that exceeds
its potential cost savings for BOP in the short term.
•

RRCs. According to BOP, RRCs, or halfway houses, provide
programs that are intended to help inmates rebuild their ties to the
community and to thereby reduce the likelihood that they will
recidivate. Run by private contractors, RRCs provide programs that
help offenders find employment, and coordinate the provision of
services to address substance abuse, medical, and mental health
care with appropriate community providers. We have previously found

49

BOP did not provide an estimate of the savings from shortening the average sentence
by 1.5 months. BOP stated that this funding level would mean that RDAP was operating at
100 percent of capacity; that is, all the available spaces would be taken.

Page 34

GAO-14-821 BOP Cost Efficiencies

that the use of RRCs is constrained by both eligibility criteria and
funding. 50 For example, inmates with detainers (other crimes for which
they were charged but were not yet sentenced), with sentences of 6
months or less, or who are in civil commitment status are all ineligible
for RRC placement. 51 In addition, BOP officials stated that objections
from local communities to halfway houses in residential
neighborhoods have hindered expanding the RRC program, even
when funds were available.
We have previously found that not all inmates are eligible to be sent to an
RRC prior to their release from prison, and that for those who are eligible,
some spend only a portion of the full 12 months’ allowable time in an
RRC because of a lack of bed space and because of eligibility criteria. 52
According to BOP, an increase in the number of offenders getting the full
12 months’ allowable time would necessitate additional bed space, which
would require both additional funding and additional RRC contracts. For
example, in fiscal year 2013, BOP reported that it had 9,455 RRC beds
available nationwide, but would have required about 30,000 beds to
provide the maximum allowable 12 months in RRCs to all participants, or
an addition of more than 20,500 beds above its current capacity. As noted
above, and as we have previously reported, BOP officials explained that
such an expansion could be challenged by local zoning restrictions and
the unwillingness of many communities to accept nearby RRCs.
Moreover, if such an expansion were to take place, BOP would need
additional funding to pay for the new RRC bed space. For fiscal year
2013, BOP reports that the average daily cost per offender in an RRC
was about $73, or $26,645 per year. This means that an increase of the
more than 20,500 beds that would be required to achieve the allowable
12 months for all participants (at current program levels) would cost about

50

GAO-12-320. During an assessment for RRC placement, BOP policy requires
prerelease RRC placement decisions be made on an individual basis and conducted in a
manner consistent with certain statutory criteria. The criteria include (1) the resources of
the facility contemplated, (2) the nature and circumstances of the offense, (3) the history
and characteristics of the prisoner, (4) any statement by the court that imposed the
sentence, and (5) any pertinent policy statement issued by the USSC. See 18 U.S.C. §
3621(b).
51

A detainer is a document issued by a law enforcement entity, a jail, or correctional
facility to seek custody of an individual for purposes of instituting legal proceedings.

52

GAO-12-320.

Page 35

GAO-14-821 BOP Cost Efficiencies

$546 million annually. 53 Expanding RRCs might help reduce recidivism
but would require a substantial funding increase, equal to almost 8
percent of BOP’s entire $6.9 billion fiscal year 2015 budget request.

Potential Actions outside
of BOP’s Authority Could
Reduce Prison
Populations and Costs,
and Experts Say There
Are Advantages and
Disadvantages to These
Options

We examined cost-saving options that are outside BOP’s authority and
that are widely discussed among criminal justice policy experts, the
USSC, and DOJ, such as instructing federal prosecutors to decline to
charge certain defendants in certain drug cases in such a manner as to
trigger mandatory minimum prison sentences, reducing the lengths of
sentences specified by federal guidelines for drug offenders, or applying
the provisions of the Fair Sentencing Act of 2010 retroactively in order to
provide sentences for cocaine base (“crack”) offenders that are less
disparate from those applied for possession of powder cocaine. 54 Toward
this end, we reviewed selected reports and analyses from entities chosen
for their expertise in criminal justice issues, and had discussions with
officials from DOJ, BOP, and the USSC. See appendix VI for more
information on the entities we selected. Through this review, we identified
eight options outside BOP’s authority that could reduce the size of the
inmate population. These involve potential legislative actions by
Congress, executive orders by the President, policy changes
implemented by the Attorney General, or changes in sentencing

53

This cost estimate is based on BOP RRC fiscal year 2013 average RRC cost of $26,645
per offender multiplied by the required 20,500 bed spaces.

54

Pub. L. No. 111-220, 124 Stat. 2372. The Anti-Drug Abuse Act of 1986, Pub. L. No. 99570, 100 Stat. 3207, included the federal criminal law distinction between “cocaine base”
and other forms of cocaine. The thresholds triggering a 10-year prison penalty—5
kilograms of powder cocaine and 50 grams of cocaine base—created a 100-to-1 quantity
ratio.

Page 36

GAO-14-821 BOP Cost Efficiencies

guidelines that could be promulgated by the USSC. 55 For each of these
options, we asked the USSC to calculate the estimated impact on the size
of BOP’s population and the potential savings that could result from
implementation of each. 56 See appendix VII for the USSC analyses of
these options. For the purposes of our analysis, we determined that the
USSC is an authoritative source for conducting such analysis; it has
conducted similar analyses in recent years and published one on its
website on July 25, 2014. 57 For each option, the time frame (number of
years) over which the option would take effect may vary, as well as the
number and type of offenders who would be eligible for a changed
sentence under the option. In general, the majority of the impact of most
options was estimated by the USSC to occur within 5 or 6 years of
implementation. However, the options are not all directly comparable;
some would have a greater impact more quickly than would others,
depending on the option and the number of potentially eligible offenders.
For options that apply only prospectively (that is, only to future
offenders)—such as the Attorney General’s August 2013 memorandum to
U.S. Attorneys—the potential impact is typically less than for options that
would apply to larger numbers of currently incarcerated offenders.

55

Of these eight options, two related to the sentencing guidelines were approved by the
USSC in April and July 2014; a third was directed by the Attorney General in August 2013.
All three changes occurred during the course of this review. The two actions approved by
the USSC would reduce lengths of prison terms for (1) most future federal drug offenders
by amending the sentencing guidelines, and (2) currently incarcerated federal drug
offenders by about 19 percent compared with current term lengths by amending a policy
statement to give retroactive effect to the proposed amendment to the sentencing
guidelines. However, under 28 U.S.C. § 994(p), Congress may modify or disapprove
amendments to the sentencing guidelines. (The second action corresponds to the first and
does not require separate congressional action.) The USSC retroactivity change would
permit eligible offenders to ask courts to reduce their sentences beginning November 1,
2014. However, offenders whose requests are granted by the courts would be released no
earlier than November 1, 2015. According to the USSC, sentences will be reduced
retroactively only after consideration by a judge in each case. The third policy change
went into effect in August 2013, when the Attorney General issued a memorandum to U.S.
Attorneys informing prosecutors to decline to charge certain defendants in certain types of
drug cases in such a manner as to trigger mandatory minimum sentences.
56

The USSC calculated the potential savings in bed years; 1 bed year is defined by the
USSC as 12 months’ prison time (e.g., one offender in prison for 1 year, or three offenders
in prison for 4 months each, and so forth.)

57

USSC: Analysis of the Impact of the 2014 Drug Guidelines Amendment If Made
Retroactive (Washington, D.C.: July 25, 2014). See this citation at http://www.ussc.gov.

Page 37

GAO-14-821 BOP Cost Efficiencies

On the basis of analyses the USSC conducted of the potential results
from implementing each of the eight options, which provided estimates of
bed year savings, we then calculated the potential cost savings. 58
Implementation of the options could range from about $8.7 million to
almost $4.1 billion, with the potential savings varying based on the option
implemented and the duration of the option. Cost savings or cost
avoidance savings were calculated using BOP’s fiscal year 2014
estimated average annual marginal cost of approximately $11,000.
Marginal costs covers such things as security, food, medical care,
clothing, unit management, education, records, and maintenance
associated with additional inmates entering existing BOP facilities. BOP
officials stated that calculating cost savings using marginal cost was more
appropriate than using per capita costs because BOP is about 36 percent
overcrowded system-wide, and therefore BOP would have to reduce its
existing population by that percentage (equal to about 78,000 offenders in
July 2014) to eliminate overcrowding system-wide. Until that number is
reached, BOP officials stated, they would not be closing existing prisons
and therefore would expect savings only in marginal costs versus
average per capita costs to be realized. However, the use of marginal
cost in calculating savings does not take into account the possibility that
reductions in BOP population could permit BOP to reduce the numbers of
offenders sufficiently to reduce staff, close some existing facilities, or
avoid opening additional ones that would otherwise have been required. If
the options achieved population reductions resulting in reductions in fixed
costs, such as reducing staff or closing existing prisons, calculating cost
savings using average annual per capita costs—$29,291 in fiscal year
2013, according to BOP—would be appropriate. Three of the options
could result in an estimated 78,000 or more offenders being eligible for
release or eligible for sentence reductions over five years, which could
greatly reduce or eliminate current overcrowding based on BOP’s
estimate within the same 5 years. 59 Given the size of these potential

58

We calculated cost savings that would result from reductions in BOP population or from
cost avoidance savings resulting from increases below what would otherwise occur if an
option was not implemented, by multiplying USSC estimates of bed years saved by the
annual marginal cost for a prisoner in BOP.

59

DOJ reports that as of February 2014 BOP prisons were 32 percent overcrowded but
medium- and high-security prisons were more overcrowded than low- and minimumsecurity institutions. Reducing overcrowding and costs would depend on which offenders
were released, from where, and in what number. Note that a reduction in bed years does
not necessarily translate into an immediate reduction in prison population.

Page 38

GAO-14-821 BOP Cost Efficiencies

reductions in sentence lengths and the number of eligible offenders, we
calculated per capita cost savings for three of these options, based on the
USSC estimates for saved bed years. We did not calculate per capita
savings for a fourth option—changing the way that BOP calculates the
amount of “good conduct” time that offenders earn for each year served—
because although the number of eligible inmates would be almost
180,000, the reduction in sentence for each inmate would be small in
contrast to the other three options and would occur over 30 or more
years. Therefore, this option may be less likely to allow BOP to reduce
staffing levels or close facilities. 60
It is important to emphasize cost savings is one of many considerations.
In particular, these options may have additional effects, such as on public
safety. The expert entities we contacted identified potential advantages
and disadvantages they believe could result from the implementation of
each option, such as making sentences more commensurate with the
crime or releasing inmates who may reoffend.
We are not taking a position on any of these options. We present the
information we obtained from the USSC and the expert entities to inform
policymakers as they weigh whether and how to address costs at BOP.
Specifically, in table 4, we summarize each option we presented to the
USSC, expert entities, and the four states selected because of their
experience with criminal justice sentencing initiatives. For each option, we
show the USSC estimate of the number of BOP offenders who would
likely be eligible for release or for a reduced sentence (leading to release)
as a result of implementation. The table also shows the USSC estimates,
if available, of the bed year savings that would result from implementation
of each option, and our calculation of the cost savings that would result. 61
In addition, we summarize the comments from the responding states and
the expert entities with regard to potential advantages and disadvantages
that could result from implementing these options.

60

See option 6.

61

We asked the USSC to estimate the number of offenders who would be eligible to be
released or to have their sentence reduced (leading eventually to release), and to
estimate the savings that would result from implementation of these eight options. Where
estimates could be made by USSC, the savings were provided in bed years.

Page 39

GAO-14-821 BOP Cost Efficiencies

Table 4: Options for Changes to Sentencing Processes, Number of Eligible Inmates, Estimated Cost of Options, and
Assessment of Potential Advantages and Disadvantages by Expert Entities and Four States with Sentencing Change
Experience
Option 1: Reduce the length of mandatory minimum sentences for incarcerated drug offenders

a

Description

Implementing this option would reduce the sentences of certain current
incarcerated drug offenders convicted of an offense carrying a
mandatory minimum sentence—for example, from 20 years to 10
years, 10 years to 5 years, and 5 years to 2 years—and would release
incarcerated offenders if they had already completed their recalculated
sentence (i.e., if the time served exceeded the new sentence length).
Offenders with time remaining on their sentences would remain in
prison until they met the new (shorter) sentence length requirement that
applies to their cases. Implementing this option would mean an average
reduction in sentence for eligible offenders of 57 months (44.2 percent).

Estimated number of inmates eligible for release or
sentence reduction

This is estimated to be 78,102 inmates. Of these, 68,620 (88 percent)
would be eligible for release during the first 5 years after
implementation, and 9,482 in the sixth year and after.

Estimated bed year reduction and cost savings

It is estimated that 370,985 bed years would be saved, or $4.1 billion.
Savings would roughly match the released numbers—about 88 percent
($3.6 billion) in savings would occur in the first 5 years after
implementation.
Given the size of these potential reductions in offenders and required
bed space, per capita savings could occur, which means savings could
be as much as $10.231 billion over the first 5 years and an additional
c
$1.395 billion in the sixth year after, or a total of $11.626 billion.

Expert-identified potential advantages

•

b

•

Expert-identified potential disadvantages

•
•
•

This could make sentences more proportionate to the crime
committed.
This could significantly reduce Bureau of Prisons (BOP)
population, thereby reducing overcrowding and improving prison
safety for prisoners and staff, as well as reducing wait lists for antirecidivism programs.
Released offenders may recidivate, which could increase crime.
The financial and societal costs of increased crime could eclipse
cost savings.
This could reduce prosecutorial leverage to get plea bargains in
exchange for evidence against those in higher positions in a
criminal organization.
This could shorten sentences too much below original intent or
shorter sentences may be too lenient.

a
The Anti-Drug Abuse Act of 1986, Pub. L. No. 99-570, 100 Stat. 3207, established the basic
framework of statutory penalties currently applicable to federal drug-trafficking offenses. The act
specified statutory penalty ranges for manufacturing, distributing, or dispensing the drug, or
possessing the drug with intent to manufacture, distribute, or dispense the drug. For each of several
drug types, the act specified separate statutory ranges for such offenses involving various quantities
of the drug. See 21 U.S.C. § 841.
b

In developing its analysis, the USSC made a number of assumptions, including that Congress would
authorize the courts (or the President using clemency powers) to reduce the sentences of offenders
incarcerated on or after a specific date in a manner consistent with the Guidelines Manual and bound
only by any mandatory minimum penalties that might apply in current law or as modified through the

Page 40

GAO-14-821 BOP Cost Efficiencies

policy option. The bed year fiscal year 2014 average marginal cost of $11,000 was not adjusted for
future inflation.
c

Estimated per capita savings were calculated using BOP fiscal year 2014 average per capita costs of
about $31,338. BOP’s fiscal year 2014 enacted Salaries and Expenses appropriation of $6.769 billion
was divided by the July 24, 2014 population of about 216,000. Future year savings were not adjusted
for inflation.

Option 2: Retroactively apply the Fair Sentencing Act of 2010 reductions in drug mandatory minimum penalties to offenders
currently incarcerated for crack cocaine–related offenses
Description
This option would retroactively apply to eligible incarcerated offenders
the reductions in mandatory minimum penalties for crack cocainerelated offenses that are provisions of the Fair Sentencing Act (FSA) of
2010, which did not apply to offenders already sentenced under the
a
original penalty guidelines. The USSC amended the sentencing
guidelines to implement these revisions effective November 1, 2011,
and also voted to apply that amendment retroactively to those already
sentenced under the guidelines that incorporated the original penalty
structure. However, because Congress did not make the statutory
changes to the mandatory minimum penalties retroactive, some
offenders were unable to receive the full or, in some cases, any benefit
from the commission’s action. This option would have the effect of
applying the new, lesser mandatory minimum penalties retroactively to
currently incarcerated offenders.
Estimated number of inmates eligible for release or
sentence reduction

An estimated 8,468 inmates would be eligible; 6685 (79 percent) would
be eligible for release over a period of 5 years following
b
implementation.

Estimated bed year reduction and cost savings

It is estimated that 37,400 bed years would be saved, or about $411
million. Of the total, about $325 million would be saved over first 5
c
years; and about another $86 million in the sixth year and after.
•
This could enhance fairness in federal sentencing by making the
sentence more commensurate with the crime. By applying the
same sentences for the same crime retroactively, it could increase
prison safety by reducing potential inmate conflict.
•
This could reduce BOP population, thereby reducing overcrowding
and improving prison safety; it could allow BOP to shift fund to
other priorities.
•
This could release offenders deemed to be at low risk for
reoffending and at high likelihood for successful reentry; it could
increase the likelihood of successful reentry and therefore save
money.

Expert-identified potential advantages

Expert-identified potential disadvantages

•
•

Released offenders may recidivate, which could increase crime.
Reductions in sentence length after conviction and sentencing
could reduce prosecutors’ ability to negotiate a plea bargain in
future cases

a
The mandatory minimums revised by the Fair Sentencing Act of 2010, Pub. L. No. 111-220, 124
Stat. 2372, had been enacted in the Anti-Drug Abuse Act of 1986, Pub. L. No. 99-570, 100 Stat.
3207, which included the federal criminal law distinction between “cocaine base” and other forms of
cocaine. The thresholds triggering the 10-year penalty—5 kilograms of powder cocaine and 50 grams
of cocaine base—created a 100-to-1 quantity ratio. The identical ratio was reflected in the 5-year

Page 41

GAO-14-821 BOP Cost Efficiencies

mandatory minimum thresholds as well: 500 grams of powder cocaine and five grams of cocaine
base both triggered the 5-year penalty. Among other things, the Fair Sentencing Act of 2010
increased the quantities of crack cocaine that trigger the 5- and 10-year statutory mandatory
minimum penalties—from 5 grams to 28 grams for 5-year mandatory minimums and from 50 to 280
grams for 10-year mandatory minimums—and eliminated the 5-year mandatory minimum for simple
possession of crack cocaine.
b

USSC data permitted calculation of bed year savings based on 8,468 cases

c

In developing its analysis, the USSC made a number of assumptions, including that Congress would
authorize the courts by statute to reduce the sentences of offenders in accordance with the
Guidelines Manual in effect on a specific date and bound only by the mandatory minimum penalties
as amended by the FSA. The bed year fiscal year 2014 average marginal cost of $11,000 was not
adjusted for future inflation.

Option 3: Avoid charging future low-level nonviolent drug offenders with a charge that results in a mandatory sentence
Description
This option has been implemented as part of the August 2013 direction
issued by the Attorney General instructing prosecutors in cases
involving the applicability of drug law mandatory minimum sentences
based on drug type and quantity to decline to charge the quantity
necessary to trigger a mandatory minimum sentence if the defendant
a
meets particular criteria.
Estimated number of inmates eligible for release or
sentence reduction

The Attorney General’s August 2013 direction applies only to future
offenders (i.e., it is not retroactive). The USSC estimates that it would
be applicable to about 620 offenders annually each year for each of the
next 5 years (or a total of about 3,100 offenders).

Estimated bed year reduction and cost savings

This would result in an estimated savings of 788 bed years over 5
b
years, or about $8.7 million.
•
This could help restore proportionality in federal drug sentences by
ensuring that sentence length is more commensurate with offense
severity. It also could serve to more often appropriately defer to
local and state prosecution in many cases.
•
It could help reduce BOP overcrowding and thereby reduce costs.
•
It could allow for more judicial or prosecutorial discretion or
alternative treatment opportunities.
•
It could allow more use of community punishments and treatment,
which are more likely to reduce recidivism than the same programs
in prison.

Expert-identified potential advantages

Expert-identified potential disadvantages

•
•

•
•

Released offenders may recidivate, which could increase crime.
This could reduce prosecutorial leverage to get plea bargains in
exchange for evidence against those higher up in a criminal
organization.
A change in policy could be quickly reversed by a different
administration or by Congress.
This could provide increased discretion to prosecutors, which could
make it difficult to ensure consistency in application.

a

According to the USSC, this option is comparable to making certain changes in application of the
safety valve.

b

Bed year fiscal year 2014 average marginal cost of $11,000 was not adjusted for future inflation.
USSC estimates for this option are based on the USSC’s understanding of the implementation of the

Page 42

GAO-14-821 BOP Cost Efficiencies

criteria established by DOJ in this policy; however, ultimate savings will depend on how DOJ
personnel interpret and apply these criteria in practice.

Option 4: Amend USSC sentencing guidelines for future drug offenders, reducing them two levels below current guidelines
Description
This option would reduce the potential sentences for all future drug
offenders by two levels below current guidelines, resulting in a roughly
b
18 percent reduction of sentence length below current lengths.

a

Estimated number of inmates eligible for release or
sentence reduction

An estimated 87,285 inmates would be eligible, based on 17,457
annually for each of the next 5 years.

Estimated bed year reduction and cost savings

Approximately 17,990 bed years would be saved over 5 years, or $198
million.
Given the size of these potential reductions in offenders’ sentences and
required bed space, per capita savings could occur; which means
c
savings could be as much as $564 million over the first 5 years.

Expert-identified potential advantages

•
•
•

Expert-identified potential disadvantages

•
•

•

This could significantly reduce average sentences, reducing BOP
overcrowding and costs.
It could make sentences more proportionate to the crime.
It could result in sentences that are individualized to the risks and
needs of the particular offender.
Released offenders may recidivate, which could increase crime.
Overall reduction of guideline length does not address the specific
offender risks and needs. Unilateral reduction in sentences could
be a relatively blunt approach to reducing sentences for all types of
drug offenders.
This may not address issues raised by statutory mandatory
minimum sentences or differentiate among crimes committed by
offenders receiving a reduced sentence.

a

This option was approved by a vote of the USSC on April 10, 2014, and transmitted to Congress on
April 30, 2014; the proposed amendment will go into effect on November 1, 2014, unless Congress
acts to modify or disapprove the amendment to the sentencing guidelines.

b

According to the USSC, if this option was implemented, and if the courts were to grant the full
reduction possible in each case, the projected new average sentence for the eligible offenders would
be 51 months versus a current average of 62 months, or a reduction of about 18 percent.

c

Per capita savings were calculated using BOP fiscal year 2014 average per capita costs. BOP’s
fiscal year 2014 enacted Salaries and Expenses appropriation of $6.769 billion was divided by the
July 24, 2014, population of about 216,000. Future-year savings were not adjusted for inflation.

Option 5: Apply the amendments to sentencing guidelines in option 4 also to incarcerated drug offenders—that is, to both
a
new drug offenders and retroactively to incarcerated drug offenders
Description
Under this option, the provisions of option 4 would apply both to new
drug offenders and also retroactively to incarcerated drug offenders.
This would reduce the sentences for all future and currently
incarcerated drug offenders by two levels below current USSC
guidelines, resulting in a roughly 19 percent reduction below current
b
lengths.

Page 43

GAO-14-821 BOP Cost Efficiencies

Option 5: Apply the amendments to sentencing guidelines in option 4 also to incarcerated drug offenders—that is, to both
a
new drug offenders and retroactively to incarcerated drug offenders
Estimated number of inmates eligible for release or
An estimated 132,635 inmates would be eligible for sentence reductions
sentence reduction
(87,285 from option 4 and 45,350 from applying option 4 retroactively.)
Estimated bed year reduction and cost savings

Expert-identified potential advantages

e

This is estimated to save 81,303 bed years over the first 5 years, saving
about $894.3 million. (17,990 bed years from future cost avoidance and
63,313 bed years saved from retroactively applying the reduction in
sentence lengths). For retroactivity over a period of 6 or more years, an
estimated additional 16,426 bed years would be saved, producing an
c
additional $181 million in savings. (Future-year cost avoidance savings
beyond the first 5 years were not calculated.)
Given the size of these potential reductions in offenders and required
bed space, per capita savings could occur; which means savings could
be as much as $2.548 billion over the first 5 years, and an additional
$515 million in the sixth year and after. This would mean a total savings
d
of about $3.1 billion.
•

•

•

Expert-identified potential disadvantages

•
•

•

This could reduce BOP population, thereby reducing overcrowding
and improving prison safety; it would increase the ability to shift
(saved) funds to other priorities such as rehabilitation programs.
It could enhance fairness and proportionality in sentencing by
ensuring that those in BOP custody for the same crime have similar
sentences.
It could reduce BOP population and therefore save money.
Released offenders may recidivate, which could increase crime.
This could reduce prosecutorial leverage to get plea bargains in
exchange for evidence against those higher up in a criminal
organization.
It could base release decisions solely upon guideline length
reduction without a separate and current risk/needs assessment
and without taking into consideration offender behavior and
programming while incarcerated, which may result in increased
f
recidivism.

a

Applying the sentencing guidelines in option 4 retroactively to eligible incarcerated offenders was
approved by a vote of the USSC on July 18, 2014; the decision amended a policy statement related
to the proposed amendment to the sentencing guidelines. As such, it corresponds to the April 30
USSC decision and does not require separate congressional action. Unless Congress acts to modify
or disapprove the corresponding April 30 amendment, beginning November 1, 2014, eligible
offenders can ask courts to reduce their sentences. Offenders whose requests are granted by the
courts may not be released earlier than November 1, 2015. The Attorney General announced after
the USSC vote that at his direction, BOP will begin notifying federal inmates of the opportunity to
apply for a reduction in sentence immediately. BOP subsequently notified inmates about the USSC
amendment.

b

According to the USSC, retroactive application of option 4 would result in an average sentence for
those eligible of 108 months, a reduction of 25 months from the current average of 133 months (18.8
percent less). The USSC also noted that sentences will be reduced retroactively only after
consideration by a judge in each case.

c

According to the USSC, 80 percent of the sentence reductions, and hence cost avoidance savings,
would occur within the first 5 years of implementation. Fiscal year 2014 average marginal costs of
$11,000 are not adjusted for future inflation. The remainder of the savings would occur over nearly 25
years, because of the length of the sentences of some of those potentially eligible.

Page 44

GAO-14-821 BOP Cost Efficiencies

d

Per capita savings were calculated using BOP fiscal year 2014 average per capita costs. BOP’s
fiscal year 2014 enacted Salaries and Expenses appropriation of $6.769 billion was divided by the
July 24, 2014, population of about 216,000. Future-year savings were not adjusted for inflation.
e

Expert entities were asked to comment on this option before it was approved. The approved version
delays implementation by 1 year (to November 2015). Since this did not affect the substance of the
proposed change, we report the experts’ comments as stated to us.

f

The USSC notes that sentences will be reduced retroactively only after consideration by a judge in
each case. The option reviewed by expert entities did not have this requirement as it was approved
by the USSC after the list of options had been sent to the expert entities for review and comment.

Option 6: Increase good time credits for good conduct time (GCT)
Description
BOP has discretionary authority to award credit toward the service of an
inmate’s sentence, beyond the time served, of up to 54 days per year of
sentence served if the inmate has displayed exemplary compliance with
institutional disciplinary regulations. As applied by BOP, this actually
results in 47 days earned per year of sentence imposed because
inmates do not earn GCT credit for years they do not ultimately serve
a
because of being released early. Under this option, the GCT credit
would be awarded to current offenders using a formula that would result
in 54 days’ credit per year of sentence, rather than the current BOP
interpretation that limits the GCT credit to 47 days per sentence year.
This would result in eligible inmates serving 85 percent of their
sentence, rather than the higher 87 percent average resulting from
BOP’s current interpretation of 18 U.S.C. § 3624(b).
Estimated number of inmates eligible for release or
sentence reduction

Approximately 179,265 inmates would be eligible.

Estimated bed year reduction and
cost savings

This would save 34,359 bed years, saving $378 million, of which $246
million (65 percent) would occur during the first 5 years after
implementation. The remaining $132 million would be saved in the sixth
year and after. (According to a DOJ 2012 analysis of this potential
change in calculating the GCT credit, it would take more than 30 years
for all eligible inmates to be affected, given the length of some
sentences.)
•
This could reduce BOP population/ overcrowding, which has
positive consequences, such as decreasing wait lists for recidivism
programs, increasing safety, or freeing up resources for other
priorities.
•
It could incentivize participation in programming aimed at reducing
recidivism, promoting successful reentry.
•
It could save money over the long term through reduced recidivism.

Expert-identified potential
b
advantages

Expert-identified potential disadvantages

•
•

•

It could require additional slots or spaces for inmates to participate
in qualifying programming, thereby requiring more funding.
It could undermine “truth in sentencing,” which was the goal of
mandatory sentences and public confidence in the criminal justice
system.
Released offenders may recidivate despite participation in
programming, thus increasing crime. For instance, inmates could
“game the system” to get admitted to antirecidivism programs in
order to get a reduced sentence.

a

See GAO-12-320 for further discussion of these issues.

Page 45

GAO-14-821 BOP Cost Efficiencies

b

Expert entities were asked to comment on both good conduct time credits and credited time that
might be earned through participation in anti-recidivism programs, including substance abuse
programs. However, because such programs incur costs, rather than producing cost savings, and
participation rates could not be specified, we asked the USSC to estimate the potential savings from
a change in the way that GCT is credited by BOP for current federal inmates.

Option 7: Make the current “safety valve” available to any drug offender with no offenses that receive 3 points under the
a
criminal history point system
Description
Under this option, the current statutory requirements for eligibility for
“safety valve” relief in drug cases involving a mandatory minimum
penalty would be expanded to cover incarcerated offenders with any
number of criminal history points under USSC guidelines, provided that
the offender does not have a conviction for which the sentence imposed
was a 3-point offense under USSC sentencing guidelines. The courts
would be authorized to resentence all federal offenders convicted of a
b
drug offense in a manner consistent with the revised statute.
Estimated number of inmates eligible for release or
sentence reduction

Approximately 11,949 inmates would be eligible (applies only
retroactively). Of these, 10,259 (86 percent) would be released within
the first 5 years following implementation.

Estimated bed year reduction and cost savings

This would save15,889 bed years, saving $175 million. Of this, about
$150.5 million (86 percent) would be saved within the first 5 years
following implementation.
•
It could reduce BOP population/overcrowding, which has positive
consequences, such as decreasing wait lists for recidivism
programs, increasing safety, or freeing up resources for other
priorities.
•
It could increases judicial discretion, resulting in less severe and
more proportionate and fair sentences.
•
It could help reduce BOP overcrowding and thereby reduce costs.

Expert-identified potential advantages

Expert-identified potential disadvantages

c

•
•

•

Released offenders may recidivate, which could increase crime.
It could reduce prosecutorial leverage to get plea bargains in
exchange for evidence against those higher up in a criminal
organization.
It may not address issues raised by statutory mandatory minimum
sentences or differentiate among crimes committed by offenders
receiving a reduced sentence. For example, point systems are
inflexible; change must recognize that three 1-point offenses are
not the same as one 3-point offense.

Source” | GAO-14-821
a

The safety valve was created by the Violent Crime Control and Law Enforcement Act of 1994, Pub.
L. No. 103-322, 108 Stat. 1796, which directed that certain defendants sentenced for certain drug
offenses who meet specified criteria would not be subject to mandatory minimum penalties.
Accordingly, only defendants sentenced after the guideline amendment’s November 1, 1987, effective
date who satisfy the safety valve requirements would be eligible for the reduced term of
imprisonment.

b

USSC guidelines are used to determine prison sentences based primarily on two factors: (1) the
conduct associated with the offense (the offense conduct, which produces the offense level) and (2)
the defendant’s criminal history (the criminal history category). In general, criminal history points are
calculated by adding 3 points for each prior sentence of imprisonment exceeding 1 year and 1 month;
adding 2 points for each prior sentence of imprisonment of at least 60 days but not more than 13
months; adding 1 point for each prior sentence of less than 60 days, and so forth. This option would

Page 46

GAO-14-821 BOP Cost Efficiencies

make more drug offenders eligible for sentencing by judges who would not be required to solely
follow mandatory sentencing guidelines.
c

The bed year fiscal year 2014 average annual marginal cost of $11,000 is not adjusted for future
inflation.

Option 8: Expand the safety valve as described in option 7 to include both nonviolent drug offenders and any nonviolent
offenders provided that they do not have a 3-point offense
Description
Under this option, the current statutory requirements for eligibility for
“safety valve” relief in drug cases involving a mandatory minimum
penalty would be expanded to cover both drug offenders and other
nonviolent (nondrug) offenders. No offender could have a conviction for
which the sentence imposed was a 3-point offense under USSC
sentencing guidelines. The courts would be authorized to resentence all
federal offenders convicted of a nonviolent offense in a manner
consistent with the revised statute.
Estimated number of inmates eligible for release or
sentence reduction

Approximately 38,930 inmates would be eligible.

Estimated bed year reduction and cost savings

This could save 40,552 bed years, saving about $446 million. Of this,
about $397 million (89 percent) would be saved in the first 5 years
following implementation. The remaining $49 million would be saved in
the sixth year and after.

Expert-identified potential advantages

•

•

•

Expert-Identified Potential Disadvantages

•

•

•

This could reduce BOP population/overcrowding and would thus
have positive consequences, such as decreasing wait lists for
recidivism programs, increasing safety, or freeing up resources for
other priorities.
It could result in sentences that are less severe and fairer, reducing
prison overcrowding and improving prison safety; cost savings
could permit shifting funds to other priorities, such as rehabilitation
programs.
It could allow more judicial discretion (judges can render sentences
focused upon the individual offender), thereby potentially
increasing the likelihood of a successful outcome while
simultaneously reducing the prison population.
It could reduce prosecutorial leverage to get plea bargains in
exchange for evidence against those higher up in a criminal
organization.
It could lead to disparities in sentencing across multiple
jurisdictions, which sentencing guidelines were intended to
address.
It could shorten sentences too much below original intent.

Sources: USSC analysis, GAO analysis of USSC and BOP data, and GAO analysis of expert responses. | GAO-14-821

As the Attorney General has noted, without action to reduce the current
size of BOP’s population, and projected future growth, BOP will continue
to require an ever greater share of DOJ’s budget, having increased from
about 14 percent of that budget in 1980 to nearly 25 percent in the fiscal
year 2015 request. As table 4 shows, a number of options outside of
BOP’s current authority exist by which the sentence lengths of

Page 47

GAO-14-821 BOP Cost Efficiencies

prospective and incarcerated offenders could be reduced (producing bed
year savings), as well as the number of offenders to be incarcerated,
such as under the Attorney General’s August 2013 Smart on Crime
direction (option 3). Hence, BOP’s costs could be reduced, but the range
of potential savings from these options is great. As illustrated in table 4,
actual savings are greater for options that apply to current offenders
compared with those that apply only prospectively. For example,
retroactively reducing drug sentences by about 44 percent (option 1)
would have the greatest estimated cost savings: $4.1 billion using
marginal cost savings, or $11.626 billion if per capita savings were to
occur. Other options would have a lesser impact, such as avoiding
charging future low-level, nonviolent drug offenders with charges that
result in mandatory minimum sentences, which could be a savings of $8.7
million in marginal costs over 5 years (option 3). However, as discussed
earlier, cost savings is one of many considerations in weighing which, if
any, option could be adopted. Equally important is consideration of the
advantages and disadvantages associated with each option.

Conclusions

According to BOP officials, BOP’s biggest challenges are managing the
continually increasing federal inmate population while providing for
inmates’ care and safety, as well as the safety of BOP staff and
surrounding communities, within budgeted levels. BOP has made
improvements in its effective use of resources a strategic goal. Toward
that end, BOP has implemented mechanisms that allow it to identify
opportunities for cost savings, such as an internal website through which
BOP staff can view cost efficiency efforts being implemented across BOP,
and undertaken cost-savings efforts, such as in energy conservation.
However, establishing a mechanism for relevant Central Office divisions
to consistently monitor bureau-wide corrective actions and assess their
progress in the presence of repeated frequent or significant findings could
help BOP better ensure that it is resolving such deficiencies promptly and,
ultimately, operating more efficiently.
BOP’s inmate population is its primary cost driver; however, BOP’s ability
to reduce its population, and thus its costs, are limited. Potential changes
outside of BOP’s authority could substantially reduce both BOP’s
population and its costs. Expert entities and states from which we
obtained views reported that there are potential trade-offs to
implementing these options. Some expert entities observed that release
of offenders poses the risk of recidivism and that reductions in sentence
length could reduce prosecutors’ leverage in extracting information from
low-level criminals that is used to prosecute higher-level ones. At the
same time, other expert entities reported potential advantages to these

Page 48

GAO-14-821 BOP Cost Efficiencies

options, such as making sentence lengths more commensurate with the
crime committed. These experts’ analyses of the possible savings and the
potential advantages and disadvantages, beyond cost savings, of these
options may be useful for policymakers to consider as they weigh whether
and how to address BOP’s population and, ultimately, its costs.

Recommendation for
Executive Action

To enable BOP to promptly address repeated frequent deficiencies and
other significant findings it identifies through its program reviews in areas
of high cost across multiple institutions, we recommend that the Director
of the Bureau of Prisons establish a mechanism for relevant Central
Office divisions to consistently monitor bureau-wide corrective actions.

Agency Comments
and Our Evaluation

We requested comments on a draft of this report from DOJ. In an email
received September 17, 2014, the DOJ liaison stated that DOJ concurred
with our recommendation. The Department did not provide official written
comments to include in our report, but did provide written technical
comments, which we incorporated as appropriate. The U.S. Sentencing
Commission also provided written technical comments, which we
incorporated as appropriate.
We are sending copies of this report to selected congressional
committees, the Attorney General, the Director of BOP, and other
interested parties. In addition, this report is available at no charge on the
GAO website at http://www.gao.gov.
If you or your staff have any further questions about this report, please
contact me at (202) 512-9627 or MaurerD@gao.gov. Contact points for
our Offices of Congressional Relations and Public Affairs may be found
on the last page of this report. Key contributors to this report are listed in
appendix VIII.

David C. Maurer
Director, Homeland Security and Justice Issues

Page 49

GAO-14-821 BOP Cost Efficiencies

List of Requesters
The Honorable Tom Coburn, M.D.
Ranking Member
Committee on Homeland Security and Governmental Affairs
United States Senate
The Honorable Charles E. Grassley
Ranking Member
Committee on the Judiciary
United States Senate
The Honorable John Cornyn
Ranking Member
Subcommittee on Immigration, Refugees and Border Security
Committee on the Judiciary
United States Senate
The Honorable Orrin G. Hatch
Ranking Member
Subcommittee on Oversight, Federal Rights and Agency Action
Committee on the Judiciary
United States Senate
The Honorable Michael S. Lee
Ranking Member
Subcommittee on Antitrust, Competition Policy and Consumer Rights
Committee on the Judiciary
United States Senate
The Honorable Bob Goodlatte
Chairman
Committee on the Judiciary
House of Representatives
The Honorable F. James Sensenbrenner Jr.
Chairman
Subcommittee on Crime, Terrorism, Homeland Security, and
Investigations
Committee on the Judiciary
House of Representatives
The Honorable Jason Chaffetz
House of Representatives

Page 50

GAO-14-821 BOP Cost Efficiencies

Appendix I: Objectives, Scope, and
Methodology
Appendix I: Objectives, Scope, and
Methodology

We addressed the following questions as part of this review:
1) What are the major costs for Bureau of Prisons (BOP) operations, and
what actions has BOP taken to implement cost savings?
2) To what extent does BOP have mechanisms to identify opportunities
for cost efficiencies and to take corrective actions that may improve
cost efficiency?
3) What potential changes, both within and outside of BOP’s authority,
could lead to cost reductions or improved efficiencies in BOP
operations, and what might be the potential impact of these changes?
To address the first question, we obtained historical obligation data from
BOP’s Financial Management Information System for fiscal years 20092013—a 5-year time frame to provide us with an adequate understanding
of trends in BOP obligation levels—to identify the major costs incurred by
BOP for its Salaries and Expenses account, which generally represents
98 percent of BOP’s budget. 1 We have assessed the reliability of these
data and determined them to be reliable for the purposes of this report.
This assessment included performing checks on the data received and
interviewing officials responsible for compiling and maintaining these
data. We also collected data (e.g., cost savings estimates prepared by
BOP) and documentation, such as memos and concept papers approved
by executive staff from fiscal years 2009 through 2013; through review of
these data and documents and interviews with relevant agency officials,
we determined that these data are also reliable for the purposes of this
report. In addition, we interviewed Central Office-, regional-, and
institutional-level officials to identify to the extent possible all existing cost
efficiency and savings initiatives adopted by BOP and their impact on its
overall budget.
To address the second question, we reviewed the processes and tools at
BOP during the same time period (fiscal years 2009-2013) that identify,
implement, and promote cost-efficiency and savings initiatives throughout
its institutions, such as executive staff meetings and a catalog compiling
cost-savings initiatives. With respect to identifying additional opportunities
to realize cost efficiencies or reduce costs, using our financial analysis as

1
BOP is appropriated funds through two accounts: Salaries and Expenses and Buildings
and Facilities. We focused our review on the Salaries and Expenses account as it
represents almost the entirety of BOP’s budget.

Page 51

GAO-14-821 BOP Cost Efficiencies

Appendix I: Objectives, Scope, and
Methodology

context, we analyzed elements of BOP’s internal control system related to
the control objective of achieving operational efficiencies and interviewed
relevant officials to assess whether BOP has a management structure
and processes to routinely assess its administrative and operational
activities for possible corrective action. Specifically, we reviewed BOP’s
mechanisms and processes leading to its internal review of operational
and administrative functions, including its process for taking corrective
action, related to high-cost areas and compared these characteristics with
those called for in Standards for Internal Control in the Federal
Government. 2
To address the third question, on potential policy changes both within and
outside BOP’s authority that could lead to BOP cost savings, we collected
analysis and documentation from the Department of Justice (DOJ), BOP,
the U.S. Sentencing Commission (USSC), and entities selected for their
expertise in criminal justice issues, and, in particular, potential changes to
federal sentencing policies. 3 We identified these expert entities through
several means, including (1) asking officials at both the USSC and the
Urban Institute who have previously worked on analyses and reports
relevant to prison population data and criminal justice to identify entities
they considered as expert in the field of sentencing reform and criminal
justice, (2) conducting a literature search to identify publications issued by
some of the entities on sentencing reform or corrections, and (3)
identifying entities with known expertise and authoritativeness (e.g., the
ABA) in the criminal justice field through literature searches. These
entities may not be representative of the universe of expert entities in the
criminal justice field and therefore may not represent all views on this
topic; however, their views provide insights. We obtained data from

2
GAO, Standards for Internal Control in the Federal Government, GAO/AIMD-00.21.3.1
(Washington, D.C.: Nov. 1, 1999). Internal control is an integral component of an
organization’s management that provides reasonable assurance that the following
objectives are being achieved: effectiveness and efficiency of operations, reliability of
financial reporting, and compliance with applicable laws and regulations.
3

The expert entities from which we requested or obtained analysis or information on
sentencing issues were the American Bar Association (ABA); the Association of State
Correctional Administrators; the Brennan Center for Justice; the Council of State
Governments/Justice Center; Families Against Mandatory Minimums; the Fraternal Order
of Police; the Heritage Foundation; the National Association of Assistant U.S. Attorneys;
the National Association of Former U.S. Attorneys; the Pew Charitable Trusts; “Right on
Crime,” a project of the Texas Public Policy Foundation in cooperation with the Justice
Fellowship; the Sentencing Project; the Urban Institute; and the Vera Institute of Justice.

Page 52

GAO-14-821 BOP Cost Efficiencies

Appendix I: Objectives, Scope, and
Methodology

calendar year 2011 through July 2014 from BOP on the impact on BOP’s
population and costs resulting from changes implemented by BOP to
broaden the criteria for compassionate release of offenders, as well as
analyses from the USSC on the estimated impact of the memorandum
accompanying the Attorney General’s August 2013 DOJ Smart on Crime
initiative that instructed prosecutors to decline to charge certain
defendants in certain types of drug cases in such a manner as to trigger
mandatory minimum sentences. For data on efforts taken under BOP’s
authority, we assessed the reliability by reviewing documentation and
interviewing knowledgeable officials; we determined the data to be
sufficiently reliable for the purposes of this report. For data on the impact
of policy changes outside BOP’s authority, we obtained analyses from the
USSC. According to our audit objectives, we determined based on its
expertise that the evidence obtained from the USSC provides a
reasonable basis for our findings and conclusions.
We reviewed sentencing reform options identified by the expert entities,
as well as analyses of these options from these entities. We identified
these entities on the basis of their demonstrated expertise as well as to
ensure that a wide range of views on the advantages or disadvantages of
potential changes to existing sentencing guidelines was covered. GAO is
not taking a position on any of these options, but presents information on
estimated cost savings and experts’ views of advantages and
disadvantages for such options to inform policymakers as they weigh
whether and how to address rising costs at BOP.
On the basis of the options we identified, we compiled a list of seven
sentencing guideline and related changes that were being widely
considered and discussed among expert entities at the time we began our
review and during the course of the review. We also included one change
resulting from a memorandum issued by the Attorney General in August
2013 to U.S. Attorneys instructing them to decline to charge certain
defendants in certain types of drug cases in such a manner as to trigger
mandatory minimum sentences. These selected options were based on
our assessment of proposals put forward by several sources, including
the Attorney General, expert entities, states with experience with criminal
justice reform, and advocacy entities with a range of views on criminal
justice reform. See appendix II for the letter sent to expert entities asking
for their assessment of options, which provides a list of the proposals that
were used to develop the selected options. During the course of our
review, two of these options were approved by the USSC, with the
amendment to the sentencing guidelines forwarded to Congress. This
amendment will become effective on November 1, 2014, unless Congress

Page 53

GAO-14-821 BOP Cost Efficiencies

Appendix I: Objectives, Scope, and
Methodology

acts to modify or disapprove the amendment. The two options approved
are closely related—the first, approved by the USSC on April 10, 2014,
reduces the sentences for future (federal) drug offenders, such that there
would be an approximately 18 percent reduction in the average length of
such sentences compared with sentences under current guidelines. The
second was approved on July 18, 2014, by USSC, and applies the
reduction retroactively—that is, to currently incarcerated drug offenders. 4
After the USSC July 18 vote to make the drug sentence reductions
retroactive (if eligibility criteria are met), the Attorney General issued a
statement that, at his direction, BOP will begin notifying federal inmates of
the opportunity to apply for a reduction in sentence immediately. Unless
Congress modifies or disapproves the amendment to the sentencing
guidelines submitted by the USSC, beginning November 1, 2014, eligible
offenders can ask courts to reduce their sentences. Offenders whose
requests are granted by the courts can be released no earlier than
November 1, 2015, under the USSC provision. For the purposes of this
report, we refer to these two changes by the USSC, and the Attorney
General’s August 2013 directive to U.S. Attorneys, and to the other
options collectively as eight options.
For each of these eight options, we asked the USSC to estimate the
impact on the size of the BOP inmate population. The USSC calculates
such impact in bed years; 1 bed year is defined by the USSC as 12
months of prison time (e.g., one offender in prison for 1 year, or three
offenders in prison for 4 months each, and so forth). Using the USSC bed
years’ savings estimates, we then calculated cost savings or cost
avoidance that could result from implementing a given option. 5 The USSC
analyses of these eight options are reproduced in appendix VII.

4

According to the USSC, the July 18, 2014, action amends a policy statement to provide
for retroactive effect to the proposed amendment to the sentencing guidelines and as
such, does not require submission to Congress.

5
We calculated savings that would result from reductions in BOP population or cost
avoidance savings resulting from increases below what would otherwise occur if an option
was not implemented, by multiplying bed years saved by the annual marginal cost for a
prisoner. BOP defines marginal costs as the cost of adding one offender for 1 year; these
cover food, medical, clothing, and utilities. For fiscal year 2014, the average annual
marginal cost per offender is about $11,000. For some options that could result in reduced
sentences and earlier releases of about 78,000 or more offenders, we also calculated per
capita savings, as such population reductions that could allow BOP to reduce its staff or
close facilities.

Page 54

GAO-14-821 BOP Cost Efficiencies

Appendix I: Objectives, Scope, and
Methodology

In addition, we asked the 14 expert entities to comment on the options
with regard to their view of the advantages or disadvantages of
implementation of the eight options. We also identified 4 states (Texas,
Louisiana, Ohio, and North Carolina) that have had at least 1 year’s
experience in implementing state-level sentencing and criminal justice
changes similar to the ones being discussed for federal offenders, and
asked them to comment on the potential options for change. 6 We
selected Louisiana, Ohio, and North Carolina also because they are
participants in the joint DOJ-Pew Charitable Trust Justice Reinvestment
Initiative (JRI), which funds implementation of changes to state
sentencing and criminal justice policies intended to reduce prison
populations and recidivism, and these states were also the largest in
terms of annual spending on corrections among the JRI states with at
least 1 year’s experience in implementing JRI-related initiatives. 7 We
selected Texas because, as of 2012, it had both the largest state prison
population in the United States and was cited by the Urban Institute as an
early leader in criminal justice reform initiatives that DOJ used as a model
for the Justice Reinvestment Initiative program. We received responses
from 17 of the 18 entities (the 14 expert entities and 4 states) to which we
sent the letter. 8
We conducted this performance audit from July 2013 to September 2014
in accordance with generally accepted government auditing standards.
Those standards require that we plan and perform the audit to obtain
sufficient, appropriate evidence to provide a reasonable basis for our

6
As of January 2014, there were 17 states that are or have been participants in DOJ’s
Justice Reinvestment Initiative (JRI) program. Of these 17, only 3—Louisiana, Ohio, and
North Carolina—had had at least 1 year’s experience implementing JRI-related programs.
Texas is not a JRI participant but is cited as one of the models for JRI programs by DOJ.
7

DOJ and the Urban Institute: Justice Reinvestment Initiative State Assessment Report
(January 2014: Washington, D.C.) The DOJ Bureau of Justice Assistance funded the JRI,
in a public-private partnership with the Pew Charitable Trusts, to provide technical
assistance and financial support for system-wide criminal justice reform efforts. Under the
JRI model, a bipartisan working group uses comprehensive data analyses to identify the
drivers of the local corrections population and costs and foster support for a set of costeffective policy options addressing those drivers.

8

Of the 18 entities to which we sent a request for comments, 17 acknowledged receipt of
our request. Of these 17, 3 did not provide substantive comments, leaving 14 that
provided comments. Of the 3 that did not provide substantive comments, 1 referred us to
multiple articles published on criminal justice issues.

Page 55

GAO-14-821 BOP Cost Efficiencies

Appendix I: Objectives, Scope, and
Methodology

findings and conclusions based on our audit objectives. We believe that
the evidence obtained provides a reasonable basis for our findings and
conclusions based on our audit objectives.

Page 56

GAO-14-821 BOP Cost Efficiencies

Appendix II: Letter Sent to Select States and
Expert Entities
Appendix II: Letter Sent to Select States and
Expert Entities

March 24, 2014
The U.S. Government Accountability Office (GAO), the investigative arm
of the U.S. Congress, is carrying out a research engagement (GAO job
#441173) on ways in which the Federal Bureau of Prisons (BOP) can
achieve cost efficiencies. As part of this work, we will be listing select
options that would require action by entities other than BOP—such as
changes enacted by Congress and the President that could affect the size
of the BOP prison population, or changes in sentencing guidelines that
could be promulgated by the U.S. Sentencing Commission (USSC) and
possibly also applied retroactively. These options include proposals
identified by a number of experts or included in pending legislation that
would likely help reduce costs at BOP. We are writing you to ask for your
comments or observations on these select options.
These options are based on an assessment of varying proposals put
forward by several sources:
•

Proposals and guidance from the U.S. Attorney General on
sentencing reform, including changes to charging procedures ordered
in August 2013 as part of the DOJ Smart on Crime Initiative to avoid
incarceration of low-level, non-violent drug offenders.

•

Potential cost-saving changes or concepts identified by expert
entities, including the USSC; the Pew Charitable Trusts; the Urban
Institute; the Heritage Foundation; the Vera Institute of Justice; the
Council of State Governments/Justice Center; and the American Bar
Association, some of which are also included in pending legislation.

•

The experiences of various states in recent years with criminal justice
reform, often intended to reduce prison populations and recidivism,
including Texas and 17 states involved in the DOJ Justice
Reinvestment Initiative (JRI) and as reported on in the January 2014
joint DOJ-Urban Institute report, Justice Reinvestment Initiative State
Assessment Report. 1

1

DOJ and the Urban Institute: Justice Reinvestment Initiative State Assessment Report
(January 2014: Washington, D.C.) The Bureau of Justice Assistance (BJA) funded the
JRI, in a public-private partnership with the Pew Charitable Trusts, to provide technical assistance and financial support for system-wide criminal justice reform efforts. Under the
JRI model, a bipartisan working group comprising key policymakers uses comprehensive
data analyses to identify the drivers of the local corrections population and costs and
foster support for a set of cost-effective policy options addressing those drivers.

Page 57

GAO-14-821 BOP Cost Efficiencies

Appendix II: Letter Sent to Select States and
Expert Entities

•

Comments and analyses by several expert/advocacy entities with a
range of views on criminal justice reform including the National
Association of Assistant United States Attorneys; the National
Association of Former United States Attorneys; the Brennan Center
for Justice at New York University School of Law; Families Against
Mandatory Minimums; and Right on Crime: A project of the Texas
Public Policy Foundation in cooperation with Justice Fellowship.

Based on common policy themes assessed or reported by the
aforementioned sources, the options listed below are grouped into three
broad categories (in most cases, there are several related proposals or
concepts that have been put forth by various expert entities or experts):
1) Sentencing Reform—Proposals to Reduce Length of Mandatory
Sentences;
2) Enhancement of “Good Conduct Time” reduction-in-sentence through
good behavior by inmates and/or through participation in programs
intended to reduce recidivism, enhance post-release employment
chances, and/or reduce substance abuse; and
3) Proposals to expand the safety valve for offenders. A “safety valve” is
a statutory exception to mandatory minimum sentencing laws that
authorizes a judge to sentence a person below a mandatory minimum
term if certain conditions are met. The current federal sentencing
guidelines provide for a reduced sentence for persons who meet the
current federal safety valve criteria (which apply only to drug
trafficking cases) even in cases where a mandatory minimum penalty
would not otherwise apply.

Instructions

For each option, we would like to ask you to briefly identify any potential
“advantage” or “disadvantage”, based on your organization’s expertise or,
for states, based on your experience in implementing some version of
them (we recognize these vary considerably in detail). For example, an
“advantage” for most of these proposals is sometimes stated to be to
reduce overcrowding of prisons, or to reduce the waiting list for antirecidivism educational or substance abuse programs. Alternatively, a
“disadvantage” for some of these, as identified by some experts, is that a
certain percentage of those released may commit crimes again and that,
therefore, early release (v. serving the full sentence) may put someone
back into society who may become a recidivist earlier than would
otherwise have occurred.

Page 58

GAO-14-821 BOP Cost Efficiencies

Appendix II: Letter Sent to Select States and
Expert Entities

GAO will generally present only the aggregate results of the responses to
these questions in our report to Congress. Although in some cases
individual responses may be discussed, the report will not include any
information that could be used to identify individual respondents.
Identifying information will not be released outside of GAO, unless
compelled by law or pursuant to a Congressional request. See 31 U.S.C.
§716.
Please start by saving the document as a Microsoft Word file to a hard
drive before completing it, as your comments will otherwise not be saved.
There are comment boxes below each of the options. Please type your
responses in the space provided, which will expand automatically to
accommodate your comments.
Group 1: Proposals to Reduce Length of Mandatory Sentences
Option A: Reduce the length of mandatory minimum sentences for drugrelated offenses –for example, reduce the mandatory minimum sentences
in drug trafficking cases from 20 years to 10 years; 10 years to 5 years;
and 5 years to 2 years.
Your Comments
Advantage:
Disadvantage:
Additional:

❏ Please check box if Not Applicable or Not Enough
Information to Comment

Page 59

GAO-14-821 BOP Cost Efficiencies

Appendix II: Letter Sent to Select States and
Expert Entities

Option B: Make the amendments to drug mandatory minimum penalties
of the Fair Sentencing Act of 2010 retroactive to those offenders currently
incarcerated for crack cocaine-related offenses.
Your Comments
Advantage:
Disadvantage:
Additional:

❏ Please check box if Not Applicable or Not Enough
Information to Comment

Option C: Avoid charging low-level non-violent drug offenders with a
charge which results in a mandatory sentence.
Your Comments
Advantage:
Disadvantage:
Additional:

❏ Please check box if Not Applicable or Not Enough
Information to Comment

Page 60

GAO-14-821 BOP Cost Efficiencies

Appendix II: Letter Sent to Select States and
Expert Entities

Option D: Amend sentencing guidelines for all drug offenders—for
example, reduce such sentences by a given percentage below current
lengths (e.g., by 20 percent; by 10 percent).
Your Comments
Advantage:
Disadvantage:
Additional:

❏ Please check box if Not Applicable or Not Enough
Information to Comment

Option E: Apply the amendments to sentencing guidelines in Option D to
both new drug offenders and retroactively to incarcerated drug offenders.
Your Comments
Advantage:
Disadvantage:
Additional:

❏ Please check box if Not Applicable or Not Enough
Information to Comment

Page 61

GAO-14-821 BOP Cost Efficiencies

Appendix II: Letter Sent to Select States and
Expert Entities

Group 2: Proposals to Reduce Sentence Time Based on Good
Conduct and/or Participation in Anti-Recidivism and Substance
Abuse Programs
Option F: Increase good time credits for good conduct and/or participation
in anti-recidivism programs, including substance abuse programs.
Your Comments
Advantage:
Disadvantage:
Additional:

❏ Please check box if Not Applicable or Not Enough
Information to Comment

Group 3: Proposals to Expand the Safety Valve for Drug and/or
Other Offenders
Overview: The “safety valve” was created by the Violent Crime Control
and Law Enforcement Act of 1994, which directed that lower-level
defendants sentenced for certain drug offenses who meet specified
criteria will not be subject to mandatory minimum penalties. Accordingly,
only defendants sentenced after the guideline amendment’s November 1,
1995 effective date who satisfy the safety valve requirements are eligible
for the reduced term of imprisonment.

Page 62

GAO-14-821 BOP Cost Efficiencies

Appendix II: Letter Sent to Select States and
Expert Entities

Select Options:
Option G: Expand the current “safety valve” for drug offenders to make it
available to offenders with more than one prior minor offense up to a
certain limit (for instance, up to three criminal history points.) 2
Your Comments
Advantage:
Disadvantage:
Additional:

❏ Please check box if Not Applicable or Not Enough
Information to Comment

Option H: Expand the safety valve as described in Option G to include
both drug offenders and other non-violent offenders (this would not apply
to any sex offenders).
Your Comments
Advantage:
Disadvantage:
Additional:

❏ Please check box if Not Applicable or Not Enough

2
USSC guidelines are used to determine prison sentences based primarily on two factors:
(1) the conduct associated with the offense (the offense conduct, which produces the
offense level) and (2) the defendant’s criminal history (the criminal history category). In
general, criminal history points are calculated by adding 3 points for each prior sentence
of imprisonment exceeding one year and one month; adding 2 points for each prior
sentence of imprisonment of at least sixty days but not more than 13 months; adding 1
point for each prior sentence of less than sixty days, and so forth. This option would make
more drug offenders eligible for sentencing by judges who would not be required to solely
follow mandatory sentencing guidelines.

Page 63

GAO-14-821 BOP Cost Efficiencies

Appendix II: Letter Sent to Select States and
Expert Entities

Information to Comment

Additional Comments or Suggested Options
Finally, in the box below, please make any other comments or
suggestions you may have, especially with regard to whether there are
any other options for making cost effective changes to criminal justice
rules or processes that you wish to bring to our attention.
Additional Comments or Observations: Other Options to Consider
This completes the requested assessment of potential options we are
considering. Thank you for your time in providing your input on these
options with regard to the federal criminal justice system.

Page 64

GAO-14-821 BOP Cost Efficiencies

Appendix III: Data on the Federal Bureau of
Prisons’ (BOP) Largest Cost Centers
Appendix III: Data on the Federal Bureau of
Prisons’ (BOP) Largest Cost Centers

Table 5 presents annual obligations data for fiscal years 2009 through
2013 for BOP’s 20 largest cost centers during fiscal year 2013.
Table 5: Annual Obligations for Each of the Bureau of Prisons’ (BOP) 20 Largest Cost Centers, Fiscal Years 2009 through
2013
Cost center

2009

Correctional Services

2011

2012

2013

$1,349,084,739 $1,429,586,006 $1,469,511,315 $1,495,434,551 $1,513,151,472

Privately Operated Secure Contract Facilities
BOP Medical Care

2010

a

Non-BOP Medical Care

$486,680,887

$544,310,237

$556,674,258

$625,636,747

$624,456,406

$493,979,657

$518,183,641

$545,304,468

$565,568,091

$595,492,696

$319,585,026

$351,572,156

$362,448,629

$385,865,248

$419,927,987

Food Service

$354,285,520

$366,808,966

$386,830,370

$403,446,856

$408,021,963

Unit Management

$287,515,668

$296,250,747

$300,984,838

$304,402,809

$305,156,264

Contract Residential Reentry Centers

$244,444,527

$259,937,098

$271,605,677

$280,878,201

$296,123,954

Construction and Mechanical Services

$246,547,837

$258,751,128

$270,133,073

$283,939,306

$283,682,155

Purchased Utilities

$231,373,396

$238,163,022

$239,485,407

$285,387,855

$250,421,579

Uncontrollables-Administrative

$120,313,499

$138,476,463

$167,144,272

$177,159,116

$177,549,624

Education

$108,573,667

$115,174,065

$120,343,124

$125,397,895

$125,822,670

Chief Executive Officer

$103,826,162

$108,540,721

$109,648,711

$108,185,902

$110,346,853

Financial Management

$97,038,156

$101,538,642

$103,963,681

$103,675,217

$107,468,147

Management and Administrative Support

$90,888,006

$96,183,777

$99,626,401

$101,044,832

$103,648,380

Inmate Services

$63,743,317

$79,161,534

$72,535,791

$86,344,508

$85,059,211

Leisure Time Activities

$65,441,546

$70,407,608

$73,291,323

$74,274,020

$76,032,065

Drug Abuse Program

$58,528,233

$64,109,527

$66,826,055

$69,733,453

$75,883,387

Psychology Services

$44,748,555

$47,206,056

$49,624,556

$51,534,118

$52,438,931

Power Plant

$45,741,334

$48,326,306

$49,020,235

$49,020,897

$50,507,908

Employee Services Department

$39,935,664

$41,446,038

$43,837,987

$44,717,642

$44,002,882

Source: GAO analysis of BOP data. | GAO-14-821

Notes: All program areas above represent obligations incurred at the institutional level, except for the
following:
Privately Operated Secure Contract Facilities (Contract Confinement Cost Centers, nationwide)
Contract Residential Reentry Centers (Contract Confinement Cost Centers, nationwide).
Uncontrollables-Administrative (National Programs). This cost center represents items that are
tracked and funded out of the Central Office on behalf of the institutions, such as DOJ phone
services, National Finance Center payroll processing, mail services, transit subsidies, and other
items.
a

BOP Medical care includes costs of salaries and other costs associated with Public Health Services
staff. According to BOP, these are medical professionals employed by the Department of Health and
Human Services (HHS) who provide medical services to BOP institutions. While these Public Health
Services staff are technically paid by HHS, BOP transfers the necessary funds over to HHS to pay for
their services.

Page 65

GAO-14-821 BOP Cost Efficiencies

Appendix IV: Data on the Federal Bureau of
Prisons’ (BOP) Largest Subobject Class
Expenses
Appendix IV: Data on the Federal Bureau of
Prisons’ (BOP) Largest Subobject Class
Expenses

Table 6 presents annual obligations data for fiscal years 2009 through
2013 for BOP’s 20 largest subobject classes in fiscal year 2013.
Compensation for full-time permanent staff has consistently been the
greatest area of expense; personnel compensation and benefits account
for most of the subobject classes listed below.
Table 6: Bureau of Prisons’ (BOP) Obligations by Its 20 Largest Subobject Classes, Fiscal Years 2009 through 2013
Object/subobject
class

2009

2010

2011

2012

2013

$1,843,183,744

$1,935,330,619

$1,982,819,749

$1,984,097,551

$2,012,540,824

Federal Employees’
Retirement System
Retirement, Law
Enforcement Officers,
and Firefighters

$493,494,905

$521,660,581

$553,491,039

$569,164,019

$542,789,373

Federal Employees
Health Benefits Act

$213,563,268

$232,974,957

$258,500,221

$272,741,587

$269,165,149

Wage Board-Permanent
Appointment

$222,097,563

$232,822,515

$239,808,681

$242,571,507

$245,682,761

Federal Insurance
Contributions Act

$133,085,312

$140,256,606

$142,415,468

$143,237,577

$136,549,567

Overtime, Other than
Wage Board

$86,672,551

$85,966,848

$78,938,623

$89,297,894

$90,438,851

b

$1,197

$4,181

Thrift Savings Plan
Matching Government
Contribution

$65,858,897

$71,003,306

$73,361,984

$73,567,039

$69,859,878

Employee
Compensation (Injury or
Death)

$50,387,420

$54,276,066

$59,573,568

$60,616,186

$63,622,470

Holiday Pay

$36,006,208

$36,497,469

$35,959,582

$36,253,718

$39,779,485

$640,758,937

$751,876,304

$781,896,029

$861,571,108

$936,791,945

$309,231,120

$290,455,682

$307,978,195

$282,334,113

$250,236,250

Personnel
Compensation and
a
Benefits
Full Time-Permanent
Appointment

Contribution Accrual

$72,455,016

Other contractual services
Contract Services

c

Housing of Prisoners,
Temporary Housing for
Federally Assisted
d
Witnesses

Page 66

GAO-14-821 BOP Cost Efficiencies

Appendix IV: Data on the Federal Bureau of
Prisons’ (BOP) Largest Subobject Class
Expenses

Medical Hospital
Services

$187,987,798

$190,474,836

$180,663,439

$196,405,226

$195,153,846

$125,846,279

$133,993,847

$154,818,892

$154,593,414

$153,185,222

$290,468,851

$303,532,259

$321,588,084

$334,613,110

$346,619,862

Food

$92,074,252

$90,759,178

$101,156,094

$104,630,698

$108,460,308

Pharmaceutical, Other

$62,758,400

$62,558,146

$63,259,284

$71,464,265

$75,396,816

Other Materials and
g
Supplies

$35,919,356

$48,650,402

$34,869,912

$44,093,428

$42,889,721

Other Services

e

Supplies and materials
Supplies/Purchase
f
Card

Rent, communications, and utilities
Electricity

$99,700,242

$103,055,828

$106,759,993

$103,184,518

$105,136,797

Water

$32,981,285

$34,776,915

$37,104,912

$38,153,278

$41,942,871

Source: GAO analysis of BOP data. | GAO-14-821

Notes: The obligations above do not include funds to pay salaries and other costs associated with
Public Health Services staff. According to BOP, these are medical professionals employed by the
Department of Health and Human Services (HHS) who provide medical services to BOP institutions.
BOP tracks these obligations only by object classification and does not break them down by
subobject class. For fiscal year 2013, the object class obligations were approximately as follows:
Personnel Compensation, $71 million; Personnel Benefits, $31 million; Transportation of Things,
$437,000; Travel and Transportation of Persons, $133,000; Other Contractual Services $100,000.
a

Items under the Personnel Compensation and Personnel Benefits object classifications are
combined in this table.

b

Contribution Accrual—month-end accrual of DOJ contributions chargeable to other subobject
classes, such as Federal Employees Health Benefits Act and Federal Insurance Contributions Act
contributions. The accrual is reversed in the following month.
c

Contract Services—contractual charges for consultants and other contractors, such as doctors,
dentists, teachers, chaplains, legal advisors, architects, engineers, surveyors, appraisers, inmate
commissary clerks, etc.

d

Housing of Prisoners, Temporary Housing for Federally Assisted Witnesses—charges for the
housing of federal prisoners under the jurisdiction of the United States Marshals Service who are
being held in nonfederal jail facilities. Charges for housing and moving expenses for witnesses under
the Emergency Witness Assistance Program. Includes, but is not limited to, prisoners being held in
state/local facilities, unsentenced prisoners, detained witnesses, the expense of 1 month’s deposit on
a rental property, moving expenses, truck rentals, etc., if such expenses are more cost-effective than
hotel/motel stays; it may also include dependents and other family members, as appropriate.
e

Other Services—charges for services not otherwise classified. Includes premiums for surety bonds,
money order fees, registration and license fees, charges for sanitation and towel service, charges for
General Services Administration information centers and water cooler services.

f

Supplies/Purchase Card —purchases made with the government-issued credit card (nontravel) for
supplies not to exceed $3,000. BOP officials noted that some contracting officers can use the cards to
make payments of up to $30,000.

g

Other Materials and Supplies—charges for materials and supplies not otherwise classified.

Page 67

GAO-14-821 BOP Cost Efficiencies

Appendix V: Additional Federal Bureau of
Prisons’ (BOP) Efforts to Achieve Cost
Savings and Efficiencies
Appendix V: Additional Federal Bureau of
Prisons’ (BOP) Efforts to Achieve Cost
Savings and Efficiencies

In addition to efforts discussed earlier, BOP also reports pursuing the
following initiatives that could provide current and future cost savings,
shown in Table 7.
Table 7: Additional Cost-Savings Efforts Under Way or Being Pursued by the Bureau of Prisons (BOP)
Cost savings initiative

Description

Electronic Inmate Central File project

This is an ongoing effort in BOP to convert inmate paper records to electronic files. In
2010, BOP estimated that the logistics of moving, managing, and accounting for inmate
paper files required approximately 1.2 million staff hours annually. BOP began this project
to modernize bureau operations and repurpose the time spent managing these paper files
back to focusing on the custody and care of inmates. BOP is currently testing an
approach to integrate its data with that from the U.S. Marshals Service that it plans to
complete by the end of August 2014. If successful, BOP officials state that they would
then transfer documents from shared directories into the electronic files, which would be
complete by the end of September 2014.

Program and administrative reductions

BOP’s fiscal year 2015 Congressional Budget Justification contains $157 million in
program offsets through program and administrative reductions, which BOP officials
reported will not be identified until 2015 funds are appropriated. According to BOP, these
funds are necessary to pay for increases in costs, including pay raises, retirement
contributions, and General Services Administration rent, among other things.

Medical Claims Adjudication

A contractor reviews the accuracy of payments for medical services provided through
contractual agreements with health care providers, currently in place at 25 BOP facilities.
BOP does not routinely track cost savings of these adjudications as they are considered
more of an operational efficiency effort; however, officials stated that when this effort
began at Federal Medical Center Butner, about $1.3 million in savings was originally
identified. As this effort has grown in scope, according to BOP officials, cost savings are
difficult to quantify as service providers become more focused on accurate billings.

Medical Contracting Initiatives for
Residential Reentry Center (RRC)
Inmates

BOP is examining the feasibility of regional contracting options at specific locations for
medical, dental, and mental health services for inmates at RRCs. Specifically, BOP
acquisition staff are formulating a procurement strategy and engaged in informal
discussions with vendors to identify an approach that will reduce medical costs while
assisting RRC inmates’ transition into their community. At this point, BOP has not
established any time frames for this acquisition of services. In addition, BOP is hiring
three Health Service Specialists to conduct medical bill adjudication for these services.

Mail order pharmacy

BOP is pursuing a Mail Order Pharmacy program to consolidate prescription drugs into
one main inventory and buy the least expensive generic medications, as well as automate
prescriptions, pending funding. However, this program has been on hold as BOP is
working to identify potential new sites for this inventory, and funding for building
renovation has been unavailable.

National contracts for medical resources

Through representation on the interagency Strategic Sourcing Leadership Council, BOP is
pursuing use of a Department of Defense prime vendor contract to purchase medical and
a
surgical supplies.

Page 68

GAO-14-821 BOP Cost Efficiencies

Appendix V: Additional Federal Bureau of
Prisons’ (BOP) Efforts to Achieve Cost
Savings and Efficiencies

Cost savings initiative

Description

Measuring cost effectiveness of health
services

BOP’s Health Services Division has tentatively identified nine metrics related to evaluating
efficiency of health care services to report to its National Governing Board. These metrics
will provide data on top diagnostic codes, top procedural codes, rate of rejected medical
claims, drug expenditures, catastrophic care costs, outside medical trips, hospital
readmission rates (catastrophic), annual health care expenditures, and average annual
percentage growth in health care expenditures. BOP reports that all nine cannot be
gathered in a systematic way, so the Health Services Division is researching alternative or
proxy measures, with no current projected implementation date.

Purchase of garbage trucks

Over the last 3 years, BOP has purchased and provided 26 new garbage trucks, and
estimates that about one-third of all institutions haul away garbage on their own, rather
than paying contractors.

Source: GAO analysis of BOP information. | GAO-14-821
a

The Strategic Sourcing Leadership Council was formed to lead governmental efforts in increasing
the use of government-wide management and sourcing of goods and services in order to save money
in contracting. This council is chaired by the Administrator for Federal Procurement Policy and
contains representatives from a number of federal agencies.

Page 69

GAO-14-821 BOP Cost Efficiencies

Appendix VI: Expert Entities Providing Input
on Prison Population Reduction Options
Appendix VI: Expert Entities Providing Input
on Prison Population Reduction Options

The expert sources that provided us with analysis or information on
sentencing issues are listed in table 8.
Table 8: Expert Entities That GAO Consulted in Reviewing Selected Policy Options
Organization

Description

American Bar Association

A professional organization with nearly 400,000 voluntary members that supports the legal
profession with resources for legal professionals while working to improve the
administration of justice, accredit law schools, establish model ethical codes, and more.

Association of State Correctional
Administrators

An organization consisting of state and Bureau of Prisons (BOP) officials focused on
improving correctional practices. Among other things, the Association of State
Correctional Administrators works to identify and promulgate correctional practices that
are likely to reduce the number of offenders returning to prison after release.

Brennan Center for Justice, New York
University School of Law

A nonpartisan law and policy institute that provides legal strategy and empirical research
for legal and policy efforts in areas including mass incarceration.

Council of State Governments/Justice
Center

A region-based forum that fosters the exchange of insights and ideas to help state officials
shape public policy. The Council of State Governments’ Justice Center is a national
nonprofit organization that provides evidence-based strategies and advice to policymakers
at the local, state, and federal government levels with the goal of increasing public
safety—including work on reentry issues and reducing spending on corrections.

Families Against Mandatory Minimums

A nonprofit organization that advocates for federal and state prison sentencing reforms.

Fraternal Order of Police

With over 325,000 sworn law enforcement officers as members, this organization works to
improve the working conditions of law enforcement officers and the safety of communities
through education, information, community involvement, and employee representation.

Heritage Foundation

An independent, tax-exempt research and education institution that studies and advises
policymakers on a variety of issues, such as federal spending and criminal law.

National Association of Assistant U.S.
Attorneys

Formed to protect, promote, and advance the mission of Assistant U.S. Attorneys; their
membership contains experts in a number of fields such as gang and narcotics
prosecutions, among other things.

National Association of Former U.S.
Attorneys

Established to promote the litigating authority and independence of the office of the U.S.
Attorney, the association provides education and continuing professional relationships for
former U.S. Attorneys.

Pew Charitable Trusts

An independent, nonprofit research and public policy organization focused on issues such
as family financial security and various legal issues, among other things.

“Right on Crime,” a project of the Texas
Public Policy Foundation in cooperation
with Justice Fellowship

The Texas Public Policy Foundation is a nonprofit, independent research institute focused
on issues of personal responsibility and free enterprise. Right on Crime is an initiative
focused on conservative ideas for criminal justice and prison reform in Texas and the
nation.

The Sentencing Project

Conducts research and advocates for criminal sentencing reform.

Urban Institute

An organization that gathers data, conducts research, evaluates programs, offers
technical assistance overseas, and educates the public on social and economic issues to
foster public policy developments in areas such as justice, health care, and housing.

Vera Institute of Justice

An independent, nonpartisan organization that conducts data and policy analysis to
provide assistance to government agencies on issues such as relations between police
and immigrant communities and policy options to safely reduce prison and jail
populations, among other things.

Source: GAO analysis of information from each entity listed above. | GAO-14-821

Page 70

GAO-14-821 BOP Cost Efficiencies

Appendix VII: Analyses of Options Provided
by the U.S. Sentencing Commission (USSC)
Appendix VII: Analyses of Options Provided
by the U.S. Sentencing Commission (USSC)

This appendix reproduces the methodology and analyses prepared by the
U.S. Sentencing Commission (USSC) in developing estimates of the
impact on the Federal Bureau of Prisons (BOP) inmate population of eight
options for cost saving that are outside BOP’s authority and that are
widely discussed among criminal justice policy experts and entities. GAO
identified these eight options based on our assessment of proposals put
forward by several sources, including the Attorney General, expert
entities, states with experience with criminal justice reform, and advocacy
entities with a range of views on criminal justice reform. GAO is not taking
a position on any of these options, but presents information on them to
inform policymakers as they weigh whether and how to address rising
costs at BOP. The USSC estimates on the inmate population are reported
in the full report.
For options 1, 2, 5, 6, 7 and 8, the USSC prepared a written analysis
describing its estimate of the impact of changes specified by the option.
For options 3 and 4, the USSC provided tables showing estimated
impact. A description of the option analyzed by the USSC and further
information relevant to specific analyses are included with each analysis
or table below.

Page 71

GAO-14-821 BOP Cost Efficiencies

1. USSC analysis of option 1: Reduce the length of mandatory
minimum sentences for incarcerated drug offenders.
Under this option, the sentences of certain incarcerated drug offenders
convicted of an offense carrying a mandatory minimum sentence would be
reduced by an average of about 44 percent—for example, from 10 years to
5 years and from 5 years to 2 years. Incarcerated offenders who have
completed their recalculated sentence (i.e., if the time served exceeded the
new sentence length) would be released.

Page 72

GAO-14-821 BOP Cost Efficiencies

UNITED STATES SENTENCING COMMISSION
ONE COLUMBUS CIRCLE, N.E.
SUITE 2-500, SOUTH LOBBY
WASHINGTON, D.C. 20002-8002
(202) 502-4500
FAX (202) 502-4699

April 11,2014

Jonathan R. Tumin
Senior Analyst
U.S. Government Accountability Office
441 G St., NW
Washington, DC 20548
Dear Mr. Tumin:
The Government Accountability Office has asked the U.S. Sentencing
Commission to assist it in connection with its engagement (number 441173). We
understand that through this engagement the GAO is examining ways in which the
operating costs of the Federal Bureau of Prisons (BoP) can be reduced. As we have
previously discussed with you and your colleagues, the Commission has agreed to
analyze on GAO's behalf several policy options which could reduce the overall size of
the prison population that the BoP administers. Specifically, our analyses of these policy
options will consider the effect on the federal prison population if the policies were
implemented in such a way that the options would apply to inmates currently incarcerated
in the BoP system.
In this memorandum, we provide to you another of these estimates. Here we
provide our estimate of the impact on the federal prison population if the current statutory
mandatory minimum penalties for drug offenses of 5, 10 and 20 years were reduced to 2,
5, and 10 years, respectively, and the courts were authorized to resentence all federal
offenders convicted of an drug offense carrying one of these mandatory minimum
penalties under a sentencing framework that incorporated the new minimum penalties.
As we discuss more fully below, we estimate that 78,106 offenders who are
incarcerated and would still be in prison as of October 1, 2014 would be eligible for a
sentence reduction. The average reduction in sentence for these offenders would be 57
months (44.2 percent). The estimated total savings to the BoP from such a policy change
applied retroactively would be 370,985 bed years.

Page 73

1

GAO-14-821 BOP Cost Efficiencies

I.

The Policy Option We Considered

One of the policy options that we have agreed to examine on behalf of the GAO is
the impact of a policy change in current law whereby the current statutory mandatory
minimum penalties for drug offenses of 5, 10 and 20 yeru:s were reduced t 2 5 and 10
years, respectively (the "Policy Option").! As part of o lIT anal ysis we have assumed that
the courts would be authorized (most likely by statute2) to resentence aU federal offenders
convicted of a drug offense carrying one of these mandatory minimum penalties under a
sentencing framework that incorporated the new minimum penalties.

II.

What We Found

Applying the Policy Option to offenders in the custody of the BoP, we estimate
that there are 78,106 offenders who would be incarcerated on October 1,2014 who would
be eligible for a reduction in their current sentence if the Policy Option were to be
retroactively applied. Ifthe courts were to grant each of these offenders the full reduction
in their sentence we estimate that the total savings to the BoP would be 370,985 bed
years.3
Table 1 provides information on some ofthe demographic characteristic ofthe
offenders eligible for a sentence reduction pursuant to the Policy Option. Table 2
provides an estimate of the number of offenders who would be released each year if the
Policy Option was applied retroactively compared to the number of offenders who will be
released each year pursuant to their current sentence. As can be seen on this table, if the
Policy Option were fully retroactive on October 1,2014, more than 24,600 offenders
would be eligible for immediate release. Within the first year of the effective date of the
Policy Option, another 8,221 offenders would be eligible for release above the number
who will be released under their current sentence. Table 3 provides information on
selected sentencing guideline characteristics and the criminal history category of these
offenders.
F or the 78,102 offenders who would be eligible to receive a sentence reduction
under the Policy Option, the current average sentence is 129 months. If the courts (or the
President) were to grant the full reduction possible in each case, the projected new
average sentence for these offenders would be 72 months, a reduction of 57 months (or
44.2 percent). Based on this reduction, the estimated total savings to the BoP from the
retroactive application of the Policy Option would be 37,400 bed years. This savings

I See, e.g., 21 U.S.C. §§ 841 (b), 960(b) (2006) (establishing five-, ten-, and twenty-year minimum penalties
in certain drug cases).

2 Another approach would be for the President to reduce the sentence for all offenders eligible for such a
change using his clemency powers. See U.S. CONST., art. II, § 2, cl. 2.
3 A "bed year" is the cost to the BoP of incarcerating one inmate for one year. For example, one inmate
who serves five years of imprisonment accounts for five bed years.

Page 74

2

GAO-14-821 BOP Cost Efficiencies

would not be realized in any single year but is the cumulative savings realized over
several years.

III.

How We Conducted This Analvsis
A.

Methodology

The methodology for this analysis is based on the Commission's Prison Impact
Model, which has been in use in some form since the guidelines were first developed.
This model is used to estimate the impact of proposed statutory and guideline
amendments on newly sentenced offenders and to project the future impact those
amendments will have on bed space in the BoP. For this analysis, those offenders who
appear to be eligible to receive a reduced sentence were hypothetically "resentenced"
with the computer program as ifthe proposed policy change had been in effect in the year
in which they were sentenced. A new release date for each offender also was calculated
in order to determine when the offender would be eligible for release if he or she were
provided the full reduction in sentence provided by the proposed policy change.
B.

The Offender Population We Studied

The Bureau of Prisons provided the Commission with a datafile of inmates who
were in the custody of the BoP on Jan 25,2014. That file contained approximately
189,000 offenders. Approximately 184,000 of these offenders were sentenced between
fiscal year 1992 and fiscal 2013. 4 USSC staff was able to match 171,765 of these
offenders to Commission records. Of these, 138,894 were estimated to remain
incarcerated on October 1,2014.
In order to approximate the group of offenders who will be sentenced in fiscal
year 2014, the Commission used the FY2013 datafile and moved all sentence dates
forward by one year. Staff then determined which of these offenders would be
incarcerated on October 1,2014. This process added another 40,178 offenders into the
analysis. Between the two groups, the Commission's analysis included data on 179,072
offenders.
C.

Our Assumptions

In performing our analysis, we have been required to make some assumptions (set
forth below) concerning the decisions that Congress and the courts (or the President)
would make in determining whether, and to what extent, to reduce the sentences of
offenders eligible to receive a modification of sentence pursuant to the Policy Option
discussed above. These assumptions may not hold in every case.
The analysis was limited to data from fiscal year 1992 through September 30, 2013 (fiscal years 1992
through 2013) because the Commission's data collection efforts prior to fiscal year 1992 were not as
complete as in later years. For example, the Commission did not collect information on the type of drug
involved in drug offenses prior to fiscal year 1992.

4

Page 75

3

GAO-14-821 BOP Cost Efficiencies

1.

Procedural assumptions

Specifically, we have assumed that Congress would authorize the courts (or the
President would issue an order) to reduce the sentences of offenders incarcerated on or
after a specific date (discussed below) in a manner consistent with the Guidelines Manual
and bound only by any mandatory minimum penalties that might apply in current law or
as modified through the Policy Option being discussed. We also assumed that the
Commission would make no further changes to the Guidelines Manual in light of the
Policy Option, so that the provisions of the 2013 Guidelines Manual would apply in these
cases. Finally, we assumed that the effective date of any such Policy Option would be
October 1,2014. Therefore, only offenders incarcerated as of that date would be eligible
to seek a reduction in sentence pursuant to the Policy Option discussed in the
memorandum.
2.

Substantive assumptions

The policy we analyzed would have the effect of lowering the current statutory
mandatory minimum penalties in drug cases from 5, 10, and 20 years to 2,5, and 10
years respectively. In fiscal year 2010, approximately 66 percent of all drug offenders
were convicted of an offense carrying a mandatory minimum penalty.s Of those,
approximately 46 percent remain subject to the penalty at the time of sentencing. 6 Some
offenders receive relief from the effect of any such penalty by providing substantial
assistance to the government in the investigation or prosecution of another offender. 7
Others receive relief from such a penalty through the statutory "safety valve" provision
which authorizes courts to sentence drug offenders without regard to any mandatory
minimum penalty provided the offender meets several criteria desirned to identify lowlevel, non-violent offenders with little or no prior criminal history. Some offenders are
eligible for both provisions.
In our analysis, we provide a full reduction for all offenders who remained subject
to a mandatory minimum penalty at the time of sentencing. That is, we reduce sentences
by 60 percent for all offenders convicted of a 5-year mandatory minimum who remained
subject to that mandatory minimum at sentencing. Similarly, we reduce sentences by 50
percent for all offenders convicted of a 1O-year or a 20-year mandatory minimum penalty
and who remained subject to that penalty at sentencing.

See U.S. SENTENCING COMMISSION, MANDATORY MINIMUM PENALTIES IN THE FEDERAL CRIMINAL
JUSTICE SYSTEM 154 (2011).

5

6

I d.

7

See 18 U.S.C. § 3553(e); USSG §5Kl.1.

8

See 18 U.S.c. § 3553(f); USSG §5K3.1.

Page 76

4

GAO-14-821 BOP Cost Efficiencies

Drug offenders who receive relief from a mandatory minimum penalty by
providing substantial assistance to the government have sentences that are, in general, 46
percent below the arplicable guideline minimum sentence,9 which often is the mandatory
minimum penalty. 1 Drug offenders who receive relief from a mandatory minimum
penalty through the statutory safety valve provision have sentences that are, in general,
39 percent below the applicable guideline minimum sentence. 11 The extent of these
reductions fluctuates somewhat each year. Because these offenders benefitted from
substantial reductions in their sentence at the time of the original sentence, we have
assumed that any benefit to them from the Policy Option would not be as great as it
would be for those offenders who remained subject to a mandatory minimum penalty at
sentencing. Therefore, for this analysis, we made additional assumptions for these
offenders, as follows:
-- offenders who were convicted of a drug offense carrying a 5-year mandatory
minimum penalty, but who were not subject to any mandatory penalty at sentencing due
to relief through substantial assistance or the safety valve would receive a further
reduction of 30 percent under the Policy Option (i.e., 50% of the 60% reduction that
offenders received if they remained subject to that mandatory minimum penalty at
sentencing);
-- offenders who were convicted of a drug offense carrying a 10-year or 20-year
mandatory minimum penalty, but who were not subject to any mandatory penalty at
sentencing due to relief through substantial assistance or the safety valve would receive a
further reduction of25 percent under the Policy Option (i.e., 50% of the 50% reduction
that offenders received if they remained subject to either ofthose mandatory minimum
penalties at sentencing).
This analysis does not assume any change in sentence for offenders who were not
convicted of a drug offense carrying a mandatory minimum penalty. It is possible, of
course, that Congress (of the President) might authorize a reduction in sentence for drug
offenders currently incarcerated but who were not convicted of an offense carrying a
mandatory minimum penalty. Any such reduction would likely be tied to a change in the
Guidelines Manual that incorporated the Policy Option. We are unable to model any
change to the sentencing guidelines from a statutory enactment of the Policy Option, and
so have excluded from this analysis any consideration of a change in sentence for drug
offenders who were not convicted of an offense carrying a mandatory minimum penalty.
Additional assumptions we have made for this analysis are that the sentence for
each offender would be reduced based on the maximum good conduct credit allowed by

9

See 2013 Sourcebook of Federal Sentencing Statistics, Table 30 (2014).

10 See USSG §5G 1.1 (which provides that the minimum guideline sentence may not be lower than any
mandatory minimum penalty).

II

See 2013 Sourcebook of Federal Sentencing Statistics, Table 30A (2014).

Page 77

5

GAO-14-821 BOP Cost Efficiencies

the Bureau of Prisons, and offenders would serve the lesser of the newly calculated
sentence or their life expectancies.
If you have any questions about our analysis, please do not hesitate to contact the
Commission. We continue to work to analyze the other policy options that we discussed
with you and your colleagues. We will provide our analyses of these options as they are
completed.
Sincerely

Noah D. Bookbinder
Director
Office of Legislative and Public Affairs

Director
Office of Research and Data

Attachments

Page 78

6

GAO-14-821 BOP Cost Efficiencies

Table 1
Demographic Characteristics of Drug Offenders
COllvicted of a Drug Offense Carrying a 5, 10, or 20-Year Mandatory Minimum Penalty
Eligible for 2, 5, and lO-Year Mandatory Minimums

DEMOGRAPIDCS
Rnec/Ethnicity

While
Black
Hispanic
Other
Total

14,773
27,512
34,096
1,590
77,971

19.0%
35.3%
43.7%
2.0%
100.0%

U.S. Citizen
Non-Citizen
Total

57,783
20,242
78,025

74.1%
25.9%
100.0%

Male
Female
Total

72,257
5,845
78,102

92.5%
7.5%
100.0%

Citizenship

Gender

Average Age
35
(as of October 1, 2014)
'The analysis involves a total of78,106 cases in which the offender was identified as eligible to seek a sentence reduction if
the drug mandatory minimum changes were made retroactive. Cases missing infonnation for any
specific analysis are excluded irom that analysis.
Total percentages may not add to exactly 100% due to rounding.
SOURCE: U.S. Sentencing Couunission, 1992 - 2013 Datafiles, USSCFY92 - USSCFY13 and projected 2014 Datafile.

Page 79

GAO-14-821 BOP Cost Efficiencies

Table 2
Estimated Projected Year of Release for Drug Offenders
Potentially Eligible for Mandatory Minimum Sentence Reduction

If Policy Option
Applied Retroactivly

CURRENT
SENTENCE

N

N

Release Date
immediate

24,601

within 1 yr

12,515

4,294

within 2 yr

12,191

9,223

within 3 yr

8,546

10,166

within 4 yr

6,404

9,256

within 5 yr

4,363

7,852

within 6 yr +

9,482

37,311

'Of the 78, 106 offenders who appear to be eligible for relief if the Polcy Option was applied
retroactivly, Commission records contained sufficient infonnatioll to perfonn
perfonTI this analysis for 78, I 02 offenders.
SOURCE: U.S. Sentencing Commission, 1992 - 2012 Datafiles, USSCFY92 - USSCFY13 and projectel

Page 80

GAO-14-821 BOP Cost Efficiencies

Table 3
Guideline Sentencing Characteristics and Criminal History of
Drug Offenders Potentially Eligible for Mandatory Minimum Sentence Reduction

CHARACTERISTICS
Ayerage Base Offense Level

32

Weapon Specific Offense Characte.-istic
Firearms Mandatory Minimum Applied
Safety Valve §5C1.2
Aggravating Role §3B1.1
Mitigating Role §3B1.2
Obstruction Adjustment §3Cl.l
Career Offender Status §4B1.1
Criminal History Category
I
II
III
IV
V
VI
Total
I

14,729
7,433
14,042
11,549
6,675
3,660
10,982

18.9%
9.5%
18.1%
14.8%
8.6%
4.7%
14.1%

28,925
9,652
12,995
6,853
4,079
15,602
78,106

37.0%
12.4%
16.6%
8.8%
5.2%
20.0%
100%

.he analysis involves a total of78,106 cases in which the offender was identified as eligible to seek a sentence reduction if

the drug mandatory minimum changes were made retroactive. Cases missing infonl1ation tor any
specific analysis are excluded from that analysis.
~The average base offense level was excluded from this analysis due to missing guideline relevant statutory information
lrom the new sentence.

Total percentages may not add to exactly 100% due to rounding.
SOURCE: U.S. Sentencing Conunission, 1992 - 2012 Datafiles, USSCFY92 - USSCFY13 and projected 2014 Datafile.

Page 81

GAO-14-821 BOP Cost Efficiencies

2. USSC analysis of option 2: Retroactively apply the Fair Sentencing
Act (FSA) of 2010 reductions in drug mandatory minimum penalties to
offenders currently incarcerated for crack cocaine-related offenses.
This option would retroactively apply to eligible incarcerated offenders the
reductions in mandatory minimum penalties for crack cocaine-related
offenses that are provisions of the Fair Sentencing Act of 2010, Pub. L. No.
111-220, 124 Stat. 2372, which did not apply to offenders already
sentenced under the original penalty guidelines.

Page 82

GAO-14-821 BOP Cost Efficiencies

UNITED STATES SENTENCING COMMISSION
ONE COLUMBUS CIRCLE, N.E.
SUITE 2-500, SOUTH LOBBY
WASHINGTON, D.C. 20002-8002
(202) 502-4500
FAX (202) 502-4699

February 20, 2014

Jonathan R. Tumin
Senior Analyst
U.S. Government Accountability Office
441 G St., NW
Washington, DC 20548
Dear Mr. Tumin:
The Government Accountability Office has asked the U.S. Sentencing
Commission to assist it in connection with its engagement (number 441173). We
understand that through this engagement the GAO is examining ways in which the
operating costs of the Federal Bureau of Prisons (BoP) can be reduced. As we have
previously discussed with you and your colleagues, the Commission has agreed to
provide to the GAO the results of our analyses of several policy options which could
reduce the overall size of the prison population that the BoP administers. Specifically,
our analyses of these policy options will consider the effect on the federal prison
population if the policies were implemented in such a way that the options would apply
to inmates currently incarcerated in the BoP system.
In this memorandum, we provide to you the first of these estimates. Here we
provide our estimate of the impact on the federal prison population if the provisions of
the Fair Sentencing Act of2010 1 (the "FSA") were authorized to apply to persons
currently incarcerated and sentenced before the enactment ofthat law. As we discuss
more fully below, we estimate that 8,829 offenders who were incarcerated on January 1,
2014 would be eligible for a sentence reduction if the FSA were applied retroactively
beginning on that date. If the courts were to reduce the sentence of each eligible offender
in accordance with the FSA and the current Guidelines Manual, the average reduction in
sentence for these offenders would be 53 months (28.8 percent). The estimated total
savings to the BoP would be 37,400 bed years.

I

Pub. L. No. 111-220, 124 Stat. 2372 (2010) [hereinafter FSA] .

Page 83

1

GAO-14-821 BOP Cost Efficiencies

I.

Statutory Changes in the FSA and the Conforming Guideline Amendment
A.

Statutory Changes in the FSA

The FSA changed the quantities of crack cocaine that trigger the five- and tenyear statutory mandatory minimum penalties. As a consequence, first-time trafficking
offenses involving less than 28 grams of crack cocaine are subject to a statutory penalty
range of zero to 20 years of imprisonment. First-time trafficking offenses involving
between 28 and 280 grams of crack cocaine are subject to a statutory penalty range of
five to 40 years of imprisonment. 2 A first-time trafficking offense involving 280 or more
grams of crack cocaine is subject to a statutory penalty range of 10 years to life
imprisonment. 3 The FSA did not make any of the statutory changes retroactive.
Therefore, offenders sentenced before the effecti ve date of the FSA4 continue to be
subject to the mandatory minimum penalties in effect on the date of sentencing. 5
B.

Temporary and Permanent Amendments to the Guidelines Manual in
Response to the FSA

On October 15,2010, the Commission promulgated a temporary, emergency
amendment to the guidelines that implemented the emergency directive in section 8 of
6
the FSA. On April 6, 2011, the Commission re-promulgated the temporary amendment
as a permanent multipart amendment, which became effective on November 1,2011. 7
(In this memorandum, the several modifications to the drug guidelines promulgated in
response to the FSA are referred to collectively as the "FSA Guideline Amendment.")
Part A of the FSA Guideline Amendment modified the base offense level for
various quantities of crack cocaine assigned by the Drug Quantity Table. 8 Part B of the
2 The new five year mandatory minimum threshold quantity of28 grams corresponds to approximately one
ounce, which has been considered to be a threshold quantity for purposes of classifying the function of
certain federal crack cocaine offenders. See UNITED STATES SENTENCING COMMISSION, REpORT TO THE
CONGRESS: COCAINE AND FEDERAL SENTENCING POLICY 18 (MAY 2007) [hereinafter 2007 Commission
Report].
3 Because it now takes approximately 18 times more powder cocaine than crack cocaine to trigger the same
statutory mandatory minimum penalties, some may refer to this penalty structure as an "18-to-1 drug
quantity ratio."

4

The FSA was enacted into law on August 3, 2010.

5 See Dorsey v. United States, 567 U.S. _ , 132 S. Ct. 2321 (2012) (FSA's lower mandatory minimum
penalties applied to offenders sentenced on and after August 3,2010).
6

USSG, 2011 Supp. to App. C, Amendment 748 (effective Nov. 1,2010).

7

USSG, 2011 Supp. to App. C, Amendment 750 (effective Nov. 1,2011).

Offenses involving 28 grams of crack cocaine were assigned a base offense level of 26, and offenses
involving 280 grams of crack cocaine were assigned a base offense level of 32. This approach is consistent
with how the guidelines incorporate the statutory mandatory minimum penalties for all other drug offenses.
8

Page 84

2

GAO-14-821 BOP Cost Efficiencies

FSA Guideline Amendment added both new mitigating and aggravating provisions to
USSG §2D 1.1 for offenses involving drugs, regardless of drug type. Part C of the FSA
Guideline Amendment eliminated a provision in the guidelines under which an offender
who possessed more than 5 grams of crack cocaine was sentenced under the drug
trafficking guideline.
C.

Retroactive Application of the FSA Guideline Amendment

The Commission voted to authorize courts to retroactively arply Parts A and C of
the FSA Guideline Amendment to offenders currently incarcerated. Pursuant to 18
U.S.C. § 3582(c)(2), courts were then authorized to reduce the sentences of those
offenders consistent with the amended sentencing guidelines. To date, the courts have
reduced the sentences of more than 7,500 offenders pursuant to retroactive application of
the FSA Guideline Amendment. 10 In more than 5,000 other cases, the courts have denied
relief to offenders seeking a reduction in sentence under the FSA Guideline Amendment.
The courts found that the offender was not eligible for any sentence reduction in more
than 3,700 of those cases. For example, in 1,458 cases, the court found that the offender
was ineligible because the applicable mandatory minimum penalty controlled the
sentence and prevented the court from reducing the sentence.
In more than 1,045 other cases, the offender was found ineligible because the
Career Offender guideline (USSG §4B1.1), and not the drug trafficking guideline (USSG
§2D 1.1), determined the guideline range at sentencing. As discussed above, the
mandatory minimum penalties applicable to offenders sentenced prior to the FSA were
unchanged by the act. The Career Offender guideline is linked to statutory maximum
penalties which, in drug cases, are also linked to statutory mandatory minimum penalties.
Because neither the FSA nor the FSA Guideline Amendments had the effect of lowering
the Career Offender guideline, the provisions of 18 U.S.C. § 3582(c)(2) allowing courts
to reduce the sentences of incarcerated offenders when their sentencing guideline is
lowered do not apply to offenders whose guideline range is determined under that
proVISIOn.
Finally, in some additional cases, the offender received only a partial reduction in
sentence under the FSA Guideline Amendment. This was because a mandatory
minimum penalty prevented the court from lowering the sentence to a point within the
amended guideline range.

The base offense level for some quantities of crack cocaine in the Drug Quantity Table did not change as a
result of the FSA Guideline Amendment. Also, offenses involving quantities of less than 500 mg of crack
cocaine were unaffected by the amendment and remained assigned to BOL 12.
9

USSG, 2011 Supp. to App. C, Amendment 759 (effective Nov. 1,2011).

10 See generally USSG, Preliminary Crack Retroactivity Data Report, Fair Sentencing Act (January 2014)
(available at
http://www.ussc.gov/Research_and_Statistics/Federal_Sentencing_Statistics/FSA_ Amendment/index.cfm)

Page 85

3

GAO-14-821 BOP Cost Efficiencies

Page 86

GAO-14-821 BOP Cost Efficiencies

B.

Our Assumptions

In performing our analysis, we have been required to make some assumptions (set
forth below) concerning the decisions that Congress and the courts would make in
determining whether, and to what extent, to reduce the sentences of offenders eligible to
receive a modification of sentence pursuant to the FSA if it were retroactively applied.
These assumptions may not hold in every case.
Specifically, we have assumed that Congress would authorize the courts by statute
to reduce the sentences of offenders in accordance with the Guidelines Manual in effect
on a specific date and bound only by the mandatory minimum penalties as they were
amended by the FSA. We also assumed that the Commission would make no further
changes to the Guidelines Manual regarding crack cocaine offenses in response to
Congress' action, so that the provisions ofthe 2012 Guidelines Manual would apply in
these cases. 12 Finally, we assumed that the effective date of any such statutory change
would be no earlier than January 1,2014.
Additional assumptions we have made for this analysis are as follows:
(1) offenders sentenced within the applicable guideline range when they were first
sentenced would be sentenced at the same point in the range under the Guidelines
Manual in effect on January 1,2014;
(2) offenders sentenced outside the applicable guideline range when they were
first sentenced would be sentenced to a new position outside the guideline range under
the Guidelines Manual in effect on January 1,2014 that is the same proportional distance
above or below the amended guideline range as their original sentence was from the
guideline range in effect at the original sentencing;
(3) offenders for whom the new estimated sentence is below the applicable
mandatory minimum (all cases in this analysis have five grams triggering a five-year
mandatory minimum or 50 grams triggering a 10-year mandatory minimum), and where
no safety valve or substantial assistance reduction was applied when the offender was
originally sentenced, would be sentenced at the new applicable mandatory minimum;
(4) offenders classified as Career Offenders would be sentenced pursuant to the
Career Offender provision of the guidelines in accordance with the statutory maximums
as amended by the FSA.
(5) the "mitigating role cap" on the base offense level of the guidelines would be
applied, if appropriate, based upon the new base offense level;
(6) offenders originally receiving relief from a mandatory minimum penalty by
operation of the safety valve provision would continue to receive relief but, if the
12 While the 2012 Guidelines Manual was used for this analysis, the provisions of the 2013 Guidelines
Manual that pertain to offenses under USSG §2D 1.1 are unchanged.

Page 87

5

GAO-14-821 BOP Cost Efficiencies

applicable statutory minimum is at least five years, the offense level determined after
applying Chapters Two (Offense Conduct) and Three (Adjustments) of the guidelines
would not be less than level 17 (pursuant to USSG §5CI.2(b»;
(7) for offenders with an original combined offense level (after application of
Chapters Two and Three) of level 16 or greater but having a new combined offense level
below level 16, the applicable reduction for Acceptance of Responsibility would be
reduced from three levels to two levels in accordance with that guideline provision;
(8) the sentence for each offender would be reduced based on the maximum good
conduct credit allowed by the Bureau of Prisons; and
(9) offenders would serve the lesser ofthe newly calculated sentence or their life
expectancies.
C.

Limitation as to Application of New Mitigating Provisions of the FSA

As discussed above, Part B of the FSA Guideline Amendment added both new
mitigating and aggravating provisions to USSG §2DI.I for offenses involving drugs,
regardless of drug type. The Commission did not vote to allow those provisions to be
applied retroactively. We are unable to assess the extent to which the courts might apply
these new mitigating provisions were the FSA to be applied retroactively as those
provisions involve the consideration of facts that are not readily available in the
sentencing documents provided to the Commission in all cases. To the extent that the
courts might consider these provisions in resentencing offenders when applying the FSA
retroactively, the sentences for some offenders could be reduced below the point
determined in our estimate, which would accelerate the release of those offenders from
incarceration.
D.

The Offender Population We Studied

In conducting this analysis, the Commission requested and received a datafile
from the BoP which identified a total of 135,536 offenders in the BoP's custody at that
time who the BoP estimated would still be incarcerated on January 1,2014 (the proposed
retroactivity date we used for this analysis.). These offenders were those sentenced
between FYI992 and FY2012 and whose BoP record could be matched with USSC
records. 13 Of those, 19,154 offenders were identified as 1) having been sentenced under
the drug trafficking guideline,14 2) having crack cocaine as one of the drug types involved
13 The analysis was limited to data from fiscal year 1992 through September 30, 2010 (fiscal years 1992
through 2010) because the Commission did not collect information on the type of drug involved in drug
offenses prior to fiscal year 1992.
14 For some of these offenders, the guideline range that applied at sentencing was derived from either the
Career Offender (USSG §4B 1.1) or Armed Career Criminal (USSG §4B 1.4) provision. However, for the
purpose of this analysis these offenders are considered to have been sentenced under the drug trafficking
guideline as the court is required to calculate a guideline range for them under USSG §2D 1.1 as well.

Page 88

6

GAO-14-821 BOP Cost Efficiencies

in the offense, and 3) having sufficient infonnation for analysis by the Commission's
prison and sentencing impact model.
The Commission's retroactivity analysis determined that of these 19,154
offenders, 10,325 would not experience a change in their sentence if the FSA were fully
retroactive. These offenders were excluded from the analysis. The majority of the
offenders (n = 9,741) were excluded because they would not experience the proposed
statutory change - either because no statutory mandatory minimum penalty applied in
their cases or the offender already was sentenced below the FSA mandatory minimum
thresholds. IS Another 362 offenders were sentenced under the Career Offender provision
and, despite a change to their drug statutory minimum penalty, their guideline range
under the Career Offender provision did not change. In another 191 cases, the current
sentence is greater than both the guideline minimum and the FSA statutory minimum,
therefore no change in their sentence would result. A small number of offenders (n=30)
were excluded because their guideline range was detennined under the Armed Career
Criminal provision, which would be unaffected by the retroactive application of the FSA.
Lastly, one offender had a base offense level of38, the top of the Drug Quantity Table,
and the quantity of drugs involved in that case continued to place the offender at that
level, so therefore no change in his or her guideline range would occur.
After these offenders were excluded, there remained 8,829 offenders who were
projected to be incarcerated on January 1,2014 and who appeared eligible to seek a
reduced sentence if the provisions of the FSA were to be made retroactive as described
above.
If you have any questions about our analysis please do not hesitate to contact the
Commission. We continue to work to analyze the other policy options that we discussed
with you and your colleagues. We will provide our analyses ofthese options as they are
completed.
Sincerely,

Noah D. Bookbinder
Director
Office of Legislative and Public Affairs

Director
Office of Research and Data

Attachment

15 Because many of the offenders currently incarcerated in the BoP for crack offenses were sentenced after
August 3, 2010, the mandatory minimum penalty provisions applied by the courts in most of those cases
were the lower penalties provided by the FSA. Some courts did continue to apply the pre-FSA minimum
penalties to offenders sentenced after August 3,2010 but who committed their crime prior to that date.
However, the Supreme Court's opinion in Dorsey clarified that this practice was in error.

Page 89

7

GAO-14-821 BOP Cost Efficiencies

Table GAO-l
Estimated Projected Year of Release for Crack Cocaine Offenders
Potentially Eligible for FSA Retroactivityl

Release Date

IFFSA
RETROACTIVITY
GRANTED
N

CURRENT
SENTENCE
N

within 1 yr

3,695

682

within 2 yr

999

847

within 3 yr

806

951

within 4 yr

682

996

within 5 yr

503

958

within 6 YI' +

1,783

4,034

IOfthe 8,829 offenders who appear to be eligible for relief if the Fair Sentencing Act
of 201 0 were retroactive, Conunission records contained sufficient information to
perform this analysis for 8,468 offenders.
SOURCE: U.S. Sentencing Commission, 1992 - 2012 Datafiles, USSCFY92 - USSCFYI2.

Page 90

GAO-14-821 BOP Cost Efficiencies

3. USSC table with data for impact of option 3: Instruct U.S. Attorneys
to decline to charge certain defendants in certain types of drug cases
in such a manner as to trigger mandatory minimum sentences.
This option has been implemented as part of the August 2013 direction
issued by the Attorney General instructing prosecutors in cases involving
the applicability of drug law mandatory minimum sentences based on drug
type and quantity to decline to charge the quantity necessary to trigger a
mandatory minimum sentence if the defendant meets particular criteria.
For option 3, the USSC provided a table composed of six subtables that
show the USSC’s analysis of the estimated impact of the changes
discussed in option 3. According to the USSC, the total number of
offenders estimated to be affected annually by option 3 can be found in line
1, column 2, of the first table (i.e., 620). For the first 5 years’ estimated
impact, we multiplied 620 by 5. According to the USSC, the bed years
saved over the first 5 years of implementation can be found by adding the
numbers shown in the first line of the third table, which total 788.

Page 91

GAO-14-821 BOP Cost Efficiencies

Comparison of Prison Impact Analysis – Safety Valve Expansion with Statutory Trumps
FY2012 Cases
Change in Sentences Imposed
Total
Cases1

Number
Affected2

Percent
Affected

Current Average
Sentence for Affected
Cases

New Average
Sentence for
Affected Cases

Number of
Months
Change

Percent
Change

2 pts

841

620

73.7

46

37

-9

-19.6

2-3 pts

2,242

1,636

73.0

49

40

-9

-18.4

Change in Incarceration Levels
Change in total years of incarceration for offenders sentenced in one fiscal year3
1st Year

2nd Year

3rd Year

4th Year

5th Year

10th Year

15th Year

Total4

2 pts

-51

-46

-64

-47

-64

-6

-1

-372

2-3 pts

-115

-119

-167

-142

-181

-34

-3

-1,065

Change in Total BOP Population in Future Years5

1Total

1st Year After
Effective Date

2nd Year After
Effective Date

3rd Year After
Effective Date

4th Year After
Effective Date

5th Year After
Effective Date

2 pts

-51

-97

-161

-208

-271

2-3 pts

-115

-234

-401

-543

-724

Cases are those with a particular sentencing factor being analyzed. 2Affected Cases are those in which the sentence is estimated to change as a result of the sentencing factor being analyzed. Not all cases
will change as a result of the application of the
sentencing
factor being analyzed. 3This table represents the number of prison beds saved each year by a cohort of offenders sentenced in a single year. 4This is the
Page
94
BOP
Efficiencies
total number of prison beds that will be saved when all offenders who were sentenced in the same year are ultimately released from prison. 5This is the annual number of GAO-14-821
prison beds saved
as Cost
ongoing
cohorts of
offenders enter the Bureau of Prisons who have been sentenced under the changed guideline.
SOURCE: U.S. Sentencing Commission Prison and Sentencing Impact Assessment Model, FY2012 datafile.

Comparison of Prison Impact Analysis – Safety Valve Expansion without Statutory Trumps
FY2012 Cases
Change in Sentences Imposed
Total
Cases1

Number
Affected2

Percent
Affected

Current Average
Sentence for Affected
Cases

New Average
Sentence for
Affected Cases

Number of
Months
Change

Percent
Change

2 pts

841

787

93.6

53

42

-11

-20.8

2-3 pts

2,242

2,090

93.2

56

45

-11

-19.6

Change in Incarceration Levels
Change in total years of incarceration for offenders sentenced in one fiscal year3
1st Year

2nd Year

3rd Year

4th Year

5th Year

10th Year

15th Year

Total4

2 pts

-52

-49

-67

-163

-61

-8

-3

-617

2-3 pts

-116

-126

-173

-447

-170

-36

-12

-1,731

Change in Total BOP Population in Future Years5
1st Year After
Effective Date

2nd Year After
Effective Date

3rd Year After
Effective Date

4th Year After
Effective Date

5th Year After
Effective Date

2 pts

-52

-101

-168

-330

-392

2-3 pts

-116

-242

-415

-862

-1,032

1Total Cases are those with a particular sentencing factor being analyzed. 2Affected Cases are those in which the sentence is estimated to change as a result of the sentencing factor being analyzed. Not all cases
will change as a result of the application of the sentencing factor being analyzed. 3This table represents the number of prison beds saved each year by a cohort of offenders sentenced in a single year. 4This is the
total number of prison beds that will be saved when all offenders who were sentenced in the same year are ultimately released from prison. 5This is the annual numberGAO-14-821
of prison bedsBOP
savedCost
as ongoing
cohorts of
Efficiencies
offenders enter the Bureau of Prisons who have been sentenced under the changed guideline.
SOURCE: U.S. Sentencing Commission Prison and Sentencing Impact Assessment Model, FY2012 datafile. Page 93

4. USSC table with data for impact of option 4: Amend USSC
sentencing guidelines for future drug offenders, reducing them two
levels below current guidelines.
This option reduces the potential sentences for all future drug offenders by
two levels below current guidelines, resulting in a roughly 18 percent
reduction of sentence length below current lengths. It was approved by the
USSC and forwarded to Congress on April 30, 2014; it will apply to future
eligible drug offenders effective November 1, 2014, in connection with
option 5 (see next option) unless amended or rejected by Congress.
For option 4, the USSC provided a table composed of 2 subtables; the
overall table is labeled “Table 5” by the USSC and shows the USSC
estimate of the changes discussed in option 4. According to the USSC, the
number of offenders estimated to be affected annually by the option
(17,457) can be found in line 1 of the first table. Some other data in the two
subtables were updated as a result of USSC actions that made option 4
retroactive (i.e., option 5), but which also delayed implementation of
retroactivity until November 1, 2015, unless amended or rejected by
Congress.

Page 94

GAO-14-821 BOP Cost Efficiencies

Table 5
Estimated Effect on Sentencing and Incarceration
of Decreasing the Drug Quantity Table by Two Levels
FY2012 Cases

Change in Sentences Imposed
Total
Cases1

Number
Affected2

Percent
Affected

Current Average
Sentence for
Affected Cases

17,457

69.9

62

24,968

New Average
Sentence for
Affected Cases

Number of
Months
Change

Percent
Change

51

-11

-17.7

Change in Sentences Served
Change in years of incarceration served for offenders sentenced in a single fiscal year 3

1st Year

2nd Year

3rd Year

4th Year

5th Year

10th Year

15th Year

Total4

-894

-1,083

-1,667

-1,281

-1,625

-634

-256

-13,938

Change in total BOP Population in Future Years5

1st Year After
Effective Date

2nd Year After
Effective Date

3rd Year After
Effective Date

4th Year After
Effective Date

5th Year After
Effective Date

-894

-1,977

-3,644

-4,925

-6,550

1Total

Cases are those with a particular sentencing factor being analyzed.
Cases are those in which the sentence is estimated to change as a result of the sentencing factor being analyzed. Not all cases will change as a result of the application of the
sentencing factor being analyzed.
3This table represents the number of prison beds saved each year by a cohort of offenders sentenced in a single year.
4This is the total number of prison beds that will be saved when all offenders who were sentenced in the same year are ultimately released from prison.
GAO-14-821 BOP Cost Efficiencies
5This is the annual number of prison beds saved as ongoing cohorts of offenders enter the Bureau of Prisons who have been sentenced under the changed guideline.
SOURCE: U.S. Sentencing Commission Prison and Sentencing Impact Assessment Model, FY2012 datafile. Page 95
2Affected

5. USSC analysis of option 5: Apply the amendments to sentencing
guidelines in option 4 also to incarcerated drug offenders—that is, to
both new drug offenders and retroactively to incarcerated drug
offenders.
Under this option, the provisions of option 4 apply to both new drug
offenders and also retroactively to eligible incarcerated drug offenders. It
reduces the sentences for all future and currently incarcerated drug
offenders by two levels below current USSC guidelines, resulting in a
roughly 19 percent reduction below current lengths. It was approved by the
USSC in July 2014 as an amendment to option 4; as such, it corresponds
to the April 2014 USSC decision and does not require separate
congressional action. Unless Congress acts to modify or disapprove the
corresponding April 30 amendment, beginning November 1, 2014, eligible
incarcerated offenders can ask courts to reduce their sentences.
Incarcerated offenders whose requests are granted by the courts may not
be released earlier than November 1, 2015.
This analysis (which also includes information relevant to option 4, to which
it is closely related) was posted to the USSC website on July 25, 2014.
(http://www.ussc.gov/research-and-publications/retroactivity-analyses-anddata-reports/retroactivity-analyses-and-data-reports)

Page 96

GAO-14-821 BOP Cost Efficiencies

UNITED STA TES SENTENCING COMMISSION
ONE COLUMBUS CIRCLE, NE
SUITE 2-500, SOUTH LOBBY
WASHINGTON, DC 20002-8002
(202) 502-4500
FAX (202) 502-4699

July 25,2014

MEMORANDUM
TO:

Chair Saris
Commissioners
Kenneth Cohen

FROM:

Office of Research and Data

SUBJECT:

Summary of Key Data Regarding Retroactive Application of the
2014 Drug Guidelines Amendment

I.

INTRODUCTION

On April 30, 2014, the United States Sentencing Commission submitted to Congress an
amendment to the federal sentencing guidelines that would revise the guidelines applicable to
drug trafficking offenses by changing how the base offense levels in the drug quantity tables in
sections 2D 1.1 and 2D 1.11 of the Guidelines Manual! incorporate the statutory mandatory
minimum penalties for drug trafficking offenses (Amendment 782).2 Specifically, the
amendment would reduce by two levels the offense levels assigned to the quantities that trigger
the statutory mandatory minimum penalties, resulting in corresponding guideline ranges that
include the mandatory minimum penalties, and make conforming changes to section 2D2.11. On
July 18, the Commission voted to give retroactive effect to Amendment 782 beginning on the
effective date of the amendment, which will be November 1, 2014, unless Congress acts to
modify or disapprove the amendment.

1 U.S. SENTENCfNG COMMISSION, GUIDELfNES MANUAL §2D 1.1 (Unlawful Manufacturing, Importing, Exporting or
Trafficking (Including Possession with Intent to Commit These Offenses); Attempt or Conspiracy) (2013)
(hereinafter USSG); USSG §2D 1.11 (Unlawfully Distributing, Importing, Exporting or Possessing a Listed
Chemical; Attempt or Conspiracy).

2 References in this memorandum to the "2014 drug guidelines amendment," "the amendment," or any similar
references mean the amendment the Commission submitted to Congress on April 30, 2014, that would modify the
drug quantity tables in USSG §§2Dl.1 and 201.11.

Page 97

GAO-14-821 BOP Cost Efficiencies

As part of the Commission's decision to retroactively apply Amendment 782, it required
that no offender may be released pursuant to the retroactive application of the amendment until
November 1, 2015 or later. This memorandum provides information concerning the effect of
this limitation.
II.

FURTHER ANALYSES OF THE IMPACT OF RETROACTIVE APPLICATION
OF THE 2014 DRUG GUIDELINES AMENDMENT
A.

Summary of Data on the Eligible Offenders

ORD previously estimated that 51,141 offenders sentenced between October 1, 1991 and
October 31,2014,3 would be eligible to seek a reduction in their current sentence if the
Commission were to make the 2014 drug guidelines amendment retroactive. 4 Of this group,
there are 46,376 offenders who would not be released under their current sentence until on or
after November 1,2015, and so could benefit from the Commission's decision to retroactively
apply Amendment 782.
The current average sentence for the 46,376 offenders who could benefit under
retroactive application of the amendment is 133 months. Applying the amendment retroactively
to these offenders the new average sentence would be 108 months. This is a difference of 25
months, which represents an 18.8 percent reduction in the sentence. The number of bed years
saved by this reduction is 79,740 bed years, which would be realized by the Bureau of Prisons
over a period of years, of course.
The most common drug types involved in these cases are methamphetamine (28.8%),
powder cocaine (27.8%), crack cocaine (19.3%), marijuana (11.6%), heroin (7.6%), and other
drugs (5.0%). Attached to this memorandum is a summary of selected offender characteristics
about these offenders as well as information about their criminal history.
B. Summary of Data on the Projected Release Dates of the Eligible Offenders
We estimate that on November 1,2015 there would be 7,953 offenders eligible for
immediate release. Another 8,550 offenders would be released during the year that begins on
November 1, 2015 and ends on October 31, 2016. That is, a total of 16,503 offenders will be
released during the first year in which offenders may be released pursuant to retroactive
application of the amendment. If the Commission had not authorized the amendment to be
applied retroactively, we estimate that 7,609 offenders would still have been released during that
year, as their current terms of incarceration expired. Attached to this memorandum is a table that
provides additional information concerning the release years of these offenders and the districts
in which they were sentenced.

The analysis was limited to data from fiscal year 1992 and later because the Commission's data collection efforts
prior to fiscal year 1992 were not as complete as in later years.

3

4 Memorandum from the Office of Research and Data and the Office of General Counsel to Chair Saris,
Commissioners, and Kenneth Cohen (May 27,2014), available at www.ussc.gov.

Page 98

GAO-14-821 BOP Cost Efficiencies

Selected Characteristics of Eligible Offenders
Retroactive Application of Amendment 782 With
Release On or After November 1, 2015
(FY1992 through FY2015)
RacelE thnicity
White
Black
Hispanic
Other
Total

10,734
14,427
19,958
1, 171
46,290

23 .2%
3 1.2%
43 .1%
2.5%
100.0%

U.S . Citizen
Non-Citizen
Total

34,928
11 ,411
46,339

75.4%
24.6%
100.0%

Male
Female
Total

42,759
3,615
46,374

92.2%
7.8%
100.0%

17,580
6,3 20
9,232
5,473
3,417
4,354
46,376

37.9%
13 .6%
19.9%
11.8%
7.4%
9.4%
100.0%

Citizenship

Gender

Criminal Histol'Y

Cate~ory

I
II
III
IV
V
VI
Total

PI'ojected Year of Release
November 1, 2015
within 1 yf
within 2 yr
within 3 yr
within 4 yr
within 5 yr
more than 5 yr
Total

Retroactive*
7,953
8,550
6,938
5,473
4,177
2,909
9,350
45,350

If Not Retroactive
7,609
7,461
6,207
5,291
3,923
14,859
45,350

The analysis involves a total of 46,376 cases, however, cases missing infonnation for any specific analysis are excluded fi'om
that analysis. Total percentages may not add to exactly 100% due to rounding.
* A total of 16,503 offenders are estimated to be elgible for release during the first year in which offenders may be released under
retroactive application of the amendment. This is 8,894 more than will be released upon expiration of their current sentence.
SOURCE: U.S. Sentencing Commission, 1992 - 2005 Datatlles, USSCFY92 - USSCFY13.

Page 99

GAO-14-821 BOP Cost Efficiencies

Possible Release Timing for Retroactive Eligible Offenders bv OistJict
(FY1992 through FY2015)
Eligible rol' Relesse

t:ligible for H.clease

f:ligible ror Relensr
in Six or More

in Year Four

in Year fin'

Years

Eligible for

Immediate Release
11101115

Eligible for Release
in YearOnc*

1:':ligiblc ro.· Release
in VtarTw"

Eligible for Release
in Ye4U' Three

TOTAL

CIRCUIT
N
7,953

17.5

N
8.S50

2S
25

12.7
12.7

37

18.S

37

18.8

FIRST CIRCUIT
Maine
Massachusetts
New Ham pshire
Puerto Rico
Rhode Island

231

412

45

193
27.1

37

II
149
14

10.8
18 .1
14.9
17.7
9.1
16 . .1

21

SECOND CIRCUIT
Connecticut
New York
Eastern

3~~

17.2

45

16.7

380
67

51

49
58

16.7
21.6

13~
5~

Distria

TOTAL
D.CClRCUIT
District oC('olumbia

30
27

%

Northem

48

Soulhern
Western
Vermont

136

17,4
17.9
18.0

55

16.3

9

11.8

J'HIRD C1RCUIT
DclaW8n!

207

15.8
22.5
17. 1

New Jersey
Pennsylvania

Eastern
Middle
Western

54
55

65
19

267

82

33.9
17.9

14
290

22.6

17.6

16..1

19

22. 1

19.0

328

16.-1

24.8

51

18.9

66
53

2~.5

19.8
\3.7
12.8

14.1

118
117

170

14 .9
20 .8

100

26.4

18.6

1.644

II

12.8

86

371
26

18.5

2.001
170

I
214
6

1.1.0
7.0

122
6

264
31

13,2

196

9.8

118

5.9

J 1.5

32

11.9

18

6.7

43
39

14.7
14.6
13 .3
12.8
9.2

21
29

14
10
48

4.8

12

7.2
10.8
8.6
11.0
is.8

12.2

120
I
26

9.2
2.5
8.3

74

101

16

65

14.5

2004

198

15. 1
20.0
18,4

160

38

I::!.l
13.6
12,8
8.8
15.2

46
31
12

43

5.0

15
5

37

147

4.8
1A
7.0

6.3
6.5
7.9

49
31
171
85
9

9.6
16.7

11.6
22.6

197

293
268

25.2

7Si
337

11 .8

76

21.7
27.5

1.310

II

284

16

5.6
0.0
5.1

41

1).0

.115

9.8
11.0
7.0

33

7.0

12. 1

)

8.8
9.1

135
43
49
5

28.7
15.)
28.7
15.:2

470
281
171
33

454
37

7.5
6.6

1,545

25.6

6.035

IO~

18.3

557

1,115
427

o

2.5
15

40

158
183

812
102

13.5
18.3

683

II 3
14 -I

559

9.3

80

49

8.8

13.2

143
50
82

128

122

109

112

10.0

340

44
60

93

9.7

40

137

11 .6

85

6.2
9.0

30.5
21 .8
21.1

118

36
63
109

6.8

116

10.3
9.2
12.5

29

12.4

11 . 7
12.6
12.3

10.0
8.4

III

20.8
18.1

264

28.0

942

11.1
156

1-15
83

12.7
11.9

121

106

98

8.6

85

7.4

1.142

114

49

7.6

44

6.8

396
160

34.7

73

24.9

643

26.1

51

15.3

III

29

8.7

13

3.9

28

296

40

17.7

37
28

12,4

14

6.2

10

4,4

25

8,4
11 . 1

333
226

127

87
67

181
62

294

12.1
11.3
17.9

17.0

22.1
18.::!

23 .1

166

40

18.6

38

147
89

2.139

7.2

16

11.1

102

17.0

1~

280

36

954

364

1104
8.8
1.6

64

17.0

6.9
4.2
5.0

12.0

II

58

197

256
19

104

26.0

25.4
25.4

16.0
13.3
JO ,5
8.1

17.4
15.4
28.9

43

50
50

6.1

13.6

12.6

'" For example, in (he District ofCc,llumbia. 62 offenders are eligible for

Page 100

18.2

4.1
4.1

6. 1

11

14.1
13.5
15.2

20.1

SOllth;;:rn

342
22
19

N
45,350

12

40
23

153

88
42

18.1
18.7

2004

387
31
33

%

20.6

13.7
13.7

64

86
150
I .U

West ViTgin ia
Northern

27
17

15.5

85

134

19.3
19.3

21.0

140

Eastern
Western

38
38

N
9,350

9.2

%

N
2,909

%
6.4

N
4.177

73

Middle
Western

12.1

59

Easlelll

Souch Carolina
Virginia

N
5.473

11.7
23.1

1.028

%

15.3

2:!5

1S.:!

Virgin Islands

FOURTH CIRCUIT
Maryland
North Carolina

295
14

'Yo

N
6.938

%

18.9

relea~e

650

between Novennber I. 2015 aIld OL1.ober 31,2016.

GAO-14-821 BOP Cost Efficiencies

Possible Release Timing for Retroactive Eligible Offenders by District
(FY1992 through FY2015)
Eligible for
Immediate Rcru~iI:
1110111;

CIRCUIT
Diotria
FIFTII CIRCUIT
Louisiana

E1it:iblc (or Releasc

~: tigibl(.'

(or Release

f.li~ibte

in

in YearT",·o

in YClirOne

for Rc1caSt

YC1lrTh~c

F.ti~ible

for Release
in Year Four

[I~iblc

for Relean
in Year Five

Eligible for Release
in Si .. or Mo~

TOTAL

Yeus

N

%

N

'/1

x

~

%

N

20.0

1,561

1.180

N
901

%

2.023

%
11 .6

N

IS.)

%
15.4

%

I,SS3

8.9

608

6.0

2.013

19.9

1U.140

Eastern

46

46
20
49

14.8
175
18.4

38
12
29

1::!.2.
10.5
10.9

10.0

28

9.0

66

24
35

18.0
21 .1
13. 1

31

22

14.8
19.3
21.3

56

Middle
Western

7.0

6

5.3

19

7.1

14

21.2
19.3
24.0

311
114
267

16.4
27.5

159
334

57

N

5.2

64

4.4

26

7.2

92

Mj~sissjppi

34
46

21.4
13 .8

33

Eastern
Northern
Southern

::!IO
220
559

14.8
16.1
19.5

209

Western

659

Nonhern
SOllthcrn

52

20.8
15.6

23
52

14.5
15.6

22
44

13.8
13. 2

14

8.8

24

7.2

24

14.8
13. 1
23 i

14.3
13.4
161
15.9

183
150
334
368

12.9

142

7.1

367
403
432
541

25.9

114
255
284

10.0
9.1

101

11.0
11.6
11. 1

16,4

1.415
1.368
2,867
3,305

Texas

19.9

179
680
755

:!l.8

203
183
462
524

739

17,9

913

22 , I

709

17. 1

514

12.4

362

Eastern

9~

22.5
24.S

116

49

123
61

21.2
15.0

11.7

Western

16.8
19.9

64

.14

13.8

48
23

75
55

lQ.3
15.9

T'
61

20.3
17.6

52
48

57
39

150
II :2

32
27

7.8

95
87

18.6

131

25.6

20.3

Inl

23.5

90
76

63
57

12 3
13.3

45
36

8.4

16.:!
21 .::!
18. 1

131
30

18.4
13.9
15.5

11.9

21
66

131
23

104
17
30

6

46

15.2
10.8

11 .3

98

20.9
19.9
23.1

~03

3::!
77

7. 1

455

167

437

16.0

389

143

292

10. 7

:!51

79
123
51

19.4

18.2

12.5
13.7

38
101
48

9.3
14.9

32
75

10.6

51
93
63

10.0

58

7.9
Il.l
12.0

56
63

17.0
14.7

73

12.2

53

12A

68
7:!

20.7
16.8

28
56

45
38

255

76
28

30 ,2
18.8

48
14

19.0
94

30
13

SIXTH CIRCUIT

29.5

8.9

109
145

8.6

174

80
5. 1
5.3

8.7

265

6.4

637

15.4

4,139

8.8
9.3

30
12

5.5
4.9

73
30

13.4

12.2

546
246

8.4

16
30

4.2

72
87

19.0

379

25.1

347

29
27

5.7

59
45

115

51::!

6.3

10.5

429

88

8,0
4.0
6.4

168
22
81

15.2

1.104

27

14.6
19. 1

151
425

CJ.3

210

7.7

689

25.3

2,724

30
59
77

7.4
8.7
16.0

29

7.1

55

8.1

148
171

36.4
25.3

677

58

I~ . O

127

263

482

8.5

21

6. 1

23

5.4

63
126

19.1

35

6.4
8.2

20

13.1

29.4

329
428

4.0
13.4

16

6.3
6.0

2)

10.)

27

18. 1

15. 1

Kentucky

MjchigaLl
Eastern
Western

Ohio
Northern
SOllthern

Tennessee
Eastern
Middle
Westem

SEVENTH CIRCUIT
Illinois
Central
Northern
Southern
Indiana
Northern
Southern
Wisconsin

Eastern
Western

Page 101

179

179

13.1

n.:
13.S
\76

17.7

II 9

10

87

20

8.8

9.4

9

8.6

407

252
149

GAO-14-821 BOP Cost Efficiencies

Possible Release Timing for Retroactive Eligible Offenders by District
(FY1992 through FY2015)
Eligible for Release

Eligible I'or
Immediate Rele.,Sc

11101115

F.ligible fOI" Release
in YearOue

Eligible for

Rele~se

in ¥c.IIrTwo

Eligible fur Release
ill YcarThree

Eligible for Release
in Ye'AT Four

Eli~ble

for Release

in Ycar Fi"'e

TOTAL

in Six or More
Years

CIRCUIT

Di,uiC1

N
739

%

l'I

~~

N

%

17.5

628

14.8

501

11.8

75
44

22.7
I i .O

45
47

13.6
18.1

45

13.6

17.4

42

16.2

123
82
75

19,8
15.0
:!4,t

SO

12.9
15.6

65
63

10.5

17~j

74
78
39

11.9

85
55

36

130
89

2),4
16A
12,6

123
88

2:! 1

96

11.3

86
113

1.5.8

14.~

38

~6

19.5

~4

16.2
17.6
18.0
18.0

889
)6
172

17.2
19,9
25.3

1, 153
3~

Central

114

Eastern

15,7
13,4

Nllrthem

92
58

Southern

77

EIGHTH CIRCUIT
:\rkan5as
Eastern
V....estern
hw..'a
Nmthem
Southern

MinnesQla
Missouri
Ea5tcm
Western
Nebraska
North Dakota
South Dako,"
NIN1H ClRCL'IT
Alaska

Arizuna
Ca lifornia

"

%

742

17.5

51
45

15. 5

91
30

In

14.3
125

15.7

6.4

5J

7.3

41

15.5
15.8

330
259

47
38

7.6
7.0

152
158

24.5
29.0

621
545

7.1

61

19.6

311

14
43
47
12
12

4.3
7.9
6.5
5.7
9.0

88
120
176
61
25

15.8
22.1
24.4
2&.9
lS.8

556
543
720
21 I
133

273

849
30
69

16.6

5. 169
lSI

4.S

5.3
6.1
2.1

16.4

11
14

10.1

681

169
114
45
69
10

23.3

724
686
289
686
37

61

50

21 .9
23.5
25.5
17.2

24

0.0
11.3

8.3

25

11.3

204

5.2

~O

8.6

232

12.9

7.4

53
70
93
21
I.l

9. 5
12.9
12.9
10.0
9.8

42
47
73
17
15

7.6
8.7
10.1
8.1

1 2.~

13.3
8. 7

465
19
33

8.1

7.5

)2

15.1

18

13.5

22.3
18.8

873

16.9

27

223

32.1

III

149
16.3

667
24
59

134
132
51
176
9
54
60
65
56

103
114
49

14.2

88

1~2

72

9,9

44

6.1

18.1
17.0

98
37
128
5
29
37

14.3

82
28
68

12.0

44
20

6.4
6.9

9.9
2.7

)3

4.8

10,0

IS
10

39

19.1
207

41
64

10.1

35

11.~

2)

48

17.6

41

18. 1

~5

18,9

21
19

80
41

Western

o

4,1'29

23

Eastern

Oregon

55
48
49
55

%
22.1

11 .6

50

Montana
Nevada
Northern Mariana Lslands

N
933

11.6

18,9

Idaho

%
6.7

12.7

40

Guam
Hawaii

N
285

%

q.s

42
21

18.5
19.2
18.0
25.7
14.3
19A
20.8
175
19.2
50.0
23,6

20.1
11.2
16.2
19,7
16,6
13,2

N
401

(),O

135

37
43
60

59
43

19.7
13.5
13_3
14.9
161
203

1~8

IV
1~5

104
I}&

50

1~4

28
23
34
17

11..1
9.0
10.5

9.7

8.0
9.1
5.8

36

1~4

0.0

o

~

20.3

28

13.2

13

6.1

11 .3
10.8

24
21

11 .8

19

IS
I
14

2.7
5.4
3.5
5.1
6.2

68
95

16.7
6.6

16.6
15.6

111.1
27.0

N

179
~g9

372

291
6
212

Wa.;hjng.t~m

Page 102

9. 1

17
12

GAO-14-821 BOP Cost Efficiencies

Possible Release Timing for Retroactive Eligible Offenders bv District
(FY1992 through FY2015)
Eligihle for
Immediate Release
111U1I15

Eligible for RcJcasc
in Year One

EJigih1c (or Relc3!11l'
in YeAr Two

Eligible for Release
in Year Three

Eli¢.blc for Release
in \ 'eal' FOllr

Eligible for Relense
in Year five

Eligible for Release
in Six or Mo.-c

TOT,U

Years

CIRCUIT

m."kl

rEN! H (I RCULT
Colorado
Kansas

New Mexico

N
419
37
84

%

:-I

%

18 , ~

428

14.6
l3 .7

48
104

89

23.3

92

18,9
19.0
17.0
24.1

21
24
SI

19.3

20
31
40

N
322
46

~.

N

14.2
18.2

263
33

130

%
11 .6

N

~ ,.

~.

%

N

21.8

2.261

21.3
25.8
14.9

613
38:::!

6.7

7.1

N
151
17

10.4

46

7.S

5,;

23

6.0

5

4,6
6.3

27

9
16
16
19

7.0
5.2
8,6

55
42

2~.0

9,9

71

32,0

109
144
229
309
222

:'16

6,3

1.12}

22 .~

5.006

1.9
5.9
7.3

36

11.1

17
25

57

19.7

171
290

60

17.4

3~4

198
153
266

22,3

33.9

5A

20,7

1.334
451
1.283
136

186
18

8.1

64
21

12,8
63
6,6
7.4

&1
61

13,2

76

1~.4

16.0

39

10,2

7.3

11

14,6

12.8
15.3
10.0

14
9
15

,.

~3

17.2

-~

10.4

14
22
23
39
17

12.6
7.7

23
12
~67

9.:'

18

10,;
10.7

6. 7

N
492
54
158
5i

253

OkLahoma

Eastern

16,7
22.3

18.3

72
41

:n.3
18.5

64
29

21.5
17.5
:20.7
13,1

1.021

10.4

807

16.1

691

13.8

580

11 .6

Middle
Nonhcm

4S

26.3

17

9,9

29
34

12,3
11.4

80

40
70

13.8

Sml{hern
Florida
MiddLe
Northern

26.9
23,3

17,0
11.7

21

78

20.3

49

14..2
14,7

Northern
W~stem

Utah

W)Oming
ELEVENTH CIRCUIT

29

12.7

28

]4.8
19,4

13.6

Alabama

235
97
249

Southern
Georgia
Middle

Nonhcm
Slllllhcm
Of the 46.376 offenders

idcnriti~d

33
29

8.4

31
31

9.0

17,6
215

207

15,5
10.6

196

155

11 .6

1~8

48

43

95

45

9.6
8.2

234

18.2

198

154

152

100
II 8

n

19.4

115

9.0

13 ,1

21
55
31

15.3

32

3.8
6.2

36

10,6

15~

]1).6

517

8,2

16

~.2

64

16.8

380

31
100

57
61

24.2

45

19. 1

37

15. 7

19.1

11.8

49

95

13.0

106

27.9

73

19.2

49

12,9

67
41

10.8

115

28
69

8.9

8.6
6.2

:lS eligible [-Or relief under the amendment, Ct)lrunisi:iion records containcd :\lIllieicnt inftxm:l1ion to pcrtllml d,is anal)'sis ft)C 4S,~SO l,rrcnders.

E.,tiU~1ttJ n:1/:;;a.\(. dALd -.t(' de1.::rmined using lht: Commission's prison and sentencing impact mood which !If)Prie,, f1~ l;."U&ldinc.' cltangl";'. to ulT~tc:d t,(ta:n&r!\:wld rc~ 11)(':<. olT~ in a propMi('lrcl munncr.
Undc'~ rno&~ .ffct'tl:d oni"td~I'S: I J receive a 11(:\...· otlcnse level; 2) haw a (~W scnl~nc:ing rang~ dcu:rm~d (d$ipg ,h;: r.lng~s from the Scnlc~--il\g r~bt. .\).J~ ~aI to the: s:tr'l'lC rel:t.tJ"c position wilhin
(<W ~fdcl ,hc- \NigitW 1;U'idclinc mn~ (c.g., an ofTl:odt:r curn:mly sentenced at the lTlic.\p<.)im ~)r\hc ~wjgJntlllukh:afk.' f·:J~...~thcn "i\1 b.! scntenced foJd\\: mi~po;r\l ,,(the ~\\ ~uich;tim; r.mg.~); lind 4) rcceive
stalUl(\r:' aJld l:,"Uideiine trwnps wben applic~lnk . Other assumplil\flS incoI"(K\rntcd into the mood include: [) t\ft~rw:ier.-; eflfll lt~ maximum allowable good-time (clirrently 54 days per Yelr served H'f imposed sentences gre!lter
than one year but not life ilTlp"isonlTII:l)t): aoo 2) ollerders sene the Ics~'fofA) the ,~ll!llCe im,>O!.Cd les ... the: maximum aUmvnble gO<'lCi cooduct time, or B}lhcir ...'Slimal<:d rernniniog life C:<pc<::tallCY. ~dupoi'l an nduary
table ino.lrporating age. (!lee. :md sex,

SOURCE: US Sl!'nt'L'ncing Commis5kHl. 1992 - ~013 Dalt=1ll1es. USSCFY92 -lJSSCFY13.

Page 103

GAO-14-821 BOP Cost Efficiencies

6. USSC analysis of option 6: Increase good time credits for good
conduct time (GCT).
Under this option, a credit for GCT would be awarded to current offenders
using a formula that would result in 54 days’ credit per year of sentence,
rather than the current BOP interpretation that limits the GCT credit to 47
days per served sentence year. This would result in eligible inmates
serving 85 percent of their sentences, rather than the higher 87 percent
average resulting from BOP’s current interpretation of 18 U.S.C. § 3624(b).

Page 104

GAO-14-821 BOP Cost Efficiencies

UNITED STATES SENTENCING COMMISSION
ONE COLUMBUS CIRCLE, N.E.
SUITE 2-500, SOUTH LOBBY
WASHINGTON, D .C. 20002-8002
(202) 502-4500
FAX (202) 502-4699

June 12,2014

Jonathan R. Tumin
Senior Analyst
U.S. Government Accountability Office
441 G St., NW
Washington, DC 20548
Dear Mr. Tumin:
The Government Accountability Office has asked the U.S. Sentencing
Commission to assist it in connection with its engagement (number 441173). We
understand that through this engagement the GAO is examining ways in which the
operating costs of the Federal Bureau of Prisons (BoP) can be reduced. As we have
previously discussed with you and your colleagues, the Commission has agreed to
analyze on GAO's behalf several policy options which could reduce the overall size of
the prison population that the BoP administers. Specifically, our analyses of these policy
options will consider the effect on the federal prison population if the policies were
implemented in such a way that the options would apply to inmates currently incarcerated
in the BoP system.
In this memorandum, we provide to you another of these estimates. Here we
provide our estimate of the impact on the federal prison population if the current method
of determining "good time credit" were to be changed so that offenders could earn up to
54 days per year for good conduct and the BoP would be authorized to apply that
expanded credit to all offenders currently incarcerated.
As we discuss more fully below, we estimate that 179,265 offenders who are
incarcerated and would still be in prison on October 1,2014 would be eligible to receive
additional good conduct credit under such a policy change. The average increase in good
conduct credit for these offenders would be 2.3 months (2.2%). The estimated total
savings to the BoP from such a policy change applied retroactively would be 34,359 bed
years.

Page 105

GAO-14-821 BOP Cost Efficiencies

1.

The Policy Option We Considered

Under current law, I a prisoner who is serving a term of imprison of more than one
year may receive credit to the service of a sentence if they behave satisfactorily while
incarcerated. Section 3 624(b) of title 18 provides:
A prisoner may receive credit toward the service of the prisoner's
sentence, beyond the time served, of up to 54 days at the end of each year
of the prisoner's term of imprisonment, beginning at the end of the first
year of the term .... [C]redit for the last year or portion of a year of the
term of imprisonment shall be prorated and credited within the last six
weeks of the sentence. 2
The goal of this provision is, in essence, that offenders serve at least 85 percent of the
sentence imposed. The BoP currently uses a formula to determine good conduct credit
that results in a maximum of 47 days of credit being given for each year of incarceration
served. The difference between the BoP approach and the 54 days described in the
statute occurs because ofthe BoP's interpretation of when the good conduct credit is
accrued (following the service of each year) and how the credit is to be applied to the
partial year remaining at the end of an offender's sentence.
An alternate method of calculating the good conduct credit could be to simply
award 54 days of credit for each year of sentence imposed at the beginning of the
sentence. Although the BoP's current interpretation of the statute has been upheld by the
Supreme Court,3 the Department of Justice has proposed to Congress that it amend the
governing statute so as to allow the BoP to award up to 54 days per year. 4
You have asked us to consider a change to the statute whereby the calculation of
the good conduct credit is changed to authorize 54 days per year to be awarded (the
"Policy Option").

II.

What We Found

Applying the Policy Option to offenders in the custody of the BoP, we estimate
that there are 179,265 offenders who would be incarcerated on October 1,2014 who
would be eligible for a reduction in the time they have left to serve on the sentence

J

18 U.S.C. § 3624(b).

2

Id.

3

Barber v. Thomas, 560 U.S. _

(2010).

4 Lanny Brewer, Assistant Attorney General, Criminal Division, U.S. Department of Justice, Remarks at
the National District Attorneys Association Summer Conference (July 23, 2012) available at
http://www.justice.gov/criminal/pr/speeches/2012/crm-speech-120723.htm!.

2
Page 106

GAO-14-821 BOP Cost Efficiencies

imposed if the Policy Option were to be retroactively applied. If the BoP were to grant
each ofthese offenders the full good time credit of the Policy Option we estimate that the
total savings to the BoP would be 34,359 bed years. 5
Table 1 provides information on some of the demographic characteristic ofthe
offenders eligible for a sentence reduction pursuant to the Policy Option. Table 2
provides an estimate of the number of offenders who would be released each year if the
Policy Option was applied retroactively compared to the number of offenders who will be
released each year pursuant to their current sentence. As can be seen on this table, if the
Policy Option were fully retroactive on October 1,2014, there would be 819 offenders
eligible for immediate release. Within the first year of the effective date of the Policy
Option, another 25,918 offenders would be eligible for release. In total, the number of
offenders eligible for release under the Policy Option in the first year is 2,878 more than
the number who will be released under their current sentence. Table 3 provides
information on selected sentencing guideline characteristics and the criminal history
category of these offenders.
For the 179,265 offenders who would be eligible to receive additional good
conduct credit under the Policy Option, the current average time served is 103.3 months.
If the BoP were to award the maximum good time credit under the Policy Option, the
projected average time served would be 101.0 months, a reduction of2.3 months (or
2.2%). Based on this reduction, the estimated total savings to the BoP from the
retroactive application of the Policy Option would be 34,359 bed years. This savings
would not be realized in any single year but is the cumulative savings realized over
several years.

III.

How We Conducted This Analysis
A.

Methodology

The methodology for this analysis is based on the Commission's Prison Impact
Model, which has been in use in some form since the guidelines were first developed.
This model is used to estimate the impact of proposed statutory and guideline
amendments on newly sentenced offenders and to project the future impact those
amendments will have on bed space in the BoP. For this analysis, those offenders who
appear to be eligible to receive a reduced sentence were hypothetically "resentenced"
with the computer program as if the proposed policy change had been in effect in the year
in which they were sentenced. A new release date for each offender also was calculated
in order to determine when the offender would be eligible for release if he or she were
provided the full reduction in sentence provided by the proposed policy change.

5 A "bed year" is the cost to the BoP of incarcerating one inmate for one year. For example, one inmate
who serves five years of imprisonment accounts for five bed years.

3
Page 107

GAO-14-821 BOP Cost Efficiencies

B.

The Offender Population We Studied

The Bureau of Prisons provided the Commission with a datafile of inmates who
were in the custody of the BoP on Jan 25,2014. That file contained approximately
189,000 offenders. Approximately 184,000 of these offenders were sentenced between
fiscal year 1992 and fiscal 2013. 6 USSC staffwas able to match 171,765 of these
offenders to Commission records. Of these, 138,894 were estimated to remain
incarcerated on October 1,2014.
In order to approximate the group of offenders who will be sentenced in fiscal
year 2014, the Commission used the FY2013 datafile and moved all sentence dates
forward by one year. Staffthen determined which of these offenders would be
incarcerated on October 1,2014. This process added another 40,178 offenders into the
analysis. Between the two groups, the Commission's analysis included data on 179,072
offenders.
C.

Our Assumptions

In performing our analysis, we have been required to make some assumptions (set
forth below) concerning the decisions that Congress and the courts would make in
determining whether, and to what extent, to reduce the sentences of offenders eligible to
receive a modification of sentence pursuant to the Policy Option discussed above. These
assumptions may not hold in every case.
1.

Procedural assumptions

Specifically, we have assumed that Congress would amend the provisions of 18
U.S.c. § 3624(b) in such a way that the BoP would award a maximum of good conduct
credit of 54 days per year of sentence imposed. We have also assumed that Congress
would authorize the statutory change to be applied retroactively to offenders incarcerated
in the BoP.
2.

Substantive assumptions

We have assumed that the sentence for each offender would be reduced based on
the maximum good conduct credit allowed under the Policy Option. This assumption
may not hold for every offender, as some offenders lose a portion or all of their earned
good conduct credit as a sanction for unacceptable behavior. As a result, our estimate of
the impact of the Policy Options may be overbroad, but not to an extent we can
determine. We have also assumed that offenders would serve the lesser ofthe newly
calculated sentence or their life expectancies.

6 The analysis was limited to data from fiscal year 1992 through September 30,2013 (fiscal years 1992
through 2013) because the Commission's data collection efforts prior to fiscal year 1992 were not as
complete as in later years. For example, the Commission did not collect information on the type of drug
involved in drug offenses prior to fiscal year 1992.

4
Page 108

GAO-14-821 BOP Cost Efficiencies

If you have any questions about this analysis, or any of our prior analyses that we
have provided to the GAO, please do not hesitate to contact the Commission. This
analysis completes the list of analyses we have agreed to provide to you. We hope that
the information the Commission has provided has been useful to the GAO in its work.
Sincerely,

Noah D. Bookbinder
Director
Office of Legislative and Public Affairs

Director
Office of Research and Data

Attachments

5
Page 109

GAO-14-821 BOP Cost Efficiencies

Table 1
Demographic Characteristics of Eligible Offenders
(FY1992 through FY2014)

DEMOGRAPHICS
RacelEthnicity
White
Black
Hispanic
Other
Total

47,363
58,531
67,051
6,045
178,990

26.5%
32.7%
37.5%
3.4%
100.0%

U.S. Citizen
Non-Citizen
Total

133,096
46,073
179,169

74.3%
25.7%
100.0%

Male
Female
Total

166,582
12,675
179,257

92.9%
7.1%
100.0%

Citizenship

Gender

Average Age

39

36

(as of November 1,2014)

(at sentencing)

The analysis involves a total of 179,265 cases, however, cases missing information for any specific analysis are
excluded from that analysis.
Total percentages may not add to exactly 100% due to rounding.
SOURCE: U.S . Sentencing Commission, 1992 - 2013 Datafiles, USSCFY92 - USSCFYI3 .

Page 110

GAO-14-821 BOP Cost Efficiencies

Table 2
Projected Y car of Release for Retroactive Eligible Offenders
(FY1992 through FY2014)

IF OPTION
RETROACTIVE

IF OPTION NOT
RETROACTIVE

Release Date

N

N

Immediate Release

819

within 1 yr

25,918

23,859

within 2 yr

31,111

30,901

within 3 yr

24,980

24,972

within 4 yr

19,287

19,395

within 5 yr

14,841

15,063

within 6 yr +

62,309

65,075

Of the 179,265 offenders who appear to be eligible to benefit under the Policy Option,
Commission records contained sufficient information to perform this analysis for 179,265 offenders.
SOURCE: U.S. Sentencing Commission, 1992 - 2013 Datafiles, USSCFY92 - USSCFYI3.

Page 111

GAO-14-821 BOP Cost Efficiencies

Table 3
Guideline Sentencing Characteristics, Crimina] History, and
Position Relative to the Guideline Range of Eligible Offenders
(FY1992 through FY2014)

CHARACTERISTICS
Average Base Offense Level
Weapon Specific Offense Characteristic
Firearms Mandatory Minimum Applied
Safety Valve §5C1.2
Aggravating Role §3B1.1
Mitigating Role §3B1.2
Obstruction Adjustment §3C1.1
Career Offender Status §4B1.1

23,888
21,962
15,970
17,959
8,705
8,416
19,532

13.3%
12.3%
16.3%
10.0%
4.9%
4.7%
10.9%

58,109
19,809
28,548
19,857
13,790
39,152
179,265

32.4%
11.1%
15.9%
11.1%
7.7%
21.8%
100%

97,940
5,615
25,871
49,478
178,904

54.7%
3.1%
14.5%
27.7%
100%

Criminal History Category
I
II

III
IV
V

VI
Total
Sentence Relative to the Guideline Ran2e
Within Range
Above Range
Substantial Assistance §5K1.1
Otherwise Below Range
Total

The analysis involves a total of 179,265 cases, however, cases missing information for any specific
analysis are excluded from that analysis.
Total percentages may not add to exactly 100% due to rounding.
SOURCE: U.S. Sentencing Commission, 1992 - 2013 Datafiles, USSCFY92 - USSCFYI3 .

Page 112

GAO-14-821 BOP Cost Efficiencies

7. USSC analysis of option 7: Make the current “safety valve”
available to any drug offender with no offenses that receive 3 points
under the criminal history point system.
Federal law generally requires a sentencing judge to impose a minimum
sentence of imprisonment following conviction for any number of federal
offenses; however, there are two statutorily provided exceptions. One of
these exceptions, commonly referred to as the “safety valve,” was created
by the Violent Crime Control and Law Enforcement Act of 1994, Pub. L. No.
103-322, 108 Stat. 1796, and is available for certain types of defendants for
certain offenses that carry minimum sentences. Under this option, the
current statutory requirements for eligibility for safety valve relief in drug
cases involving a mandatory minimum penalty would be expanded to cover
incarcerated offenders with any number of criminal history points, provided
that the offender does not have a conviction for which the sentence
imposed was a 3-point offense under USSC sentencing guidelines. Under
USSC sentencing guidelines, 3 criminal history points are assigned if an
offender has been previously sentenced or is being sentenced to a prison
term of 13 months or more.

Page 113

GAO-14-821 BOP Cost Efficiencies

UNITED STATES SENTENCING COMMISSION
ONE COLUMBUS CIRCLE, N.E.
SUITE 2-500, SOUTH LOBBY
WASHINGTON, D.C. 20002-8002
(202) 502-4500
FAX (202) 502-4699

April 24, 2014

Jonathan R. Tumin
Senior Analyst
U.S. Government Accountability Office
441 G St., NW
Washington, DC 20548
Dear Mr. Tumin:
The Government Accountability Office has asked the U.S. Sentencing
Commission to assist it in connection with its engagement (number 441173). We
understand that through this engagement the GAO is examining ways in which the
operating costs of the Federal Bureau of Prisons (BoP) can be reduced. As we have
previously discussed with you and your colleagues, the Commission has agreed to
analyze on GAO's behalf several policy options which could reduce the overall size of
the prison popUlation that the BoP administers. Specifically, our analyses of these policy
options will consider the effect on the federal prison population if the policies were
implemented in such a way that the options would apply to inmates currently incarcerated
in the BoP system.
In this memorandum, we provide to you another of these estimates. Here we
provide our estimate of the impact on the federal prison population if the current statutory
requirements for eligibility for "safety valve" relief in drug cases involving a mandatory
minimum penalty! were to be expanded. Specifically, you have asked us to consider a
change to current law whereby offenders with any criminal history not involving a
conviction for which the sentence imposed exceeded 13 months would be eligible for
such relief and the courts would be authorized to resentence all federal offenders
convicted of a drug offense in a manner consistent with the revised statute.
As we discuss more fully below, we estimate that 11,949 offenders who are
incarcerated and would still be in prison as of October 1, 2014 would be eligible for a

I

18 U.S.c. § 3553(f); see also USSG §5C1.2.

Page 114

GAO-14-821 BOP Cost Efficiencies

sentence reduction. The average reduction in sentence for these offenders would be 18
months (19.1 percent). The estimated total savings to the BoP from such a policy change
applied retroactively would be 15,889 bed years.

1.

The Policy Option We Considered

One of the policy options that we have agreed to examine on behalf of the GAO is
the impact of a policy change in current law whereby the current statutory requirements
for eligibility for "safety valve" relief from the application of mandatory minimum
penalties in drug cases would be expanded. That provision currently requires courts to
sentence offenders who have been convicted of a drug offense carrying a mandatory
minimum penalty without regard for that penalty when all of several criteria are met. 2
One ofthose criteria involves the criminal history score 3 of the offender. Currently,
offenders are eligible for safety valve relief only if they have been assessed no more than
one criminal history point under the sentencing guidelines. 4 You have asked us to
consider a change to the statute whereby offenders would be eligible for safety valve
relief regardless of their criminal history score provided that none of their criminal
history involves a conviction for which the sentence imposed exceeded one year and one
month,5 and further provided they meet the remaining safety valve criteria (the "Policy
Option").

II.

What We Found

Applying the Policy Option to offenders in the custody of the BoP, we estimate
that there are 11,949 offenders who would be incarcerated on October 1, 2014 who would
be eligible for a reduction in their current sentence if the Policy Option were to be
retroactively applied. If the courts were to grant each of these offenders the full reduction
in their sentence we estimate that the total savings to the BoP would be 15,889 bed
years.6
Table 1 provides information on some of the demographic characteristic of the
offenders eligible for a sentence reduction pursuant to the Policy Option. Table 2
2

Id.

3

See USSG ChA, Pt.A, for a discussion on how the criminal history score is determined.

4

Criminal history points are assessed by the court at the time of sentencing. Id.

Offenders convicted of an offense for which the sentence imposed exceeded one year and one month are
assessed three points under the sentencing guidelines. See USSG §4A 1.1 (a). Therefore, this Policy Option
would expand safety valve relief to any offender who did not have any conviction that was assigned three
points under the criminal history provisions ofthe guidelines.
5

6 A "bed year" is the cost to the BoP of incarcerating one inmate for one year. For example, one inmate
who serves five years of imprisonment accounts for five bed years.

Page 115

2

GAO-14-821 BOP Cost Efficiencies

provides an estimate of the number of offenders who would be released each year if the
Policy Option was applied retroactively compared to the number of offenders who will be
released each year pursuant to their current sentence. As can be seen on this table, if the
Policy Option were fully retroactive on October 1,2014, more than 1,123 offenders
would be eligible for immediate release. Within the first year of the effective date of the
Policy Option, another 865 offenders would be eligible for release above the number who
will be released under their current sentence. Table 3 provides information on selected
sentencing guideline characteristics and the criminal history category of these offenders.
For the 11,949 offenders who would be eligible to receive a sentence reduction
under the Policy Option, the current average sentence is 94 months. If the courts were to
grant the full reduction possible in each case, the projected new average sentence for
these offenders would be 76 months, a reduction of 18 months (or 19.1 percent). Based
on this reduction, the estimated total savings to the BoP from the retroactive application
of the Policy Option would be 15,889 bed years. This savings would not be realized in
any single year but is the cumulative savings realized over several years.

III.

How We Conducted This Analysis
A.

Methodology

The methodology for this analysis is based on the Commission's Prison Impact
Model, which has been in use in some form since the guidelines were first developed.
This model is used to estimate the impact of proposed statutory and guideline
amendments on newly sentenced offenders and to project the future impact those
amendments will have on bed space in the BoP. For this analysis, those offenders who
appear to be eligible to receive a reduced sentence were hypothetically "resentenced"
with the computer program as if the proposed policy change had been in effect in the year
in which they were sentenced. A new release date for each offender also was calculated
in order to determine when the offender would be eligible for release if he or she were
provided the full reduction in sentence provided by the proposed policy change.
B.

The Offender Population We Studied

The Bureau of Prisons provided the Commission with a datafile of inmates who
were in the custody of the BoP on Jan 25,2014. That file contained approximately
189,000 offenders. Approximately 184,000 of these offenders were sentenced between
fiscal year 1992 and fiscal 2013. 7 USSC staff was able to match 171,765 of these
offenders to Commission records. Of these, 138,894 were estimated to remain
incarcerated on October 1,2014.

7 The analysis was limited to data from fiscal year 1992 through September 30, 2013 (fiscal years 1992
through 2013) because the Commission's data collection efforts prior to fiscal year 1992 were not as
complete as in later years. For example, the Commission did not collect information on the type of drug
involved in drug offenses prior to fiscal year 1992.

Page 116

3

GAO-14-821 BOP Cost Efficiencies

In order to approximate the group of offenders who will be sentenced in fiscal
year 2014, the Commission used the FY2013 datafile and moved all sentence dates
forward by one year. Staff then determined which of these offenders would be
incarcerated on October 1,2014. This process added another 40,178 offenders into the
analysis. Between the two groups, the Commission's analysis included data on 179,072
offenders.
C.

Our Assumptions

In performing our analysis, we have been required to make some assumptions (set
forth below) concerning the decisions that Congress and the courts would make in
determining whether, and to what extent, to reduce the sentences of offenders eligible to
receive a modification of sentence pursuant to the Policy Option discussed above. These
assumptions may not hold in every case.
1.

Procedural assumptions

Specifically, we have assumed that Congress would amend the provisions of 18
U.S.C. § 3553(f)(1) to the effect that "the defendant does not have a criminal history
score, as determined under the sentencing guidelines, that includes points added for any
conviction for which the sentence imposed exceeded one year and one month." We have
further assumed that the Commission would amend USSG §5C1.3 to conform it to the
provisions of the amended statutory safety valve provision8 but make no further changes
to the Guidelines Manual, so that the remaining provisions of the 2013 Guidelines
Manual would apply in these cases. Further, we have assumed that Congress would
authorize the statutory change to be applied retroactively to offenders incarcerated in the
BoP and that the Commission would amend USSG §IBl.I0 to authorize courts to apply
the amendment to USSG §5Cl.3 retroactively to those offenders on some specific date. 9
We also assumed that the courts would resentence offenders in a manner that is consistent
with the amended Guidelines Manual. 10 Finally, we assumed that the effective date of
any such Policy Option would be October 1,2014. Therefore, only offenders
incarcerated as of that date would be eligible to seek a reduction in sentence pursuant to
the Policy Option discussed in this memorandum.
See 28 U.S.C. § 994(0), which requires the Commission to periodically review and revise the sentencing
guidelines.

8

9 See 28 U.S.C. § 994(u), which requires that when the Commission "reduces the term of imprisonment
recommended in the guidelines applicable to a particular offense or category of offenses, it shall specify in
what circumstances and by what amount the sentences of prisoners serving terms of imprisonment for the
offense may be reduced." The Commission does this through amendments to USSG § 1B 1.1 0, which lists
all of the amendments to the Guidelines Manual that may be retroactively applied.
10 See 18 U.S.c. 3582(c)(2), which authorizes the courts to resentence an incarcerated offender in light of
the Commission's decision under 28 U.S.C. §§ 994(0) and (u) to retroactively apply an amendment to the
Guidelines Manual if the sentencing range that applied at the time the offender was sentenced would be
lower under the amended guideline.

4
Page 117

GAO-14-821 BOP Cost Efficiencies

2.

Substantive assumptions

In fiscal year 2010 approximately 66 percent of all drug offenders were convicted
of an offense carrying a mandatory minimum penalty. 11 Of those approximately 46
percent remain subject to the penalty at the time f sentencing. 12 S me offenders receive
relief from the effect of any such penalty by providing substantial assistance to the
government in the investigation or prosecution of another offender. 13 Others receive
relief from such a penalty through the statutory "safety valve" provision. 14 Some
offenders are eligible for both provisions.
Under the sentencing guidelines, drug offenders who meet the requirements of the
statutory safety valve provision, which is incorporated into the guidelines at USSG
SCI.3, are entitled to a two-level reduction in offense level. 15 This provision applies
regardless of whether the offender was convicted of an offense carrying a mandatory
minimum penalty. That is, for offenders who are convicted of an offense carrying a
mandatory minimum penalty, eligibility for the safety valve provides two benefits: 1)
statutory relief from application of the mandatory minimum penalty, and 2) a two-level
reduction in the offense level calculated under the sentencing guidelines. For drug
offenders who were not convicted of an offense carrying a mandatory minimum penalty,
eligibility for safety valve relief also results in the two-level reduction in offense level.
For the analysis described in this memorandum, we have assumed that this structure
would be unchanged, so that all drug offenders could benefit from the Policy Option if
they meet the revised safety valve criteria and regardless of whether a mandatory
minimum penalty had ever applied in the case.
Additional assumptions we have made for this analysis are that the sentence for
each offender would be reduced based on the maximum good conduct credit allowed by
the Bureau of Prisons, and offenders would serve the lesser of the newly calculated
sentence or their life expectancies.

II See U.S. SENTENCING COMMISSION, MANDATORY MINIMUM PENALTIES IN THE FEDERAL CRIMINAL
JuSTICE SYSTEM 154 (2011).

12Id.
13

See 18 U.S .C. § 3553(e); USSG §5Kl.1.

14

See 18 U.S.C. § 3553(f); USSG §5K3.1.

15

USSG §2Dl.1(b)(16).

Page 118

S

GAO-14-821 BOP Cost Efficiencies

If you have any questions about our analysis, please do not hesitate to contact the
Commission. We continue to work to analyze the other policy options that we discussed
with you and your colleagues. We will provide our analyses of these options as they are
completed.
Sincerely

)J~k (J ~$kto~fky
Noah D. Bookbinder
~
Director
Office of Legislative and Public Affairs

Director
Office of Research and Data

Attachments

Page 119

6

GAO-14-821 BOP Cost Efficiencies

Table 1
Demographic Characteristics of Drug Offenders
Potentially Eligible for Retroactive Expanded Safety Valve (No 3-Point Events)

DEMOGRAPHICS
RacelEthnicity

%

N

White
Black
Hispanic
Other
Total

3,496
3,463
4,667
309
11,935

29.3%
29.0%
39.1%
2.6%
100.0%

U.S. Citizen
Non-Citizen
Total

9,639
2,308
11,947

80.7%
19.3%
100.0%

Male
Female
Total

10,458
1,490
11,948

87.5%
12.5%
100.0%

Citizenship

Gender

33

Average Age

'The analysis involves a total of 11,949 cases in which the offender was identified as eligible to seek a sentence reduction if
the drug mandatory minimum changes were made retroactive. Cases missing information for any specific analysis are
excluded from that analysis.
Total percentages may not add to exactly 100% due to rounding.

SOURCE: U.S. Sentencing Commission, 1992 - 2013 Datafiles, USSCFY92 - USSCFYI3 and projected 2014 Datafile.

Page 120

GAO-14-821 BOP Cost Efficiencies

Table 2
Estimated Projected Year of Release for Drug Offenders
l>otentially Eligible for Retroactive Expansion of Safety Valve eN 0 3-Point Events)

If Policy Option Applied

Release Date

Retroactivly

Current Sentence

N

N

immediate

I 123

within 1 yr

2,108

1,243

within 2 yr

2,520

2,242

within 3 yr

2,118

2,009

within 4 yr

1,446

1,888

within 5 yr

912

1,327

within 6 yr +

1,690

3,208

'Of the 11 ,949 offenders who appear to be eligible for relief if the Poley Option was applied retroactively,
Commission records contained sufficient information to perform this analysis for 11 ,917 offenders.

SOURCE: U.S. Sentencing Commission, 1992 - 2013 Datafiles, USSCFY92 - USSCFY13 and
projected 2014 Datafile.

Page 121

GAO-14-821 BOP Cost Efficiencies

Table 3
Guideline Sentencing ChaTactel"istics and Criminal History of
Drug Offenders Potentially Eligible for Retroactive Expanded Safety Valve (No 3-Point Events)

CHARACTERISTICS

N

0/0

Average Base Offense Level

30

Weapon Specific Offense Characteristic

0

0.0%

Firearms Mandatory Minimum Applied

0

0.0%

Safety Valve §5C1.2

0

0.0%

Aggravating Role §3Bl.l

0

0.0%

1,339

11.2%

Obstruction Adjustment §3Cl.l

0

0.0%

Career Offender Status §4B1.1

0

0.0%

0

0.0%

II

4,658

39.0%

III

4,936

41.3%

IV

1,452

12.2%

599

5.0%

Mitigating Role §3B1.2

Criminal History Category
I

V

VI
Total

304

2.5%

11,949

100%

IThe analysis involves a total of 11,949 cases in which the offender was identified as eligible to seek a sentence reduction
if the expanded safety valve changes were made retroactive. Cases missing information for any specific analysis are
excluded from that analysis.
Total percentages may not add to exactly 100% due to rounding.

SOURCE: U.S. Sentencing Commission, 1992 - 2013 Datafiles, USSCFY92 - USSCFY13 and projected 2014 Datafile.

Page 122

GAO-14-821 BOP Cost Efficiencies

8. USSC analysis of option 8: Expand the safety valve as described in
option 7 to include both non-violent drug offenders and any
nonviolent offenders provided that they do not have a 3-point offense.
Under this option, the current statutory requirements for eligibility for "safety
valve" relief in drug cases involving a mandatory minimum penalty would
be expanded to cover both drug offenders and other non-violent (non-drug)
offenders. No offender could have a conviction for which the sentence
imposed was a 3-point offense under USSC sentencing guidelines.

Page 123

GAO-14-821 BOP Cost Efficiencies

UNITED STATES SENTENCING COMMISSION
ONE COLUMBUS CIRCLE, N.E.
SUITE 2-500, SOUTH LOBBY
WASHINGTON, D.C. 20002-8002
(202) 502-4500
FAX (202) 502-4699

June 11,2014

Jonathan R. Tumin
Senior Analyst
U.S. Government Accountability Office
441 G St., NW
Washington, DC 20548
Dear Mr. Tumin:
The Government Accountability Office has asked the U.S. Sentencing
Commission to assist it in connection with its engagement (number 441173). We
understand that through this engagement the GAO is examining ways in which the
operating costs of the Federal Bureau of Prisons (BoP) can be reduced. As we have
previously discussed with you and your colleagues, the Commission has agreed to
analyze on GAO's behalf several policy options which could reduce the overall size of
the prison population that the BoP administers. Specifically, our analyses of these policy
options will consider the effect on the federal prison population if the policies were
implemented in such a way that the options would apply to inmates currently incarcerated
in the BoP system.
In this memorandum, we provide to you another of these estimates. Here we
provide our estimate of the impact on the federal prison population if the current statutory
requirements for eligibility for "safety valve" relief in drug cases involving a mandatory
minimum penalty) were to be expanded to all offenders regardless of the type of offense
committed, subject to some limitations. Specifically, you have asked us to consider a
change to current law whereby any non-violent offender would be eligible for safety
valve relief regardless of the offender's criminal history, provided that the offender had
not been convicted of any offense for which the sentence imposed exceeded 13 months,
and the courts would be authorized to resentence those offenders in a manner consistent
with the revised statute.
As we discuss more fully below, we estimate that 38,930 offenders who are
incarcerated and would still be in prison as of October 1, 2014 would be eligible for a
I

18 U.S.c. § 3553(f); see also USSG §5CI.2.

Page 124

GAO-14-821 BOP Cost Efficiencies

sentence reduction under such a policy change. The average reduction in sentence for
these offenders would be 15 months (20.5%). The estimated total savings to the BoP
from such a policy change applied retroactively would be 40,552 bed years.

1.

The Policy Option We Considered

One of the policy options that we have agreed to examine on behalf of the GAO is
the impact of a policy change in current law whereby the current statutory requirements
for eligibility for "safety valve" relief from the application of mandatory minimum
penalties in drug cases would be expanded. That provision currently requires courts to
sentence offenders who have been convicted of a drug offense carrying a mandatory
minimum penalty without regard for that penalty when all of several criteria are met. 2
One of those criteria involves the criminal history score3 of the offender. Currently,
offenders are eligible for safety valve relief only if they have been assessed no more than
one criminal history point under the sentencing guidelines. 4
You have asked us to consider a change to the statute whereby the safety valve
would be available to all offenders, not just those committing drug offenses, provided
that they offender was not a violent offender. You also asked us to consider a further
change to the statute so that any such non-violent offender would be eligible regardless of
their criminal history score, provided that none of their criminal history involves a
conviction for which the sentence imposed exceeded one year and one month, 5 and
further provided they meet the remaining safety valve criteria (the "Policy Option").

II.

What We Found

Applying the Policy Option to offenders in the custody of the BoP, we estimate
that there are 38,930 offenders who would be incarcerated on October 1,2014 who would
be eligible for a reduction in their current sentence if the Policy Option were to be
retroactively applied. If the courts were to grant each of these offenders the full reduction
in their sentence we estimate that the total savings to the BoP would be 40,552 bed
years.6
Table 1 provides information on some of the demographic characteristic of the
offenders eligible for a sentence reduction pursuant to the Policy Option. Table 2
2

ld.

3

See USSG ChA, Pt.A, for a discussion on how the criminal history score is determined.

4

Criminal history points are assessed by the court at the time of sentencing. !d.

5 Offenders convicted of an offense for which the sentence imposed exceeded one year and one month are
assessed three points under the sentencing guidelines. See USSG §4A l.l(a). Therefore, this Policy Option
would expand safety valve relief to any offender who did not have any conviction that was assigned three
points under the criminal history provisions of the guidelines.

6 A "bed year" is the cost to the BoP of incarcerating one inmate for one year. For example, one inmate
who serves five years of imprisonment accounts for five bed years.

2
Page 125

GAO-14-821 BOP Cost Efficiencies

provides an estimate of the number of offenders who would be released each year if the
Policy Option was applied retroactively compared to the number of offenders who will be
released each year pursuant to their current sentence. As can be seen on this table, if the
Policy Option were fully retroactive on October 1, 2014, more than 5,623 offenders
would be eligible for immediate release. Within the first year of the effective date of the
Policy Option, another 12,067 offenders would be eligible for release. In total, the
number of offenders eligible for release under the Policy Option in the first year is 6,321
more than the number who will be released under their current sentence. Table 3
provides information on selected sentencing guideline characteristics and the criminal
history category of these offenders.
For the 38,931 offenders who would be eligible to receive a sentence reduction
under the Policy Option, the current average sentence is 73 months. If the courts were to
grant the full reduction possible in each case, the projected new average sentence for
these offenders would be 58 months, a reduction of 15 months (or 20.5%). Based on this
reduction, the estimated total savings to the BoP from the retroactive application of the
Policy Option would be 40,552 bed years. This savings would not be realized in any
single year but is the cumulative savings realized over several years.
III.

How We Conducted This Analysis
A.

Methodology

The methodology for this analysis is based on the Commission's Prison Impact
Model, which has been in use in some form since the guidelines were first developed.
This model is used to estimate the impact of proposed statutory and guideline
amendments on newly sentenced offenders and to project the future impact those
amendments will have on bed space in the BoP. For this analysis, those offenders who
appear to be eligible to receive a reduced sentence were hypothetically "resentenced"
with the computer program as if the proposed policy change had been in effect in the year
in which they were sentenced. A new release date for each offender also was calculated
in order to determine when the offender would be eligible for release if he or she were
provided the full reduction in sentence provided by the proposed policy change.
B.

The Offender Population We Studied

The Bureau of Prisons provided the Commission with a datafile of inmates who
were in the custody of the BoP on Jan 25,2014. That file contained approximately
189,000 offenders. Approximately 184,000 of these offenders were sentenced between
fiscal year 1992 and fiscal 2013. 7 USSC staff was able to match 171,765 of these
offenders to Commission records. Of these, 138,894 were estimated to remain
incarcerated on October 1,2014.
7 The analysis was limited to data from fiscal year 1992 through September 30,2013 (fiscal years 1992
through 2013) because the Commission's data collection efforts prior to fiscal year 1992 were not as
complete as in later years. For example, the Commission did not collect information on the type of drug
involved in drug offenses prior to fiscal year 1992.

3
Page 126

GAO-14-821 BOP Cost Efficiencies

In order to approximate the group of offenders who will be sentenced in fiscal
year 2014, the Commission used the FY2013 datafile and moved all sentence dates
forward by one year. Staff then determined which of these offenders would be
incarcerated on October 1,2014. This process added another 40,178 offenders into the
analysis. Between the two groups, the Commission's analysis included data on 179,072
offenders.
C.

Our Assumptions

In performing our analysis, we have been required to make some assumptions (set
forth below) concerning the decisions that Congress and the courts would make in
determining whether, and to what extent, to reduce the sentences of offenders eligible to
receive a modification of sentence pursuant to the Policy Option discussed above. These
assumptions may not hold in every case.
1.

Procedural assumptions

Specifically, we have assumed that Congress would amend the provisions of 18
U.S.C. § 3553(f) to provide that the subsection would apply at sentencing in any case in
which the offender was not convicted of a violent offense and had not used violence in
the commission ofthe offense. We have also assumed that Congress would amend 18
U.S.C. § 3553(f)(1) to expand eligibility under that section to any federal offender to the
extent that: "the defendant does not have a criminal history score, as determined under
the sentencing guidelines, that includes points added for any conviction for which the
sentence imposed exceeded one year and one month."
We have further assumed that the Commission would amend USSG §5C1.3 to
conform it to the provisions of the amended statutory safety valve provision8 but make no
further changes to the Guidelines Manual, so that the remaining provisions of the 2013
Guidelines Manual would apply in these cases. Further, we have assumed that Congress
would authorize the statutory change to be applied retroactively to offenders incarcerated
in the BoP and that the Commission would amend USSG §IBl.I0 to authorize courts to
apply the amendment to USSG §5C1.3 retroactively to those offenders on some specific
date. 9 We also assumed that the courts would resentence offenders in a manner that is
consistent with the amended Guidelines Manual. 10 Finally, we assumed that the effective

See 28 U.S.C. § 994(0), which requires the Commission to periodically review and revise the sentencing
guidelines.

8

9 See 28 U.S.C. § 994(u), which requires that when the Commission "reduces the term of imprisonment
recommended in the guidelines applicable to a particular offense or category of offenses, it shall specify in
what circumstances and by what amount the sentences of prisoners serving terms of imprisonment for the
offense may be reduced." The Commission does this through amendments to USSG § 1B 1.1 0, which lists
all of the amendments to the Guidelines Manual that may be retroactively applied.

10 See 18 U.S.c. § 3582(c)(2), which authorizes the courts to resentence an incarcerated offender in light of
the Commission's decision under 28 U.S.C. §§ 994(0) and (u) to retroactively apply an amendment to the

4
Page 127

GAO-14-821 BOP Cost Efficiencies

date of any such Policy Option would be October 1, 2014. Therefore, only offenders
incarcerated as of that date would be eligible to seek a reduction in sentence pursuant to
the Policy Option discussed in this memorandum.
2.

Substantive assumptions

In fiscal year 2010, approximately 66 percent of all drug offenders were convicted
of an offense carrying a mandatory minimum penalty. I I Of those, approximately 46
percent remain subject to the penalty at the time of sentencing. 12 Some offenders receive
relief from the effect of any such penalty by providing substantial assistance to the
government in the investigation or prosecution of another offender. 13 Others receive
relief from such a penalty through the statutory "safety valve" provision. 14 Some
offenders are eligible for both provisions.
Under the sentencing guidelines, drug offenders who meet the requirements of the
statutory safety valve provision, which is incorporated into the guidelines at USSG
5Cl.3, are entitled to a two-level reduction in offense level. 15 This provision applies
regardless of whether the offender was convicted of an offense carrying a mandatory
minimum penalty. That is, for offenders who are convicted of an offense carrying a
mandatory minimum penalty, eligibility for the safety valve provides two benefits: 1)
statutory relief from application of the mandatory minimum penalty, and 2) a two-level
reduction in the offense level calculated under the sentencing guidelines. For drug
offenders who were not convicted of an offense carrying a mandatory minimum penalty,
eligibility for safety valve relief also results in the two-level reduction in offense level.
For the analysis described in this memorandum, we have assumed that this
structure would be unchanged, but now would be applied in all cases (not just drug cases)
in which the offender was not convicted of a violent offense and had not used violence in
connection with the offense. As a result, all non-violent offenders could benefit from the
Policy Option if they meet the revised safety valve criteria and regardless of whether a
mandatory minimum penalty had ever applied in the case.
We made several assumptions in for the purpose of determining which offenders
were violent offenders. Specifically, we considered any ofthe following to be a violent
offense or an indication that violence had been used in connection with the offense:
Guidelines Manual if the sentencing range that applied at the time the offender was sentenced would be
lower under the amended guideline.
11 See U.S. SENTENCING COMMISSION, MANDATORY MfNIMUM PENALTIES fN THE FEDERAL CRIMINAL
JUSTICE SYSTEM 154 (2011).

12

!d.

13

See 18 U.S.C. § 3553(e); USSG §5Kl.1.

14

See 18 U.S.c. § 3553(f); USSG §5K3.1.

15

USSG §2Dl.1(b)(16).

5
Page 128

GAO-14-821 BOP Cost Efficiencies

the guideline applied in the case was from Part A of Chapter 2 of the Guidelines
Manual (which involves crimes against the person);
the guideline applied was USSG §2B2.1 (burglary of a residence);
the guideline applied was USSG §2B3.1 (robbery);
the guideline applied was USSG §2B3.2 (extortion by force or threat of injury);
the guideline applied was USSG §2Dl.9 (placing dangerous devices to protect
drug production);
the guideline applied was from Subpart 1 of Part E of Chapter 2 (racketeering);
the guideline applied was from Subpart 1 of Part G of Chapter 2 (promoting
commercial sex acts or other prohibited sexual conduct);
the guideline applied was from Subpart 2 of Part G of Chapter 2 and the conduct
involved the production of child pornography;
the guideline applied was USSG §2H4.1 (peonage, slave trade);
the guideline applied was from Part K of Chapter 2 (firearms);
the guideline applied was from Subparts 1,2,3,5, or 6 of Part M of Chapter 2
(treason, sabotage, espionage, and cases involving nuclear, biological and chemical
weapons and materials);
the guideline applied was USSG §2Nl.l (tampering with consumer products with
risk of death);
the guideline applied was USSG §2P 1.3 (prison riot);
the guideline applied was USSG §2Q 1.1 (knowing endangerment from
HAZMATs);
the guideline applied was USSG §2Ql.4 (tampering with public water system); or
the guideline applied was USSG §2X6.1 (use of a minor in a crime of violence).
Additional assumptions we have made for this analysis are that the sentence for
each offender would be reduced based on the maximum good conduct credit allowed by
the Bureau of Prisons, and offenders would serve the lesser of the newly calculated
sentence or their life expectancies.
If you have any questions about our analysis, please do not hesitate to contact the
Commission. We continue to work to analyze the other policy options that we discussed
with you and your colleagues. We will provide our analyses of these options as they are
completed.
Sincerely,

.~ ~m,, /7fNoah D. Bookbinder
Director
Office of Legislative and Public Affairs

Glenn R. Schmitt
Director
Office of Research and Data

Attachments
6
Page 129

GAO-14-821 BOP Cost Efficiencies

Table 1
Demographic Characteristics of Eligible Offenders
(FY1992 through FY2014)

DEMOGRAPHICS
Race/Ethnicity
White
Black
Hispanic
Other
Total

13,640
6,540
17,586
997
38,763

35 .2%
16.9%
45.4%
2.6%
100.0%

U.S. Citizen
Non-Citizen
Total

25,707
13,211
38,918

66.1%
33.9%
100.0%

Male
Female
Total

34,574
4,341
38,915

88.8%
11.2%
100.0%

38
(as of November 1,2014)

37
(at sentencing)

Citizenship

Gender

Average Age

The analysis involves a total of38,931 cases, however, cases missing information for any specific analysis are
excluded from that analysis.
Total percentages may not add to exactly 100% due to rounding.
SOURCE: U.S. Sentencing Commission, 1992 - 2013 Datafiles, USSCFY92 - USSCFY13.

Page 130

GAO-14-821 BOP Cost Efficiencies

Table 2
Projected Year of Release for Retroactive Eligible Offenders
(FY1992 through FY2014)

Release Date

IF AMENDMENT
RETROACTIVE

IF AMENDMENT
NOT
RETROACTIVE

N

N

Immediate Release

5,623

within 1 yr

12,067

11,369

within 2 yr

7,113

7,605

within 3 yr

4,724

5,476

within 4 yr

3,168

4,115

within 5 yr

1,975

2,816

within 6 yr +

4,260

7,549

Of the 38,931 offenders who appear to be eligible for relief under the amendment, Commission
records contained sufficient information to perform this analysis for 38,930 offenders.
SOURCE: U.S. Sentencing Commission, 1992 - 2013 Datafiles, USSCFY92 - USSCFY13.

Page 131

GAO-14-821 BOP Cost Efficiencies

Table 3
Guideline Sentencing Characteristics, Criminal History, and
Position Relative to the Guideline Range of Eligible Offenders
(FY1992 through FY2014)

CHARACTERISTICS
Average Base Offense Level
Weapon Specific Offense Characteristic
Firearms Mandatory Minimum Applied
Safety Valve §5C1.2
Aggravating Role §3B1.1
Mitigating Role §3B1.2
Obstruction Adjustment §3C1.1
Career Offender Status §4B1.1
Criminal

0
0
0
0
2,J 08
0
845

0.0%
0.0%
0.0%
0.0%
5.4%
0.0%
2.2%

17,290
7,895
8,492
2,848
1,095
1,311
38,931

44.4%
20.3%
21.8%
7.3%
2.8%
3.4%
100%

20,748
823
4,895
12,438
38,904

53.3%
2.1%
12.6%
32.0%
100%

Histor~ C ategor~

I
II

III
IV
V
VI
Total
Sentence Relative to the Guideline Range
Within Range
Above Range
Substantial Assistance §5K1.1
Otherwise Below Range
Total

The analysis involves a total of 38,931 cases, however, cases missing information for any specific
analysis are excluded from that analysis.
Total percentages may not add to exactly 100% due to rounding.
SOURCE: U.S . Sentencing Commission, 1992 - 2013 Datafiles, USSCFY92 - USSCFYI3 .

Page 132

GAO-14-821 BOP Cost Efficiencies

Appendix VIII: GAO Contact and Staff
Acknowledgments
Appendix VIII: GAO Contact and Staff
Acknowledgments

GAO Contact

David C. Maurer, (202) 512-9627 or MaurerD@gao.gov.

Staff
Acknowledgments

In addition to the contact named above, Jill Verret (Assistant Director),
Pedro Almoguera, Billy Commons, Adam Couvillion, Michele Fejfar, Eric
Hauswirth, Susan Hsu, Linda Miller, Michael Tropauer, and Jonathan R.
Tumin made key contributions to this report.

(441173)

Page 133

GAO-14-821 BOP Cost Efficiencies

GAO’s Mission

The Government Accountability Office, the audit, evaluation, and
investigative arm of Congress, exists to support Congress in meeting its
constitutional responsibilities and to help improve the performance and
accountability of the federal government for the American people. GAO
examines the use of public funds; evaluates federal programs and
policies; and provides analyses, recommendations, and other assistance
to help Congress make informed oversight, policy, and funding decisions.
GAO’s commitment to good government is reflected in its core values of
accountability, integrity, and reliability.

Obtaining Copies of
GAO Reports and
Testimony

The fastest and easiest way to obtain copies of GAO documents at no
cost is through GAO’s website (http://www.gao.gov). Each weekday
afternoon, GAO posts on its website newly released reports, testimony,
and correspondence. To have GAO e-mail you a list of newly posted
products, go to http://www.gao.gov and select “E-mail Updates.”

Order by Phone

The price of each GAO publication reflects GAO’s actual cost of
production and distribution and depends on the number of pages in the
publication and whether the publication is printed in color or black and
white. Pricing and ordering information is posted on GAO’s website,
http://www.gao.gov/ordering.htm.
Place orders by calling (202) 512-6000, toll free (866) 801-7077, or
TDD (202) 512-2537.
Orders may be paid for using American Express, Discover Card,
MasterCard, Visa, check, or money order. Call for additional information.

Connect with GAO

Connect with GAO on Facebook, Flickr, Twitter, and YouTube.
Subscribe to our RSS Feeds or E-mail Updates. Listen to our Podcasts.
Visit GAO on the web at www.gao.gov.

To Report Fraud,
Waste, and Abuse in
Federal Programs

Contact:
Website: http://www.gao.gov/fraudnet/fraudnet.htm
E-mail: fraudnet@gao.gov
Automated answering system: (800) 424-5454 or (202) 512-7470

Congressional
Relations

Katherine Siggerud, Managing Director, siggerudk@gao.gov, (202) 5124400, U.S. Government Accountability Office, 441 G Street NW, Room
7125, Washington, DC 20548

Public Affairs

Chuck Young, Managing Director, youngc1@gao.gov, (202) 512-4800
U.S. Government Accountability Office, 441 G Street NW, Room 7149
Washington, DC 20548

Please Print on Recycled Paper.