Drug Offenders: Various Factors May Limit the Impacts of Federal 
Laws That Provide for Denial of Selected Benefits (26-SEP-05,	 
GAO-05-238).							 
                                                                 
Several provisions of federal law allow for or require certain	 
federal benefits to be denied to individuals convicted of drug	 
offenses in federal or state courts. These benefits include	 
Temporary Assistance for Needy Families (TANF), food stamps,	 
federally assisted housing, postsecondary education assistance,  
and some federal contracts and licenses. Given the sizable	 
population of drug offenders in the United States, the number and
the impacts of federal denial of benefit provisions may be	 
particularly important if the operations of these provisions work
at cross purposes with recent federal initiatives intended to	 
ease prisoner reentry and foster prisoner reintegration into	 
society. GAO analyzed (1) for selected years, the number and	 
percentage of drug offenders that were estimated to be denied	 
federal postsecondary education and federally assisted housing	 
benefits and federal grants, contracts, and licenses and (2) the 
factors affecting whether drug offenders would have been eligible
to receive TANF and food stamp benefits, but for their drug	 
offense convictions, and for a recent year, the percentage of	 
drug offenders released who would have been eligible to receive  
these benefits. Several agencies reviewed a draft of this report,
and we incorporated the technical comments that some provided	 
into the report where appropriate.				 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-05-238 					        
    ACCNO:   A38257						        
  TITLE:     Drug Offenders: Various Factors May Limit the Impacts of 
Federal Laws That Provide for Denial of Selected Benefits	 
     DATE:   09/26/2005 
  SUBJECT:   Convictions					 
	     Criminals						 
	     Data collection					 
	     Drugs						 
	     Eligibility criteria				 
	     Eligibility determinations 			 
	     Federal aid programs				 
	     Federal law					 
	     Higher education					 
	     Housing programs					 
	     Licenses						 
	     Statistical data					 
	     Food Stamp Program 				 
	     Temporary Assistance for Needy Families		 
	     Program						 
                                                                 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-05-238

United States Government Accountability Office

GAO

                       Report to Congressional Requesters

September 2005

DRUG OFFENDERS

Various Factors May Limit the Impacts of Federal Laws That Provide for Denial of
                               Selected Benefits

GAO-05-238

[IMG]

September 2005

DRUG OFFENDERS

Various Factors May Limit the Impacts of Federal Laws That Provide for Denial of
Selected Benefits

                                 What GAO Found

For the years for which it obtained data, GAO estimates that relatively
small percentages of applicants but thousands of persons were denied
postsecondary education benefits, federally assisted housing, or selected
licenses and contracts as a result of federal laws that provide for
denying benefits to drug offenders. During academic year 2003-2004, about
41,000 applicants (or 0.3 percent of all applicants) were disqualified
from receiving postsecondary education loans and grants because of drug
convictions. For 2003, 13 of the largest public housing agencies in the
nation reported that less than 6 percent of 9,249 lease terminations that
occurred in these agencies were for reasons of drug-related criminal
activities-such as illegal distribution or use of a controlled
substance-and 15 large public housing agencies reported that about 5
percent of 29,459 applications for admission were denied admission for
these reasons. From 1990 through the second quarter of 2004, judges in
federal and state courts were reported to have imposed sanctions to deny
benefits such as federal licenses, grants, and contracts to about 600
convicted drug offenders per year.

Various factors affect which convicted drug felons are eligible to receive
TANF or food stamps. This is because state of residence, income, and
family situation all play a role in determining eligibility. Federal law
mandates that convicted drug felons face a lifetime ban on receipt of TANF
and food stamps unless states pass laws to exempt some or all convicted
drug felons in their state from the ban. At the time of GAO's review, 32
states had laws exempting some or all convicted drug felons from the ban
on TANF, and 35 states had laws modifying the federal ban on food stamps.
Because of the eligibility requirements associated with receiving these
benefits, only those convicted drug felons who, but for their conviction,
would have been eligible to receive the benefits could be affected by the
federal bans. For example, TANF eligibility criteria include requirements
that an applicant have custodial care of a child and that income be below
state-determined eligibility thresholds. Available data for 14 of 18
states that fully implemented the ban on TANF indicate that about 15
percent of drug offenders released from prison in 2001 met key eligibility
requirements and constitute the pool of potentially affected drug felons.
Proportionally more female drug felons than males may be affected by the
ban, as about 27 percent of female and 15 percent of male drug offenders
released from prison in 2001 could be affected.

Federal Benefits That May Be Denied to Drug Offenders

Federal benefit Description

TANF Cash assistance designed to meet a needy family's ongoing basic needs

Food stamps Food assistance payments to low-income households

Postsecondary
education Federal Pell Grants, Stafford loans, and work-study assistance

Federally assisted Public housing primarily for low-income families with
children and housing vouchers for private-market assistance for very
low-income families

Denial of Federal Federal postsecondary student loans, federal licenses
(e.g., for
Benefits Program physicians, pilots, and others), and procurement
contracts, among others

Source: GAO analysis of federal law.

United States Government Accountability Office

Contents

Letter

Results in Brief
Background
Estimated Number of Drug Offenders Denied Federal

Postsecondary Education, Housing, and Certain Contracts and
Grants Varies by Type of Benefit

State Exemptions to the Federal Bans on TANF and Food Stamps
and Other Factors Affect the Percentage of Drug Offenders That
Would Have Been Eligible to Receive the Benefits

Concluding Observations
Agency Comments and Our Evaluation

                                       1

                                      3 5

12

17 21 22

Appendix I

Objectives, Scope, and Methodology

Assessing How Many Drug Offenders Were Estimated to Be Denied
Federal Postsecondary Education and Federally Assisted
Housing Benefits and Certain Grants and Contracts

Estimating the Percentage of Convicted Drug Felons That Would
Have Been Eligible to Receive TANF or Food Stamps and the
Factors Contributing to These Estimates

Data Sources Used in Preparing This Report

                                       24

                                       25

                                     26 27

                                  Appendix II

Denial of TANF and Food Stamps Benefits Legal
Framework and Methodologies

Legal and Administrative Framework
State Exemptions to the Federal Bans on TANF and Food Stamps
Methods for Estimating the Percentage of Drug Felons Potentially

Affected by the TANF and Food Stamp Bans
Data on the Race of Drug Offenders Who Meet Selected TANF
Eligibility Criteria

                                       30

                                     30 32

                                       39

                                       49

Appendix III Denial of Federal Postsecondary Education Benefits

to Drug Offenders 51

Legal Framework for Denying Federal Higher Education Benefits 51
Administration of the Denial of Federal Higher Education Benefits 52
Methods for Estimating the Numbers of Persons and Amounts of

Federal Assistance Affected by the Provisions 53
Estimates of the Numbers and Amounts of Federal Assistance
Affected by Controlled Substances Convictions 57
Limited Data on the Long-Term Impacts of Denying Federal
Postsecondary Education Assistance 59

Limited Data on Racial Minorities and the Denial of Federal Postsecondary
Education Benefits 61

Appendix IV Denial of Federally Assisted Housing Benefits for Drug-Related
Criminal Activity

Legal Framework Relating to the Denial of Housing Benefits for
Drug-Related Criminal Activity HUD and PHA Roles in Denying Federally
Assisted Housing

Benefits for Reasons of Drug-Related Criminal Activities Number of Denials
of Federally Assisted Housing Benefits Several Factors Can Contribute to
the Reported Low Number of

Denials because of Drug-Related Criminal Activities, but Potential Impacts
Are Unclear Data on the Race of Persons Denied Federally Assisted Housing
Were Infrequently Provided

                                       62

                                       62

                                     64 66

                                     76 77

Appendix V	Legal and Administrative Framework for the Denial of Federal
Benefits Program

About the Law
Administration of the Law
Use of the Denial of Federal Benefits Program Provisions
Assessing the Impacts of the Denial of Federal Benefits Program
Reliable Data by Race of Persons Receiving the DFB Are Not

Available 78

78 80 81 83

84

Appendix VI GAO Contact and Staff Acknowledgments

Tables

Table 1: Provisions of Selected Federal Law and the Corresponding Benefits
That May Be Denied to Certain Drug Offenders 5 Table 2: Selected Federal
Legal Provisions Providing for Denial of Benefits to Drug Offenders 8

Table 3: General Eligibility Criteria to Receive Selected Federal Benefits
and Federal Agencies Responsible for Administering Programs 10

Table 4: Number and Percentage of Tenants and Applicants for Federally
Assisted Housing Benefits That Were Denied Benefits for Reasons of
Drug-Related Criminal Activities in Selected Large PHAs in 2003 14

Table 5: Status of States on the TANF and Food Stamp Ban as of 2004 33

Table 6: Types of State Exemptions (Modifications) to the Federal Ban on
TANF, by States within the Types of Modifications, as of 2004 35

Table 7: Types of State Exemptions or Modifications of the Federal Ban on
Food Stamps, by States within the Types of Modifications, as of 2004 37

Table 8: Number of States in Each TANF and Food Stamp Ban Status Category
and Percentage of Drug Offense Arrests in 2002 Occurring in the States
within Each Ban Status Category 38

Table 9: Estimated Number of Drug Felons Released from Prison during 2001,
in 14 of 18 States That Currently Maintain the Ban on TANF 42

Table 10: Estimated Parental Status of Drug Offenders, by Gender and
Region of the Country 43

Table 11: Estimated Percentages of Drug Offenders Released from Prison in
2001 That Met Selected Guardianship and Earned Income Limits to Receive
TANF benefits 45

Table 12: Periods of Ineligibility for Federal Higher Education Assistance
Based on Type and Number of Controlled Substance Offense Convictions 52

Table 13: Data Used to Estimate the Number of Persons Denied Federal
Education Assistance because of Drug Convictions and Estimated Amount of
Federal Assistance Denied 56

Table 14: Estimated Number of Persons Denied Federal Education Assistance
because of Drug Convictions and Estimated Amount of Federal Assistance
Affected 57

Table 15: Evictions from and Denials of Admission into Public Housing for
Reasons of Criminal Activity, Fiscal Years 2002 and 2003 67

Table 16: Public Housing Lease Terminations at Selected Public Housing
Agencies, by Reason for Termination (Criminal Activities and Drug-Related
Criminal Activities), 2003 69

Table 17: Housing Choice Voucher Program Terminations of Assistance at
Selected Public Housing Agencies, by Reason for Termination (Criminal
Activities and Drug-Related Criminal Activities) of Assistance, 2003 71

Table 18: Selected Public Housing Agencies' Actions Reported on
Applications for Admission and Denials of Admission into Public Housing
for Reasons of Criminal Activities and Drug-Related Criminal Activities,
2003 73

Table 19: Housing Choice Voucher Program, Selected Public Housing
Agencies' Actions Reported on Applications for Assistance and Denials of
Assistance for Reasons of Criminal Activities and Drug-Related Criminal
Activities, 2003 75

Table 20: Periods of Ineligibility for Which Benefits Can or Must Be
Denied under Section 5301 of the Anti-Drug Abuse Act, as Amended 79

Abbreviations

ABAWD Able-Bodied Adults without Dependent
ACF Administration for Children and Families
AOUSC Administrative Office for United States Courts
BJA Bureau of Justice Assistance
BJS Bureau of Justice Statistics
DOJ Department of Justice
DFB Denial of Federal Benefits
ED Department of Education
EFC Expected Family Contribution
FAFSA Free Application for Federal Student Aid
FNS Food and Nutrition Service
GED general equivalency degree
HCV Housing Choice Voucher
HEA Higher Education Act
HHS Department of Health and Human Services
HUD Department of Housing and Urban Development
INDCP Office for National Drug Control Policy
MASS Management Operations Certification Assessment System
NCRP National Corrections Reporting Program
PHA public housing agency
PRWORA Personal Responsibility and Work Opportunity

Reconciliation Act of 1996 TANF Temporary Assistance for Needy Families
USDA U.S. Department of Agriculture USSC United States Sentencing
Commission

This is a work of the U.S. government and is not subject to copyright
protection in the United States. It 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.

United States Government Accountability Office Washington, DC 20548

September 26, 2005

The Honorable Robert C. Scott
Ranking Minority Member
Subcommittee on Crime, Terrorism, and Homeland Security
Committee on the Judiciary
House of Representatives

The Honorable Bobby L. Rush
House of Representatives

During recent years, several hundred thousand persons per year in the
United States were convicted in federal and state courts of drug offenses
such as felony drug trafficking and misdemeanor drug possession.
Provisions of federal law allow that these convictions may render these
individuals ineligible to receive selected federal benefits. The benefits
include Temporary Assistance for Needy Families (TANF), food stamps,
federally assisted housing, postsecondary education grants and loans, and
certain federal contracts and licenses. Depending upon the federal
benefit,
drug offenders may be prohibited from receiving assistance for periods
that range from 1 year to life.1

You asked us to assess the impacts of these provisions. An important first
step in assessing the impacts is to determine how many people could be
affected by these provisions. The number and the impacts of federal denial
of benefit provisions may be particularly important if the operations of
these provisions work at cross purposes with recent federal initiatives
intended to ease prisoner reentry and foster prisoner reintegration into
society. This report analyzes two interrelated questions about the number
or percentage of drug offenders that could be affected by the provisions:

(1) In specific years, how many drug offenders were estimated to be
denied federal postsecondary education benefits, federally assisted

1In this report, we use the term "drug offenders" broadly to refer to any
person that may be denied federal benefits under the provisions that allow
or require them to be denied specifically to drug offenders. This term
applies whether the offender was convicted of a felony or misdemeanor drug
offense, or whether the offender engaged in a type of drugrelated criminal
activity that could nonetheless result in the loss of federally assisted
housing benefits.

housing, and selected grants, contracts, and licenses? (2) What factors
affect whether drug offenders would have been eligible to receive TANF and
food stamp benefits, but for their drug offense convictions, and for a
recent year, what percentage would have been eligible to receive these
benefits?

You also asked us to address the impact of federal benefit denial laws on
racial minorities and the long-run consequences of having benefits denied.
The final sections of appendices III, IV, and V in this report include
discussions of the data limitations related to estimating the impacts on
racial minorities. Where information was available, we also identify in
the appendices some of the possible long-run consequences of denial of
benefits.

To estimate the extent to which drug offenders were denied federal
postsecondary education, federally assisted housing benefits, and selected
federal licenses, grants, and contracts, we obtained and analyzed data
from several agencies. We analyzed Department of Education data on persons
who applied for federal postsecondary education assistance such as Pell
Grants and who reported that they had convictions that prohibited them
from receiving federal postsecondary education benefits.2 We analyzed
Bureau of Justice Statistics (BJS) data on the educational attainment of a
nationally representative sample of drug offenders on probation. From
officials at 17 large public housing agencies (PHA) we obtained
information about their experiences in denying federally assisted housing
benefits to persons because of their drug-related criminal activities.3 We
analyzed data on evictions and denials of admission into public housing
that the Department of Housing and Urban Development (HUD) obtains from
PHAs that manage units in the Public Housing Program. From the Bureau of
Justice Assistance (BJA) we obtained data on persons reported to be denied
selected federal licenses, grants, and contracts by federal and state
judges.

To estimate the extent to which convicted drug felons would have been
eligible to receive TANF and food stamp benefits apart from their felony

2The Pell Grant program provides grants (i.e., aid that does not need to
be repaid) to needy undergraduates.

3Our review covered both the Public Housing Program and the Housing Choice
Voucher Program. Not all 17 of the PHAs provided responses to all of our
questions. Below, we report data from the PHAs that provided relevant
information.

drug offense conviction, we obtained and analyzed the most recently
available data compiled by BJS on the familial and economic
characteristics of drug offenders in prison. We analyzed these data to
estimate the number of drug offenders that met selected eligibility
criteria to receive these benefits, and we applied our estimates from the
BJS prisoner data to data on drug offenders released from prison in 14 of
18 states that fully implemented the federal bans on TANF and food stamps.
Because of the limited data on persons actually denied TANF and food stamp
benefits, we provide rough estimates of the proportions of drug offenders
that met selected eligibility criteria to receive these benefits rather
than of the number of such offenders actually denied these benefits.

To identify the various factors that contributed to our estimates of those
potentially affected by federal provisions that provide for denial of
benefits, we interviewed officials from the federal agencies that
administer the federal benefits that may be denied, researchers who have
studied the denial of federal benefits and related issues, and officials
of various associations who represent PHAs, which are the agencies that
manage federally assisted benefits. We also reviewed published studies on
the denial of federal benefits.

We assessed the reliability of the data that we used in preparing this
report by, as appropriate, interviewing agency officials about their data,
reviewing documentation about the data sets, and conducting electronic
tests. We used only the portions of the data that we found to be
sufficiently reliable for our purposes in this report. We provide a more
detailed discussion of our objectives, scope, and general methodology in
appendix I. In appendices II through V, we further discuss our
methodologies for each of the federal benefits that we reviewed.

We conducted our work from March 2004 through July 2005 in accordance with
generally accepted government auditing standards.

Relatively small percentages of applicants but thousands of persons were
reported to be denied postsecondary education benefits, federally assisted
housing, or selected licenses and contracts as a result of federal laws
that provide for denying benefits to drug offenders during the years for
which we had data. In relation to postsecondary education loans and
grants, during academic year 2003-2004-one of the years in our
review-about 41,000 applicants (about 0.3 percent of total applicants)
either reported that they had drug offense convictions that disqualified
them from receiving assistance or left the question blank, which also
disqualified

Results in Brief

them from receiving federal assistance. In relation to public housing,
among 13 of the largest PHAs in the nation that responded to our request
for information about both lease terminations and reasons for denial of
housing assistance that occurred during 2003, less than 6 percent of the
more than 9,200 lease terminations that occurred in these agencies were
reportedly for reasons of drug-related criminal activities-such as illegal
distribution or use of a controlled substance. About 5 percent of the
roughly 29,000 applications for admission into public housing in 15 of the
largest PHAs in the nation were reportedly denied admission for reasons of
drug-related criminal activities. From 1990 through the second quarter of
2004, sentencing judges in federal and state courts reported that they
denied benefits such as licenses, grants, and contracts to an average of
about 600 convicted drug offenders per year.

Various factors affect which convicted felony drug offenders may be
eligible to receive TANF or food stamps. This is because state of
residence, income, and family situation all play a role in determining
eligibility. Federal law mandates that convicted felony drug offenders
face a lifetime ban on receipt of TANF and food stamps unless states pass
laws to exempt some or all such drug felons from the ban. At the time of
our review, 32 states had laws exempting some or all convicted felony drug
offenders from the ban on TANF, and 35 states had done so for the ban on
food stamps. TANF eligibility requires that an applicant have custodial
care of a child and that income be below eligibility thresholds. For food
stamps, income and work requirements are used to determine eligibility.
Thus, only those drug offenders who meet these requirements could be
denied benefits because of their convictions. From available data for 14
of the 18 states that fully implemented the ban on TANF, we estimated that
about 15 percent of drug offenders released from prison in 2001 met key
eligibility requirements and constitute the pool of potentially affected
felony drug offenders. The custodial parent requirement results in
proportionally more female drug felons potentially affected by the TANF
ban than males. For food stamps, we estimated that in 12 of the 15 states
that fully implemented the federal ban on food stamps, about 23 percent of
drug offenders released from prison during 2001 were custodial parents of
minor children with incomes that could make them eligible to receive food
stamps.

We provided a draft of this report to the Attorney General; the
Secretaries of the Departments of Education, Agriculture, and Housing and
Urban Development; the Assistant Secretary of the Administration for
Children and Families, the Director of the Office of National Drug Control
Policy; the Research Director of the United States Sentencing Commission;
and

the Director of the Administrative Office of the United States Courts for
their review and comment. We received technical comments from the
Departments of Justice, Agriculture, and Education, and from the United
States Sentencing Commission and the Administrative Office of the United
States Courts, which we incorporated into the report where appropriate.

Background 	Selected provisions of federal law explicitly prohibit
specific categories of drug offenders from receiving certain federal
benefits for specified periods.4 Table 1 identifies key provisions of
federal law that provide for denial of benefits specifically to drug
offenders and the corresponding benefits that may or must be denied to
drug offenders.

Table 1: Provisions of Selected Federal Law and the Corresponding Benefits
That May Be Denied to Certain Drug Offenders

Federal programs identified in provisions allowing for denial of benefits

Description of the benefits that may or must be denied

                           Provisions of federal law

    Section 115 of the Personal Temporary Assistance for Needy Families Cash
                     assistance, vouchers, and other forms

Responsibility and Work Opportunity Reconciliation Act of 1996a (21 U.S.C.
S: 862a) of support designed to meet a needy family's ongoing basic needs

Food Stamp Program	Food assistance payments to low-income households and
those transitioning from welfare to work.

Section 438 of the Higher Education Act Postsecondary education assistance
Federal Pell Grants, Stafford loans, and
Amendments of 1988 (20 U.S.C. S: work-study assistance.
1091(r))b

Section 5101 of the Anti-Drug Abuse Act of 1988 (42 U.S.C. S:
1437d(l)(6));c Section 428 of the Housing and Urban Development
Appropriations Act for fiscal year 1999 (42 U.S.C. S: 1437n(f)); and
section 577 of the Quality Housing Work and Responsibility Act of 1998 (42
U.S.C. S: 13662) Federally assisted housing: Low-rent Public Housing
Program

Federally assisted housing: Housing Choice Voucher Program (formerly known
as Section 8 Housing) Public housing primarily for low-income families
with children.

Private market housing assistance for very low-income families, the
elderly, and the disabled.

4In addition to federal provisions explicitly denying federal benefits for
drug offenders, other federal and state provisions may restrict convicted
drug offenders from accessing certain types of occupational opportunities,
participating in jury service, obtaining drivers' licenses, or exercising
voting rights.

Federal programs identified in provisions Description of the benefits that
may or Provisions of federal law allowing for denial of benefits must be
denied

Section 5301 of the Anti-Drug Abuse Act Denial of Federal Benefits Program
Federal postsecondary student loans,

of 1988 (21 USC S: 862) 	federal licenses (e.g., for physicians, pilots,
and others), and procurement contracts, among other benefits

Source: GAO analysis of federal law.

aPub. L. No. 104-93, 110 Stat. 2105 (1996).

bPub. L. No. 105-244, 112 Stat. 1581 (1998).

cPub. L. No. 100-690, 102 Stat. 4181 (1988).

dPub. L. No. 105-276, 112 Stat. 2461, 2511, 2640 (1998). See appendix V
for details about other federal provisions that affect the denial of
federally assisted housing benefits.

Except for federal licenses, procurement contracts, and grants under
Denial of Federal Benefits Program, the benefits that may or must be
denied are benefits that are generally provided to low-income individuals
and families. TANF, food stamps, federally assisted housing, and Pell
Grants are low-income programs. The Denial of Federal Benefits Program,
established under Section 5301 of the Anti-Drug Abuse Act of 1988, as
amended, provides that federal and state court judges may deny all or some
of certain specified federal benefits to individuals convicted of drug
trafficking or drug possession offenses involving controlled substances.
Additional details on each of the programs may be found in appendices II,
III, IV, and V.

The provisions differ on key elements. For example, they establish
different classes of drug offenders that may or must be denied benefits,
and they provide for different periods that drug offenders are rendered
ineligible to receive a benefit and whether or not benefits can be
restored. Some of the provisions allow that drug offenders may become
eligible for benefits upon completing a recognized drug treatment program.

Provisions established by the Personal Responsibility and Work Opportunity
Reconciliation Act of 1996 (PRWORA), as amended, which govern the TANF and
food stamp programs, provide that benefits must be denied to persons
convicted of a state or federal felony drug offense that involves the
possession, use, or distribution of a controlled substance that occurred
after August 22, 1996, the effective date of these provisions. Students
become ineligible to receive federal postsecondary education benefits upon
a conviction of either a misdemeanor or a felony controlled substances
offense. Loss of federally assisted housing benefits can occur if
individuals, relatives in their household, or guests under a tenant's
control engage in drug-related criminal activity, regardless of whether
the activity

resulted in a conviction.5 Local public housing authorities, which
administer federally assisted housing benefits, have discretion in
determining the behaviors that could lead to loss of certain federal
housing benefits. Under the Denial of Federal Benefits Program, judges in
federal and state courts may deny a range of federal licenses, contracts,
and grants to persons convicted of controlled substances drug trafficking
and drug possession offenses.

The period of ineligibility to receive benefits varies. Under PRWORA, as
amended, unless states enact laws that exempt convicted drug offenders in
their state from the federal ban, TANF and food stamp benefits are
forfeited for life for those convicted of disqualifying drug offenses.
State laws may also result in a shorter period of denial of these
benefits. Students are disqualified from receiving federal postsecondary
education benefits for varying periods depending on the number and type of
disqualifying drug offense convictions. A first conviction for possession
of a controlled substance, for example, results in a 1-year period of
ineligibility, while a first conviction for sale of controlled substance
results in a 2-year period of ineligibility. Upon subsequent convictions,
the period of ineligibility can extend indefinitely.

Federally assisted housing benefits may also be denied for varying periods
of time, depending upon the number and types of drug-related criminal
activities. The minimum loss of benefit is 3 years in certain
circumstances, and the maximum is a lifetime ban. For example, for persons
convicted of certain methamphetamine offenses, the ban is mandatory and
for life. Under the Denial of Federal Benefits Program, the denial of
certain other types of benefits by judges, such as federal grants and
contracts, can range from 1 year to life depending on the type of offense
and number of convictions.

In some cases, the period of benefit ineligibility may be shortened if
offenders complete drug treatment. For example, students may have their
postsecondary education benefits restored if they satisfactorily complete
a drug treatment program that satisfies certain criteria and includes two
unannounced drug tests. Under the Denial of Federal Benefits Program, the
denial of benefits penalties may, for example, be waived if a person

5The term "drug-related criminal activity" means the illegal manufacture,
sale, distribution, use, or possession with intent to manufacture, sell,
distribute, or use a controlled substance-as defined in section 102 of the
Controlled Substances Act (21 U.S.C. S: 802), 42 U.S.C. S:1437d(1).

successfully completes a drug treatment program. Other than offenders who
were convicted of methamphetamine offenses, drug offenders that
successfully complete drug treatment may receive federally assisted
housing benefits prior to the end of their period of ineligibility. In
states that have passed laws so specifying, drug offenders may shorten the
period of ineligibility for TANF and food stamp benefits by completing
drug treatment. (See table 2.)

Table 2: Selected Federal Legal Provisions Providing for Denial of
Benefits to Drug Offenders

                                                             Conditions under 
                                                                 which either 
                                                         the loss of benefits 
                                                         may be               
                    Drug offense that may                limited or benefits  
                                  lead to                       may be        
                                            Period of                         
Federal benefit   denial of benefits   benefit denial       restored

TANF   Persons convicted of Lifetime, upon a first  Loss of benefit is     
                  felonies for         felony          limited when           
          certain controlled     conviction, unless       offenders reside in 
              substance               offenders                  states whose 
         related offenses that  reside in states that  exemptions either opt  
                      occurred       have opted        out of the             
        after August 22, 1996  out of or modified the  federal ban or provide 
                               federal ban                      for           
                                                           restoration of     
                                                              benefits        

Food stamps Persons convicted of   Lifetime, upon a    Loss of benefit is  
                       felonies for     first felony      limited when        
                certain controlled   conviction, unless   offenders reside in 
                    substance             offenders              states whose 
                 related offenses   reside in states that exemptions either   
                  that occurred          have opted       opt out of the      
                 after August 22,   out of or modified      federal ban or    
                       1996         the federal ban           provide for     
                                                            restoration of    
                                                               benefits       

Postsecondary education Students convicted of certain Varies from 1 year
to an indefinite Benefits may be restored after assistance misdemeanor or
felony controlled period for repeated convictions completion of drug
treatment or if

substance-related offenses 	the conviction is reversed, set aside, or
otherwise rendered nugatory

Federally assisted housing 	Varies, but includes applicants, Varies, but a
minimum of 3 years tenants, their household family for those with a prior
eviction for members, and guests under their drug-related criminal
activities control who engage in drug-and mandatory lifetime denial for
related criminal activities whether persons convicted of certain convicted
or not methamphetamine offenses

Certain federal licenses, Persons convicted of controlled Varies from 1
year to life,
contracts, loans, grants, substance trafficking and depending upon the
type of
and other federal benefits, possession offenses in federal or offense and
number of
under the Denial of Federal state courts convictions
Benefits Program

Benefits may be restored after completion of drug treatment, with the
exception of those persons convicted of certain methamphetamine offenses

Benefits may be restored after the individual has completed a drug
treatment program, has otherwise been rehabilitated, or has made good
faith efforts to participate in treatment but is unable to complete
treatment because of lack of programs or ability to pay

Source: GAO analysis of selected provisions of federal law and agency
documents.

The legislative history of these provisions is silent as to whether they
were intended to do more than provide for denying federal benefits to drug
offenders, such as deterring drug offenders from committing future

criminal acts. For example, our 1992 report indicated that in the floor
debates over the Denial of Federal Benefits Program, some members of
Congress expressed the opinion that even casual drug use should result in
serious consequences, such as the loss of federal benefits.6 With respect
to prohibiting drug offenders from public housing, congressional findings
made in 1990 and amended in 1998 address the extent of drug-related
criminal activities in public housing and the federal government's duty to
provide public and other federally assisted low-income housing that is
decent, safe, and free from illegal drugs.

How Benefit Eligibility Requirements May Affect whether a Drug Offender
Loses Benefits

TANF, food stamps, federally assisted housing, and Pell Grants are
meanstested benefits. To receive the benefits, individuals must meet
certain eligibility criteria. These criteria vary with the benefit. For
instance, states determine maximum earned income limits for TANF, but to
receive food stamps, the federal poverty guidelines are generally used in
determining eligibility. To receive federally assisted housing, local area
median income is used. Additionally, most adults eligible for TANF and
some adults eligible for food stamps must meet specified work requirements
to participate in the programs. Table 3 summarizes the general eligibility
requirements for the federal benefits discussed in this report and
identifies the federal, state, and local agencies responsible for
administering the programs.

6GAO, Drug Control: Difficulties in Denying Federal Benefits to Convicted
Drug Offenders, GAO/GGD-92-56 (Washington, D.C.: April 21, 1992).

Table 3: General Eligibility Criteria to Receive Selected Federal Benefits
and Federal Agencies Responsible for Administering Programs

Federal benefit General eligibility requirements to receive the federal
benefit Federal, state, or local agency responsible for administering the
benefit

TANF States determine which needy families receive benefits. The Administration
for Children and Families in the Department of Health and Human Services (HHS).

A needy family must consist of at least one child living with a relative
or may consist of a pregnant woman. Eligibility for benefits is based on a
variety of factors, including earned income limits, and work and education
requirements.

Food stamps 	In most states, household gross income cannot exceed The Food
and Nutrition Services in the U.S. 130 percent of the federal poverty
level, and net income Department of Agriculture (USDA). cannot exceed 100
percent of the poverty guidelines,

                                       a

and most states place limits on household assets. Unless exempted by law,
all food stamp recipients are subject to some type of work requirements.
Able-bodied adults without dependent children must meet more stringent
work requirements to receive benefits.

Postsecondary Requirements include a high school diploma or general U.S.
Department of Education.

education assistance 	equivalency degree (GED); acceptance into or active
enrollment in an eligible, degree-granting institution; and family income
requirements.b

Federally assisted Income eligibility is based on HUD's local area's
median Office of Public and Indian Housing within the

housing 	family income determinations. PHAs that manage HUD. Local PHAs
have responsibility for managing federal housing may establish nonincome
preference federally assisted housing rental units. factors.

        Licenses,      Professional competency The Department of Justice      
       contracts,      and specific licensure  (DOJ) maintains                
and loans under the requirements as defined  records of the information it 
                           by regulations.                receives from state 
    Denial of Federal                             and federal court officials 
                                                         about persons denied 
    Benefits Program                           benefits under this program,   
                                               and the General                
                                                      Services Administration 
                                                    includes this information 
                                               on its list of persons         
                                               excluded from federal          
                                               procurement or nonprocurement  
                                               programs, to                   
                                               which other federal agencies   
                                               have access.                   

Source: GAO analysis of federal and state requirements and agency
documents.

aFor 2003, 130 percent of the federal poverty level was roughly equivalent
to $1,654 per month for a family of three, and the net income threshold
for a family of three was about $1,272 per month. Net income is determined
by taking into account a number of approved deductions from gross income.

bStudents must be considered as either dependents or independent of their
parents for the purposes of financial aid, and the share of family income
and assets that are expected to be available for a student's education is
determined by certain formulas.

Not all persons who meet the general eligibility requirements to receive
federal benefits participate in the respective programs. Our recent study
on programs that aim to support needy families and individuals shows that

the portion of those eligible to receive the benefits that actually
enrolled in the programs varied among programs.7 Among families eligible
to participate in TANF in 2001, between 46 percent and 50 percent were
estimated to be participating in the program. For food stamps in 2001,
between 46 percent and 48 percent of eligible households were estimated to
participate in the program. For federally assisted housing, between 13
percent and 15 percent of eligible households in 1999 were estimated to be
covered by the Housing Choice Voucher (HCV) Program and between 7 percent
and 9 percent of eligible households in 1999 were estimated to be covered
by the Public Housing Program. Further, the Department of Education
reports that among all applicants for federal postsecondary education
assistance in academic year 2001-2002, about 77 percent of the applicants
that were eligible to receive Pell Grants applied for and received them.

Drug offenders would be directly affected by the federal provisions that
allow for denial of low-income federal benefits when, apart from their
disqualifying drug offense, they would have qualified to receive the
benefits. For example, if a drug offender is not in a financially needy
family and living with her dependent child, the drug offender would not be
eligible for TANF benefits aside from the drug offense conviction.8 To be
directly affected by the ban on food stamps, a drug offender would have
had to meet income tests and work requirements, unless the work
requirements are, under certain specified circumstances, identified as not
applicable by federal food stamp laws; otherwise, the offender's
ineligibility to receive the benefit would disqualify him, as opposed to
his drug offense. Because the ban on the receipt of TANF and food stamps
is for life, an offender who is not otherwise eligible to receive the
benefits at one point in time might become otherwise eligible to receive
the benefits at a later point in time and at that time be affected by the
provisions of PRWORA.

To be otherwise eligible to receive federal postsecondary education
assistance, a person convicted of a disqualifying drug offense would, at a
minimum, have to be enrolled in or accepted at an institution of higher
education, as well as meet certain income tests. To be otherwise eligible

7GAO, Means-Tested Programs: Information on Program Access Can Be an
Important Management Tool, GAO-05-221(Washington, D.C.: February 25,
2005).

8Needy pregnant women can also be eligible for TANF benefits, as they
constitute a family unit for the purposes of TANF.

Estimated Number of Drug Offenders Denied Federal Postsecondary Education,
Housing, and Certain Contracts and Grants Varies by Type of Benefit

for federally assisted housing benefits, a person would have to meet
income tests.

We estimated that among applicants for federal postsecondary education
assistance, drug offenders constituted less than 0.5 percent on average of
all applicants for assistance during recent years. In general, the
education attainment level of drug offenders is lower than that of the
general population, and this lower level affects drug offenders'
eligibility for federal postsecondary assistance. Among selected large
PHAs that reported denying applicants admission into public housing during
2003, less than 5 percent of applicants were denied admission because of
drugrelated criminal activities. PHAs have discretion in developing
policies to deny offenders for drug-related criminal activities. Federal
and state court sentencing judges were reported to impose sanctions to
federal benefits to fewer than 600 convicted drug offenders in 2002 and
2003, or less than 0.2 percent of felony drug convictions on average.

An Estimated 17,000 to 41,000 Drug Offenders Lost Selected Federal
Postsecondary Education Benefits in Selected Years

According to Department of Education data on applicants for federal
postsecondary education assistance for the academic years from 20012002
through 2003-2004, less than 0.5 percent on average of the roughly 11
million to 13 million applicants for assistance reported on their
applications that they had a drug offense conviction that made them
ineligible to receive education assistance in the year in which they
applied. These numbers do not take into account the persons who did not
apply for federal postsecondary education assistance because they thought
that their prior drug convictions would preclude them from receiving
assistance or any applicant who falsified information about drug
convictions. Using these data and Department of Education data on
applicants that received assistance for the academic years 2001-2002
through 2003-2004, we estimated that between 17,000 and 20,000 applicants
per year would have been denied Pell Grants, and between 29,000 and 41,000
would have been denied student loans if the applicants who self-certified
to a disqualifying drug offense were eligible to receive the benefits in
the same proportion as the other applicants. (See app. III for details on
our methods of estimating these figures.)

In general, the educational attainment levels of persons convicted of drug
offenses is less than that of persons in the general population. This
results in proportionately fewer persons eligible for these education
benefits than in the general population. Our analysis of data from the
only national survey of adults on probation that also reports on their
educational

attainment indicates that among drug offenders on probation during 1995,9
less than half had completed high school or obtained a general equivalency
degree (GED)-prerequisites for enrolling in a postsecondary institution.
By comparison, according to a Bureau of Justice Statistics report, about
18 percent of adults in the general population had less than a high school
degree. More recent data from the U.S. Sentencing Commission on roughly
26,000 drug offenders sentenced federally during 2003 indicate that half
of them had less than a high school degree, about one-third had graduated
from high school, and about 18 percent had at least some college.10 In
addition, our analysis of BJS data on drug offenders released from prisons
in 23 states during 2001 indicate that about 57 percent of these drug
offenders had not completed high school by the time they were admitted
into prison; about 36 percent had completed high school or obtained a GED
as their highest level of education completed; and the remainder had
completed some postsecondary education.

In Selected PHAs Less Than 6 Percent of Residents and Applicants Were
Denied Federally Assisted Housing during 2003 because of Drug-Related
Criminal Activities

We obtained data from 17 of the largest PHAs in the nation on the
decisions that they made to deny federally assisted housing benefits to
residents or applicants during 2003.11 Thirteen of the 17 PHAs reported
data on both (1) the number of leases in the Public Housing Program units
that they manage that ended during 2003 and (2) the number of leases that
were terminated for reasons of drug-related criminal activities. These 13
PHAs reported terminating leases of 520 tenants in the Public Housing
Program because of drug-related criminal activities. The termination of a
lease is the first step in evicting tenants from public housing. Tenants
whose leases were terminated for reasons of drug-related criminal
activities constituted less than 6 percent of the 9,249 leases that were

9The Bureau of Justice Statistics' Survey of Adults on Probation in 1995
is its most recent nationally representative survey of probationers that
collects data on education level. These data also report the type of
offense that led to the probation term.

10United States Sentencing Commission, 2003 Sourcebook of Federal
Sentencing Statistics, Washington, D.C.: United States Sentencing
Commission. Published annually.

11We contacted 40 of the largest PHAs in the nation and requested
information about denials of federally assisted housing assistance. The 17
that provided us with information are shown in appendix IV. These 17 are
not statistically representative of the more than 3,000 PHAs nationwide.
While our review covered both the Public Housing Program and the Housing
Choice Voucher Program, not all 17 of the PHAs provided us with
information on both types of programs.

terminated in these 13 PHAs during 2003.12 Among these PHAs, the
percentage of terminations of leases for reasons of drug-related criminal
activities ranged from 0 percent to less than 40 percent. These PHAs also
reported that the total number of lease terminations in 2003 and the
volume of denials for drug-related activities were generally comparable
with similar numbers for the 3 prior years. (See app. IV for data for each
PHA that responded to our request for information.)

Fifteen PHAs acted on 29,459 applications for admission into the Public
Housing Program during 2003. Among these applicants seeking residency, we
estimated that less than 5 percent were denied admission because of their
drug-related criminal activities. The PHAs also reported that they acted
on similar numbers of applicants and made similar numbers of denial
decisions in the prior 3 years. Table 4 shows the data on lease
terminations and denials of admission in two federally assisted housing
programs.

Table 4: Number and Percentage of Tenants and Applicants for Federally
Assisted Housing Benefits That Were Denied Benefits for Reasons of
Drug-Related Criminal Activities in Selected Large PHAs in 2003

                                                                 Among PHAs,a 
                                                                     range of 
                                                     Estimated  percentage of 
                                                       overall 
                                               percentage of   actions due to 
                                               actions         
              Type of    Number of                      due to                
              action        PHAs     Number of    drug-related   drug-related
              taken                                            
    Housing              reporting     actions        criminal       criminal 
    program    by PHAs      data                    activities     activities 

    Public Housing    Lease terminations   13   9,249    5.6%b    0% to 39.3% 
       Program                                                 
                     Admissions decisions  15  29,459     < 5%   0.4% to 6.9% 
     HCV Program       Terminations of      9  12,703     < 2%  0.07% to 6.1% 
                          assistance                           
                     Admissions decisions   9  21,996   < 1.5%          0.7%c 

Source: GAO analysis of data from selected public housing agencies.

aAmong PHAs that reported the data.
bThis figure represents the actual percentage of actions taken due to
drug-related criminal activities.
cOnly one PHA reported denying HCV assistance for reasons of drug-related
criminal activities.

12Leases may end or be terminated for a variety of reasons, including a
tenant's failure to make payments due under the lease.

We also obtained and analyzed data from HUD on the number of evictions and
denials of admission into public housing during fiscal years 2002 and 2003
that occurred for reasons of criminal activity, of which drug-related
criminal activity is a subset. More than 3,000 PHAs reported to HUD that
in each of these years there were about 9,000 evictions for reasons
related to criminal activity and about 49,000 denials of admission for
reasons of criminal activity. As a percentage of units managed by these
PHAs, evictions for reasons of criminal activity in each of these years
amounted to less than 1 percent of units managed, and denials of admission
amounted to about 4 percent of units managed. Evictions and denials for
reasons of drug-related criminal activities would have to be equal to or,
much more likely, less than these percentages.

On the basis of data that 9 PHAs were able to report about terminating
participation in the HCV Program during 2003, we estimated that less than
2 percent of the decisions to terminate assistance in the HCV program (of
the roughly 12,700 such decisions) were for reasons of drug-related
criminal activities. In addition, 9 PHAs reported that they acted on
21,996 applications for admission into the HCV Program and that less than
1.5 percent of applicants were denied admission for reasons of
drug-related criminal activities.

Discretionary Authority May Allow Some to Receive Housing Benefits, but
Supply Limitations May Be a Significant Barrier

Local PHAs that administer federally assisted housing benefits have
discretion in determining whether current tenants or applicants for
assistance have engaged in drug-related criminal activities that
disqualify them from receiving housing benefits. HUD requires PHAs to
develop guidelines for evicting from or denying admittance into federally
assisted housing to individuals who engage in drug-related criminal
activity. A November 2000 HUD study on the administration of the HCV
Program described the variation in PHA policies on denying housing to
persons who engaged in drug-related criminal activities. HUD concluded
that because of the policy differences, some PHAs could deny applicants
who could be admitted by others.13 For example, some PHAs consider only
convictions in determining whether applicants qualify for housing
benefits, while others look at both arrests and convictions. Some look for
a pattern of drug-related criminal behavior, while others look for
evidence that any

13Deborah Devine, Barbara A. Haley, Lester Rubin, and Robert W. Gray, "The
Uses of Discretionary Authority in the Section 8 Tenant-Based Program: A
Baseline Inventory of Issues, Policy, and Practice," Washington, D.C.:
U.S. Department of Housing and Urban Development, November 2000.

drug-related criminal activities occurred. In addition, among PHAs, the
period of ineligibility for assistance arising from a prior eviction from
federally assisted housing because of drug-related criminal activities
ranged from 3 to 5 years. (See app. IV for a summary of selected PHA
policies.)

Any imbalance between the supply of and demand for federally assisted
housing may also affect whether drug offenders are denied access to this
benefit. The stock of available federally assisted housing units in the
Public Housing Program is generally insufficient to meet demand. PHAs may
have long waiting lists, up to 10 years in some cases, for access to
federally assisted housing. As PHAs generally place new applicants at the
end of waiting lists, a drug offender who might be disqualified from
federally assisted housing but who applies for housing assistance could go
to the end of a PHA's waiting list. Until that applicant moved to the top
of the waiting list, the limited supply of federally assisted housing, and
not necessarily a drug offense conviction, would effectively deny the
applicant access to the benefit.

Limited Use of the Denial of Federal Benefits Program Sanctions

Between 1990 and the second quarter of 2004, BJA received reports from
state and federal courts that 8,298 offenders were sanctioned under the
Denial of Federal Benefits Program in federal and state courts. This
amounted to an average of fewer than 600 offenders per year. The Denial of
Federal Benefits Program provides judges with a sentencing option to deny
federal benefits such as grants, contracts, and licenses. About 62 percent
of the cases reported to be sanctioned under the Denial of Federal
Benefits Program occurred in federal courts, and the remaining 38 percent
occurred in state courts. For recent years (2002 and 2003), BJA reported
that fewer than 600 persons were denied federal benefits under the
program. In 2002, there were more than 360,000 drug felony convictions
nationwide. On average, less than 0.2 percent of these convicted drug
felons were sanctioned under this program.

According to the BJA data, state court judges in 7 states have imposed the
sanction, and state court judges in Texas accounted for 39 percent of all
cases in which drug offenders were reportedly denied benefits under this
program by state court judges. Federal judges in judicial districts in 26
states had reportedly imposed denial of benefits sanctions, and federal
judges in Texas accounted for 21 percent of the cases in which federal
judges reportedly denied benefits. The pattern of use of sanctions under
this program, with substantially more use in some jurisdictions than in
others, may indicate that there are drug offenders in some locations who

State Exemptions to the Federal Bans on TANF and Food Stamps and Other
Factors Affect the Percentage of Drug Offenders That Would Have Been
Eligible to Receive the Benefits

could have received the sanction but did not. (See app. V for more
information about this program.)

We previously reported on the relatively limited use of this sanction.14
We reported then that many offenders who could be denied access to federal
benefits would also be sentenced to prison terms that exceed the benefit
ineligibility period; therefore, upon release from prison, the offenders
would not necessarily have benefits to lose. BJA officials reported that
as of 2004, about 2,000 convicted drug offenders were still under sanction
under the Denial of Benefits Program, as the period of denial had expired
for the other sanctioned offenders.

Most states have acted on the discretionary authority provided them under
federal law to enact legislation that exempts some or all convicted drug
felons in their states from the federal bans on their receipt of TANF and
food stamps. That is, these state laws allow that convicted drug felons
may not be banned for life from receiving TANF and food stamps provided
they meet certain conditions. For states that had not modified the federal
ban on TANF, we estimated that about 15 percent of all offenders and 27
percent of female offenders released from prison during 2001 would have
met selected eligibility requirements and would therefore potentially be
affected by the ban. We also estimated that among drug offenders released
from prison during 2001 in states that had not modified the federal ban on
food stamps about a quarter were custodial parents whose reported income
was below federal poverty thresholds for food stamps. While food stamps
are not limited to custodial parents, and the ban could affect other drug
offenders, we limited our analysis to this group.

Eighteen States Fully Implement the Federal Ban on TANF, as the Remainder
Have Enacted Legislation to Exempt Some or All Convicted Drug Felons

A total of 32 states have enacted laws that exempt all or some convicted
drug felons from the federal ban on TANF benefits. Of these states, 9 have
enacted laws that exempt all convicted drug felons from the federal ban,
and these persons may receive TANF benefits provided that they meet their
state's general eligibility criteria. Another 23 states have passed laws
that exempt some drug felons from the TANF ban. The modifications allow
that some convicted drug felons may receive benefits and generally fall
into any of three categories: (1) Some states permit felons convicted of

14See GAO, Drug Control: Difficulties in Denying Federal Benefits to
Convicted Drug Offenders, GAO/GGD-92-56, (Washington, D.C.: April 21,
1992).

drug use or simple possession offenses to continue to receive TANF
benefits but deny them to felons convicted of drug sales, distribution, or
trafficking offenses; (2) some states allow convicted felons to receive
TANF benefits only after a period of time has passed; and (3) some states
allow convicted drug felons to receive TANF benefits conditioned upon
their compliance with drug treatment, drug testing, or other conditions.
(See app. II for the status of states' exemptions to the TANF ban.)

Using state-level data on drug arrests as a proxy for state-level data on
drug convictions, we estimated that the 9 states that completely opted out
of the TANF ban and exempted all convicted drug felons from the ban
accounted for about 10 percent of drug arrests nationally in 2002.15 The
23 states whose exemptions modified the TANF ban accounted for about 45
percent of drug arrests nationally. For these states with various
exemptions, it is difficult to determine to which drug felons the ban
might apply, as participation in the program is contingent upon a felon's
behavior (such as abiding with conditions of probation or parole
supervision, or participating in drug treatment). Finally, the 18 states
that fully implemented the TANF ban accounted for about 45 percent of all
drug arrests nationwide.

In States That Fully Implement the Ban on TANF, about 15 Percent of All
Drug Offenders Released from Prison in 2001 Were Estimated to Meet TANF
Eligibility Criteria

Using Bureau of Justice Statistics survey data on the family and economic
characteristics of drug offenders in prison and state-level data on the
number of drug offenders released from prison during 2001 in 14 of the 18
states that fully implement the ban on TANF, we estimated that about 15
percent of those released from prison were parents of minor children,
lived with their children, and had earned income below the maximum levels
permitted by their states of residence.16 That is, but for the ban, they
may have been eligible to receive TANF benefits. We estimated that the
majority of drug felons-who are single males and not custodial parents-

15State-level data on the number of felony drug convictions are not
collected by federal agencies or by organizations that monitor state court
activities.

16In the 14 states, about 96,000 drug offenders were released from prison
during 2001, and of these, about 51,000 had been released from a new
sentence for a drug conviction. The 4 omitted states, for which we could
not obtain data, accounted for about 3 percent of the population of the 18
states that fully implemented the TANF ban. Our estimates of the
percentage that may have been eligible to receive TANF are based in part
on data from the Bureau of Justice Statistics Survey of Inmates of State
Correctional Facilities. The survey provides data that prisoners report
about their family and economic characteristics prior to their admission
into prison. See appendix II for details on our methodology.

did not meet these TANF eligibility requirements and would therefore not
have been qualified to receive the benefit even in the absence of the
provisions of PRWORA. (See app. II for additional information about the
methods used to estimate these quantities.)

Female drug offenders released from prison in the 14 states constituted
about 13 percent of drug offenders released from prison in 2001. We
estimated that between 25 percent and 28 percent of these female offenders
were parents of minor children who lived with their children and whose
incomes were below state thresholds, and therefore stood to lose TANF
benefits. This percentage among female drug offenders released from prison
is about twice that for males. From the available data, we estimated that
less than 15 percent of male prisoners were parents who lived with their
children and had earned incomes that would qualify them to receive TANF
benefits.

Other factors, which we could not take into account to estimate the
percentages of drug offenders that could be eligible to receive TANF
benefits, include citizenship status and total family income. Noncitizens
with fewer than 5 years of residence in the United States are generally
ineligible to receive TANF. Several of the states for which we obtained
data on drug offenders released from prison have relatively large
noncitizen populations. Therefore, among those drug offenders that we
estimated could have been eligible to receive TANF benefits might be some
ineligible noncitizens. In addition, the data that we used to estimate
whether drug offenders met state income eligibility requirements included
individual income rather than total family income. It is possible that
some prisoners would join family units with incomes above state TANF
eligibility earned income limits and would thus be disqualified for
benefits.

Among the drug offenders released from prison during 2001, the percentage
that may be affected by the TANF ban at any time during their lifetimes
would be greater than our estimate of those initially affected. This is
because at a later date some of these offenders may meet the general
eligibility criteria for receiving benefits. Thus, the percentage ever
affected by the bans would grow over time.

Because of data availability, our estimates focus on convicted drug felons
who were in prison. We do not have data to assess the effect of the TANF
ban on drug felons who received probation or who were sentenced to time in
local jails. According to BJS data, nationwide, about one-third of
convicted drug felons are sentenced to probation. Moreover, our estimates
apply to the states that fully implemented the ban on TANF. Because of

complexities associated with state exemptions to the federal ban and the
lack of sufficiently detailed data, we cannot provide an estimate of the
percentage of convicted drug offenders who could be affected by the ban in
the 23 states that modified the TANF ban. We note, however, that state
modifications to the ban may allow convicted drug felons to participate in
TANF if they abide by the conditions set in the state exemptions (such as
abide by conditions of parole or probation supervision or participate in
drug treatment). In these states, unlike in the states that fully
implement the federal ban, the post-conviction behavior of offenders would
help to determine whether they could receive the benefit. Other state
modifications allow drug offenders to receive TANF benefits at some point
in the future (such as after completing drug treatment or receiving a
sufficient number of negative drug test results). In states that require
that drug felons wait before becoming eligible to participate in TANF, the
federal ban is in effect until the waiting period ends. We would therefore
expect estimates of the percentage affected during the waiting period to
be similar to the estimates of the percentage affected in the states that
fully implemented the federal ban.

In States That Fully Implement the Ban on Food Stamps, About One-Fourth of
Drug Offenders Released from Prison in 2001 Were Parents Who Could Have
Been Eligible for Food Stamps

At the time of our review, 15 states fully implemented the federal ban on
food stamp benefits to convicted drug felons, and 35 states had passed
laws to exempt all or some convicted drug felons in their own states from
the federal ban on food stamps. Of the 35 states with exemptions, 14
states exempt all convicted drug felons from the food stamp ban, and 21
have laws that exempt some convicted drug felons from the food stamp ban
provided that they meet certain conditions.17 In the 21 states that
modified the food stamp ban, the modifications are similar to those for
TANF and generally include (1) exempting persons convicted of drug
possession from the ban, while retaining it for persons convicted of drug
sales, distribution, or trafficking; (2) requiring a waiting period to
pass before eligibility is restored; and (3) conditioning food stamp
eligibility upon compliance with drug treatment, drug testing, or other
conditions. (See app. II for the status of states' exemptions to the food
stamps ban.)

States' decisions to exempt all or some convicted drug felons in their
states from the ban on food stamps affect the proportion of drug felons

17The number of states exempting some or all convicted drug felons from
the food stamp ban has increased over time. USDA reports, for example,
that in 2001, 19 states had exemptions for some or all convicted drug
felons. By 2004, 35 states had some type of exemption.

that can be affected by the ban. Using the state-level drug arrest data as
the proxy for felony drug convictions (as we did for TANF), we find that
the 15 states that fully implemented the ban on food stamps accounted for
about 22 percent of all drug arrests nationally. Using data from the BJS
inmate survey on the family and economic characteristics of drug offenders
in prison and state-level data on the number of drug offenders released
from prison during 2001 in 12 of the 15 states that fully implemented the
ban on food stamps, we estimated that about 23 percent of those released
from prison were parents of minor children whose incomes were below the
federal poverty guidelines. Among male drug offenders, we estimated that
about 22 percent met these conditions, while among female drug offenders,
we estimated that about 36 percent did.18 We are unable to provide an
estimate of the percentage of drug offenders that could be eligible to
receive food stamps as able-bodied adults without dependent children.
According to USDA, in 2003, this class of food stamp recipients
constituted about 2.5 percent of food stamp recipients nationwide. Food
stamps are not limited to custodial parents. However, we limited our
assessment to custodial parents because of data limitations.

Because the denial of food stamps is a lifetime ban, the number of drug
offenders affected by the ban will increase over time, as additional
convicted drug felons are released from prison.

Also, as with the TANF estimates, data limitations precluded our providing
estimates for the felony drug offenders that were sentenced to probation
in 2001 or for the states that modified the federal ban.

A complex array of provisions of federal law allow or require federal
benefits be denied to different classes of drug offenders. There is also a
good deal of discretion allowed in implementing these laws that can exempt
certain drug offenders from their application. Our estimates

                                   Concluding

                                  Observations

18Within these 12 states, there were about 67,000 drug offenders released
from prison during 2001, and of these, about 30,000 were released from a
new sentence for a drug conviction. Data for the other 3 states were not
available. We have no reason to believe that these 3 states are
substantially different from the other 12 in relation to the impact s of
the ban. Our estimates of the percentage that may have been eligible to
receive food stamps are based in part on data from the Bureau of Justice
Statistics Survey of Inmates of State Correctional Facilities. The survey
provides data that prisoners report about their family and economic
characteristics prior to their admission into prison. See appendix II for
details on our methodology.

indicate that denial of benefit laws potentially affect relatively small
percentages of drug offenders, although the numbers potentially affected
in given years may be large. There are a number of reasons why the
percentages affected may be relatively small. First, large numbers of drug
offenders would not be eligible for these benefits regardless of their
drug offender status. For example, those who lack a high school diploma
are ineligible for postsecondary educational loans or grants, and many do
not meet eligibility requirement for TANF and food stamps. Also, in the
case of TANF and food stamps, the majority of states have used their
discretion to either partially or fully lift the ban on these benefits for
certain drug offenders.

It is important to note that although the overall numbers of drug
offenders that could be affected by the TANF and food stamp bans are
relatively small in comparison with the numbers of drug offenders, our
estimates suggest that the effects of the bans disproportionately fall on
female offenders. This is because they are more likely to be custodial
parents with low incomes and thus otherwise eligible for the benefits.

We provided a draft of this report to the Attorney General; the
Secretaries of the Departments of Education, Agriculture, and Housing and
Urban Development; the Assistant Secretary of the Administration for
Children and Families; the Director of the Office of National Drug Control
Policy; the Research Director of the United States Sentencing Commission;
and the Director of the Administrative Office of the United States Courts
for their review and comment. We received technical comments from the
Departments of Justice, Agriculture, and Education, and from the United
States Sentencing Commission and Administrative Office of the United
States Courts, which we incorporated into the report where appropriate.

Agency Comments
and Our Evaluation

We are sending copies of this report to the Attorney General; the
Secretaries of the Departments of Education, Health and Human Services,
Agriculture, and Housing and Urban Development; the Director of the Office
of National Drug Control Policy; the Research Director of the United
States Sentencing Commission, and the Director of the Administrative
Office of the United States Courts. We will also make copies available to
others upon request. In addition, the report will be available at no
charge on GAO's Web site at http://www.gao.gov.

If you or your staff have any questions concerning this report, please
contact me at (202) 512-8777 or by e-mail at [email protected]. Contact

points for our Offices of Congressional Relations and Public Affairs may
be found on the last page of this report. GAO staff who made major
contributors to this report are listed in appendix VI.

Laurie E. Ekstrand, Director Homeland Security and Justice Issues

Appendix I: Objectives, Scope, and Methodology

Federal law provides that certain drug offenders may or must be denied
selected federal benefits, such as Temporary Assistance for Needy Families
(TANF), food stamps, federally assisted housing, postsecondary education
grants and loans, and certain federal contracts and licenses. Our
objectives were to analyze and report on two interrelated questions about
the number or percentage of drug offenders that could be affected by the
provisions: (1) In specific years, how many drug offenders were estimated
to be denied federal postsecondary education benefits, federally assisted
housing, and selected grants, contracts, and licenses? (2) What factors
affect whether drug offenders would have been eligible to receive TANF and
food stamp benefits, but for their drug offense convictions, and for a
recent year, what percentage would have been eligible to receive these
benefits?

In addition, we were asked to address the impact of federal benefit denial
laws on minorities and the long-term consequences of denying federal
benefits on the drug offender population and their families. Because of
severe data limitations, we were unable to provide a detailed response to
this matter. The final section of appendixes II, III, IV, and V in this
report include discussions of the data limitations that precluded us from
estimating the impacts on minorities. Where information was available, we
also identify in the appendices some of the possible long-run consequences
of denial of benefits.

We limited our analysis of federal laws to those that explicitly included
provisions that allowed for or required the denial of federal benefits to
drug offenders. We excluded other provisions that provide for denial of
benefits to all offenders, of which drug offenders are a subset. We also
excluded from our analysis provisions that applied to offenders only while
they are incarcerated and provisions that applied to fugitive felons.
Other federal laws relating to drug offenders but not within the scope of
our review include provisions such as those making a person ineligible for
certain types of employment, denying the use of certain tax credits, and
restricting the ability to conduct certain firearms transactions. Further,
because of the limited data available on persons actually denied federal
benefits, we provide rough estimates of either the number or the
percentage of drug offenders affected by the relevant provisions. We
provide an overview of these methodologies below but we discuss the
specifics of our methodologies for analyzing and estimating the impacts of
denying specific federal benefits in appendices II through V.

We assessed the reliability of the data that we used in preparing this
report by, as appropriate, interviewing agency officials about their data,

Appendix I: Objectives, Scope, and Methodology

reviewing documentation about the data sets, and conducting electronic
tests. We used only the portions of the data that we found to be
sufficiently reliable for our purposes in this report.

We conducted our work primarily in Washington, D.C., at the headquarters
of five federal agencies-the Departments of Justice (DOJ), Agriculture
(USDA), Housing and Urban Development (HUD), Education (ED), and Health
and Human Services (HHS)-responsible for administering the denial of
federal benefit laws. We also conducted work at the Office for National
Drug Control Policy (ONDCP)-which has responsibilities for national drug
control policy-the Administrative Office for United States Courts
(AOUSC)-which provides guidance to the courts for the implementation of
statutory requirements-and the United States Sentencing Commission
(USSC)-which has responsibilities for monitoring federal sentencing
outcomes.

Assessing How Many Drug Offenders Were Estimated to Be Denied Federal
Postsecondary Education and Federally Assisted Housing Benefits and
Certain Grants and Contracts

To estimate how many or what percentage of drug offenders were reported to
be denied federal postsecondary education and federally assisted housing
benefits and certain grants and contracts under the Denial of Federal
Benefits Program, we obtained and analyzed data from agency officials.
From ED, we obtained data for several years on the number of applicants
using the Free Application for Federal Student Aid (FAFSA), the number of
these who reported a disqualifying drug offense conviction, the number
eligible for Pell Grants, and the number receiving Pell Grants and student
loans. We analyzed these data to generate our estimates of the number of
those that reported disqualifying drug offenses that would have been
eligible to receive Pell Grants and student loans. We also obtained Bureau
of Justice Statistics (BJS) data that reported on the educational
attainment of nationally representative sample of offenders on probation.
We used these data, along with USSC data on sentenced drug offenders and
BJS data on drug offenders released from prison, to assess the education
levels of drug offenders. To identify factors that could contribute to the
number of drug offenders denied federal postsecondary education benefits,
we interviewed officials at ED about federal regulations, guidance, and
rulings pertaining to the eligibility to receive benefits. Appendix III
describes in more detail our methods for estimating the education of those
denied education benefits.

From a nonprobability sample of some of the largest public housing
agencies (PHA) in the United States, we obtained information about
reported actions taken in 2003 in these PHAs to deny persons federally
assisted housing benefits for reasons of drug-related criminal activities.

Appendix I: Objectives, Scope, and Methodology

We selected large agencies because of the volume of actions that they take
in a given year and to provide indications of the range of outcomes in
PHAs in different settings with different populations. We also obtained
and analyzed data from HUD on persons reportedly evicted from or denied
admission into public housing for reasons of criminal activities. From
selected PHAs, we obtained, analyzed, and compared termination and
admissions policies and procedures used during 2003 or 2004 to deny
federally assisted housing to persons involved in drug-related criminal
activities. We also spoke with staff from selected research organizations,
national associations, and PHAs to review the eligibility criteria to
receive federal benefits. Appendix IV describes our methods for assessing
denials of federally assisted housing.

From the Bureau of Justice Assistance (BJA), we obtained data on drug
offenders reported to have been denied federal benefits under the Denial
of Federal Benefits Program. We spoke with officials at BJA about the
current operations and plans to enhance the program, and we interviewed
officials from USSC and AOUSC about the operations of this program. We
also interviewed ONDCP officials about the array of federal provisions
that provide for denial of federal benefits and federal programs that
provide for drug treatment for drug offenders. Appendix V describes our
methodology for analyzing the Denial of Federal Benefits Program.

Data limitations concerning the actual number of persons denied TANF and
food stamp benefits required us to develop estimates of the drug offenders
that could be denied these benefits in that they had characteristics that
would have allowed them to qualify to receive the benefits except for
their drug offense convictions. To determine the extent to which drug
offenders were otherwise qualified or eligible to receive federal
benefits, we identified key elements of the eligibility to receive federal
benefits. We met with officials at the federal agencies responsible for
administering TANF-the Department of Health and Human Services- and food
stamps-the U.S. Department of Agriculture-to discuss issues related to
eligibility to receive these benefits. We obtained and analyzed data from
BJS on the characteristics of drug offenders in prison, and we applied
this information to the number of drug offenders released from prison
during 2001 in states that fully implemented the ban on TANF. To determine
the current status of states that have opted out of or modified federal
provisions banning TANF and food stamp benefits to persons convicted of
drug felony offenses, we reviewed state laws and contacted officials at
USDA (which annually surveys states about the status of their laws in
relation to the ban on food stamps) and state officials in states that

Estimating the Percentage of Convicted Drug Felons That Would Have Been
Eligible to Receive TANF or Food Stamps and the Factors Contributing to
These Estimates

Appendix I: Objectives, Scope, and Methodology

Data Sources Used in
Preparing This Report

have modified the federal ban on TANF or food stamps to discuss the status
of their provisions regarding the exemptions under their state laws.
Appendix II provides detailed information on our methodology for assessing
the TANF and food stamps bans.

From the following sources, we obtained, assessed the reliability of, and
analyzed data related to denial of federal benefits that we used in
developing estimates of the impacts of the federal provisions. To assess
the reliability of the data, as needed, we interviewed agency officials
about the data systems, reviewed relevant documentation, and conducted
electronic tests of the data. We determined that the data were
sufficiently reliable for the purposes of this report. The data sources
included

o  	Bureau of Justice Assistance: Data on the number of drug offenders
reported to BJA by state and federal courts as having been denied federal
benefits under the Denial of Federal Benefits Program from 1991 to 2004.

o  Bureau of Justice Statistics:

o  	Survey of Inmates of State Correctional Facilities in 1997: We used
these data to estimate the number of convicted drug felons in prison that
were parents of minor children, lived with their children prior to their
incarceration, and had incomes within state earned income limits. We used
these estimates to assess the impacts of the provisions allowing for the
denial of TANF and food stamp benefits.

o  	National Corrections Reporting Program, 2001: We used these data to
obtain counts of the number of drug offenders released from prison during
2001 in selected states. We also used these data to provide estimates of
the level of education completed from drug offenders released from prison
during 2001 and in developing our estimates of the impacts of the TANF and
food stamps provisions.

o  	Survey of Adults on Probation, 1995: We used these data, from the only
national source of data on the characteristics of adults on probation of
which BJS is aware, to learn about the education levels of drug offenders
on probation and in developing estimates of the impact of denying federal
postsecondary education assistance.

o  	Selected state corrections and court officials: For selected states
that fully implemented the ban on TANF and food stamps, we obtained data
on the numbers of convicted drug felons released from prison in during
2001. We

Appendix I: Objectives, Scope, and Methodology

used these data in developing estimates of the impacts of TANF and food
stamps.

o  	Department of Housing and Urban Development: We obtained and analyzed
data from HUD's Public Housing Assessment System (PHAS) and Management
Operations Certification Assessment System (MASS) for fiscal years 2002
and 2003 on the number of public housing residents evicted because of
criminal activities (of which drug-related criminal activities form a
subset), and of the numbers denied admission into the Public Housing
Program for reasons of criminal activities.

o  	Seventeen of the 40 largest PHAs in the nation: We requested
information from the 40 largest PHAs about the number of decisions they
made during 2003 to deny federally assisted housing to tenants and
applicants for reasons of drug-related criminal activities, and we
obtained data from 17 of these PHAs. Not all 17 PHAs provided responses to
all of our questions; therefore, we reported data only on the PHAs that
were able to provide data relevant to the question under review. We
selected these PHAs from among the 1,531 PHAs that managed both Public
Housing and Housing Choice Voucher (HCV) programs as of August 31, 2004.
We asked them for information about denials of federally assisted housing
for reasons of drug-related criminal activities, and we also asked them to
provide these data based on the race of tenants and applicants. HUD does
not collect this information. We used these data in describing the number
of persons denied federally assisted housing and in providing information
about the race of persons denied federal housing benefits.

o  	Department of Education: We obtained and analyzed data on the number
of students applying for federal postsecondary assistance for academic
years 2001-2002, 2002-2003, and 2003-2004. In addition, we obtained data
on the percentage of these applicants who were eligible to receive Pell
Grants and of these, the percentage that received them, and we also
obtained data on the percentage of applicants who received student loans.
We used these data in developing estimates of the impact of the denial of
federal postsecondary education assistance.

In addition, we used published statistical reports from various agencies
such as BJS; Uniform Crime Reports data on drug abuse violation arrests by
state; Department of Health and Human Services reports on the
characteristics of TANF recipients; USDA reports on food stamp recipients;
and the United States Sentencing Commission's 2003 Sourcebook of Federal
Sentencing Statistics.

Appendix I: Objectives, Scope, and Methodology

Impacts of the Federal Provisions on Racial Minorities and Long-Run
Impacts of Denying Federal Benefits

We were asked to address the impacts of the federal benefit denial laws on
racial minorities and the long-term impacts of denying federal benefits on
individuals that were denied, their families, and their communities.
Although very limited, the available information on these issues is
summarized in appendices II through V.

To determine the extent of data on the race of persons affected by the
denial of federal benefit provisions, we asked the officials that we
interviewed about their knowledge of data on the race of persons denied
federal benefits. We also spoke with researchers and officials at various
organizations about their knowledge of available data. To address data
limitations of HUD data on persons denied federally assisted housing
because of drug-related criminal activities, we requested, obtained, and
analyzed data provided by 17 of the largest PHAs in the nation on the race
of persons denied housing for reasons of drug-related criminal activity.

To determine the current research and data on the potential economic and
social impacts of the loss of federal benefits on individuals, families,
and communities, we conducted literature searches to identify and review
existing studies that have measured the impacts of the denial of federal
benefits on drug offenders and families. We interviewed experts to
understand how the incentives for drug treatment, as provided in the laws
that deny benefits, are likely to affect drug addicts' behavior, and we
obtained their views regarding the effects that incarceration and drug
convictions might have on a drug felon's potential employment and
earnings.

We conducted our work from March 2004 to July 2005 in accordance with
generally accepted government auditing standards.

Appendix II: Denial of TANF and Food Stamps Benefits: Legal Framework and
Methodologies

Legal and Administrative Framework

This appendix describes the legal and administrative framework for denying
TANF and food stamp benefits to convicted drug felons and our methods for
estimating the percentage of convicted drug offenders that would have been
eligible to receive TANF and food stamps but for their drug felony
convictions.

The Personal Responsibility and Work Opportunity Reconciliation Act
(PRWORA)1 of 1996 provides that persons convicted of certain drug felony
offenses are banned for life from receiving TANF and food stamp benefits.
Specifically, Section 115 of PRWORA, as amended, provides that an
individual convicted (under federal or state law) of any offense that is
classified as a felony by the law of the jurisdiction involved and that
has as an element the possession, use, or distribution of a controlled
substance shall not be eligible to receive TANF assistance or food stamp
benefits. The prohibition applies if the conviction is for conduct
occurring after August 22, 1996.2

TANF assistance includes benefits designed to meet a need family's
ongoing, basic needs (for example, for food, clothing, shelter, utilities,
household goods, and general incidental expenses) and includes cash
payments, vouchers, and other forms of benefits.3 TANF assistance excludes
short-term episodic benefits that are not intended to meet recurrent or
ongoing needs and that do not extend beyond 4 months. The federal
prohibition on TANF assistance to convicted drug felons does not apply to
TANF "nonassistance" benefits, which include benefits meant to assist an
individual's nonrecurring emergency needs.4 TANF nonassistance can include
drug treatment, job training, emergency Medicaid medical services,
emergency disaster relief, prenatal care, and certain public health
assistance.

1Pub. L. No. 104-193, 110 Stat. 2105 (1996) (codified at 21 U.S.C. S:
862a).

2See 21 U.S.C. S: 862a(d)(2).

3See 45 C.F.R. S: 260.31. The federal prohibition does not extend to
assistance from a state's own separate assistance funds such as State
Maintenance of Effort (MOE) funds. MOE funds are funds that states are
required maintain at certain historical state expenditure levels in order
to receive their full TANF block grants.

4As defined at 45 C.F.R. S: 260.31(b).

Appendix II: Denial of TANF and Food Stamps Benefits: Legal Framework and
Methodologies

The Food Stamp Program provides benefits in the form of electronic benefit
cards, which can be used like cash for food products at most grocery
stores. Eligible households receive a monthly allotment of food stamps
based on the Thrifty Food Plan, a low-cost model diet plan based upon
National Academy of Sciences' Recommended Dietary Allowances. For persons
between the ages of 18 and 50 who are also viewed as fit to work and who
are not the guardians of dependent children, PRWORA provides for a work
requirement or a time limit for receiving food stamp benefits. The
provision is known as the Able-Bodied Adults without Dependent (ABAWD)
provision. ABAWD participants in the food stamp program are limited to 3
months of benefits in a 3-year period unless they meet certain criteria.5

PRWORA provides that states may enact a legislative exemption removing or
limiting the classes of convicted drug felons that are otherwise affected
by the federal ban on TANF and food stamps. State laws providing for
exemptions need to have been enacted after August 22, 1996.

The Office of the Administration for Children and Families (ACF) within
the U.S. Department of Health and Human Services provides federal
oversight of the TANF program. TANF is funded by both federal block grants
and state funds, but states are responsible for determining benefit levels
and categories of families that are eligible to receive benefits. State
eligibility requirements establish earned income limits, and other rules,
and these requirements may vary widely among the states.

The U.S. Department of Agriculture's Food and Nutrition Service (FNS)
provides oversight for the Food Stamp Program, which is the primary
federal food assistance program that provides support to needy households
and to those making the transition from welfare to work. Eligibility for
participation is based on the Office of Management and Budget federal
poverty guidelines for households. Most households must meet gross and net
income tests unless all members are receiving TANF or selected other forms
of assistance.6 Gross income cannot exceed 130

5To be exempt from the ABAWD provisions, an otherwise able-bodied person
must either (1) live in local areas that provide for a waiver to the ABAWD
provision; (2) live in a state that utilizes provisions from the Balanced
Budget Act of 1997 (Pub. L. No. 105-33, 111 Stat. 251, 252 (1997)
(codified at 7 U.S.C. S: 2015(o)(6)) that allows for the exemption of up
to 15 percent of those ineligible to receive benefits under the ABAWD
provision; or (3) meet her or her local jurisdiction's work requirements
that provide for an exception to the ABAWD limits.

6Eligibility requirements are less stringent for elderly disabled persons.

Appendix II: Denial of TANF and Food Stamps Benefits: Legal Framework and
Methodologies

State Exemptions to the Federal Bans on TANF and Food Stamps

percent of the federal poverty guideline (or about $1,313 per month for a
family of two and $1,654 per month for a family of three in 2004), and net
income cannot exceed 100 percent of the poverty guideline (or about $1,010
per month for a family of two and $1,272 per month for a family of three
in 2004). "Gross income" means a household's total, nonexcluded income
before any deductions have been made. "Net income" means gross income
minus allowable deductions. Allowable deductions include a 20 percent
deduction from earned income, dependent child care deductions, and medical
expenses, among others.

According to officials at ACF and FNS, states may implement the provisions
to deny convicted drug felons TANF and food stamps in a variety of ways.
Some states administer the denial of benefits by requiring applicants to
admit to disqualifying felony drug offense convictions at the time that
they apply for benefits. Also according to agency officials, neither
agency regularly collects and assesses data on the number of persons that
self-certify disqualifying drug offenses.7

We reviewed documentation provided by USDA and for states that exempted
some or all convicted drug felons from the federal ban on food stamps, we
reviewed states' laws pertaining to the exemption and we contacted
officials to determine the status of their state's exemptions to the
federal bans on TANF and food stamps. Table 5 shows these statuses and for
states that have enacted exemptions, provides citations to the state laws.

7Both agencies review case files to determine if individuals that were
ineligible to receive the benefits received them in error, and these
reviews may uncover whether a disqualified drug offender received
benefits. In its 2002 report to ACF on closed cases, the state of Maryland
reported that it closed 3 cases (out of the 17,861 cases that it closed
during 2001) because of an individual's disqualifying felony drug
conviction.

Appendix II: Denial of TANF and Food Stamps Benefits: Legal Framework and
Methodologies

      Table 5: Status of States on the TANF and Food Stamp Ban as of 2004

       State      TANF status   Food stamp  State exemptions laws (effective  
                                  status                  date)               
      Alabama         Ban           Ban     
      Alaska          Ban           Ban     
      Arizona         Ban           Ban     
     Arkansas    Modify the ban Modify the  Ark. Code Ann. S: 20-76-409 (July 
                                    ban                   1997)               
    California        Ban       Modify the         Cal. Welf. & Inst. Code S: 
                                    ban              18901.3 (September 2004) 
     Colorado    Modify the ban   Opt Out     Colo. Rev. Stat. S:S: 26-2-305, 
                                                         26-2-706 (July 1997) 
    Connecticut  Modify the ban Modify the    Conn. Gen. Stat. S: 17b-112d    
                                    ban                (July 1997)            
     Delaware         Ban       Modify the   Del. Code Ann. tit. 31, S: 605   
                                    ban                (July 2003)            
      Florida    Modify the ban aModify the Fla. Stat. Ann. ch. 414.095 (May  
                                    ban                   1997)               
      Georgia         Ban           Ban     
      Hawaii     Modify the ban Modify the  Haw. Rev. Stat. S: 346-53.3 (June 
                                    ban                   1997)               
       Idaho     Modify the ban Modify the  Idaho Code S: 56-202 (July 2000)  
                                    ban     
     Illinois    Modify the ban   Opt out   730 Ill. Comp. Stat 5/1-10 (June  
                                                          1997)               
      Indiana    Modify the ban Modify the    Ind. Code S: 12-20-16-6 (July   
                                    ban                   2003)               
       Iowa      Modify the ban Modify the    Iowa Code S: 239B.5 (October    
                                    ban                   1997)               
      Kansas          Ban           Ban     
     Kentucky    Modify the ban Modify the   Ky. Rev. Stat Ann. S: 205.2005   
                                    ban                (July 1998)            
     Louisiana   Modify the ban Modify the  La. Rev. Stat. Ann. 46:233.2(July 
                                    ban                   1997)               
       Maine                                Me. Rev. Stat. Ann. tit. 22, S:S: 
                    Opt out       Opt out   3104(14), 3762(17)                
                                                      (April 2002)            
     Maryland    Modify the ban Modify the  Md. Ann. Code 88A, S:S: 50A, 65   
                                    ban     (July 2000)                       
Massachusetts                            2001 Mass. Adv. Legis. Serv. 177, 
                 Modify the ban   Opt out   S: 4400-1000                      
                                               (December 2001) (reenacted     
                                                       annually )             
                                Modify the  1997 Mich. Pub. Acts 109, S: 622  
     Michigan    Modify the ban     ban     (August 1997)                     
                                                  (reenacted annually)        
     Minnesota   Modify the ban Modify the    Minn. Stat. S: 256J.26 (July    
                                    ban                   1997)               
    Mississippi       Ban           Ban     
     Missouri         Ban           Ban     
      Montana         Ban           Ban     
     Nebraska         Ban       Modify the  Neb. Rev.Stat. S: 68-1017.02      
                                    ban     (August 2003)                     
      Nevada     Modify the ban Modify the  Nev. Rev. Stat S: 422.29316       
                                    ban     (January 1998)                    
New Hampshire    Opt out       Opt out   N.H. Rev. Stat. Ann. S: 167:81-a  
                                            (August 1997)                     
                  bModify the   cModify the    N.J. Stat. Ann. S: 44:10-48    
    New Jersey        ban          banb,             (January 1997)           
                                                   N.J. Stat. Ann. S:         
                                                44:10-48.1(January 2000)      
    New Mexico      Opt out       Opt out   N.M. Stat. Ann. S: 27-2B-11(C)    
                                            (March 2002)                      
     New York                               1997 N.Y. Laws S: 121436 (August  
                   Opt out d     Opt out d                1997)               
                                                  (reenacted annually)        

Appendix II: Denial of TANF and Food Stamps Benefits: Legal Framework and
Methodologies

       State        TANF status     Food stamp       State exemptions laws    
                                      status           (effective date)       
North Carolina Modify the ban  Modify the ban      N.C. Gen. Stat. S:      
                                                     108A-25.2 (July 1997)    
    North Dakota        Ban             Ban       
        Ohio          Opt out         Opt out          Ohio Rev. Code Ann. S: 
                                                     5101.84 (September 1997) 
      Oklahoma        Opt out         Opt out       1997 Okla. Sess. Laws 414 
                                                       S: 28 (September 1997) 
       Oregon         Opt out         Opt out      Or. Rev. Stat. S: 411.119  
                                                          (July 1997)         
    Pennsylvania  Modify the ban  Modify the ban   62 Pa. Stat. S: 405.1(i)   
                                                        (February 2004)       
    Rhode Island                                      R.I. Gen. Laws S:S:     
                    fOpt oute ,     fOpt oute ,        40-5.1-8, 40-6-8       
                                                      (Modify: July 1997,     
                                                      Opt-out: July 2004)     
South Carolina      Bang            Bang       
    South Dakota        Ban             Ban       
                                                         Tenn. Code Ann. S:S: 
     Tennessee    Modify the ban  Modify the ban     71-3-154, 71-5-308 (July 
                                                                        2002) 
       Texas            Ban             Ban       
        Utah      Modify the ban      Opt out     Utah Code Ann. S: 35A-3-311 
                                                          (July 1997)         
      Vermont                                      1997 Vt. Laws 61, S: 131   
                      Opt out         Opt out             (June 1997)         
                                                     (reenacted annually)     
      Virginia          Ban             Ban       
                                                      Wash. Rev. Code S:      
     Washington   hModify the ban   iOpt outh ,            74.08.025          
                                                   (modify: April 1997; opt   
                                                       out: March 2004)       
West Virginia        Ban             Ban       
                                                       Wis. Stat. S:S: 49.79, 
     Wisconsin                                        49.145, 49.148 (October 
                  Modify the ban  Modify the ban                        1997) 

Wyoming Ban Ban

Source: GAO analysis of state laws as of January 2005 and U.S. Department
of Agriculture state reports

aOur analysis of Florida law and conversations with state officials led to
our classification of Florida as a state that has exemptions that modified
the food stamp ban. This classification differs from the USDA
classification, which identifies Florida as a state that fully implements
a lifetime ban on food stamps for convicted drug felons. The discrepancy
in classifications may arise from the different methods used to determine
Florida's status. Annually, USDA surveys states and asks them to report
their status of implementing the ban. Our analysis of Florida law, on the
other hand, identifies exemptions that may not have been reported in the
survey.

bProvision enacted in 1997.

c1997 provision modified in 2000.

dNew York provides for drug use screening of all assistance recipients and
conditions eligibility on further assessment and treatment where social
services officials find them necessary. N.Y. Soc. Serv. S: 132.

eModified in July 1997.

fOpted out in July 2004.

gSome authorities have classified South Carolina as a modification state
(see S.C. Code Ann. S: 43-51190). However state officials report that no
specific exemption to the federal ban has been enacted in South Carolina.

hModified in April 1997.

iOpted out in March 2004.

Appendix II: Denial of TANF and Food Stamps Benefits: Legal Framework and
Methodologies

There are several general types of modifications to the federal ban on
TANF and food stamps among the states that have modified the ban. These
modifications may include one or more of the following elements: (1)
removing from the ban drug felons convicted for drug use or simple
possession, but implementing the ban for drug sellers or traffickers
(e.g., possession with intent to distribute offenses); (2) restoring
benefits to drug felons complying with drug treatment program
requirements; (3) restoring benefits so long as drug felons have negative
drug test results over some period of time; and (4) restoring benefits to
drug felons after various waiting periods, such as a number of years after
conviction or release from prison. State modifications may also include
other conditions. For example, Michigan allows convicted drug felons to
receive benefits provided they do not violate the terms of their parole or
probation and other conditions are met. Tables 6 and 7 show the types of
modifications that states have adopted for the TANF and food stamp bans,
respectively. These tables present general categories of different
modifications, not an exhaustive listing of all specific requirements. For
more detail consult the statutes listed in table 5.

Table 6: Types of State Exemptions (Modifications) to the Federal Ban on
TANF, by States within the Types of Modifications, as of 2004

                  Eligibility to participate in TANF restored

To drug users but not to drug traffickers

Arkansas

Florida

a

         New Jersey Conditional upon undergoing required drug treatment

b

                                    Colorado

Hawaii

Iowa

c

Illinois

Kentucky

a

                              Nevada Pennsylvania

d

                                   Tennessee

Utah

e

                                   Washington

Conditional upon not failing required drug tests Marylandf

Appendix II: Denial of TANF and Food Stamps Benefits: Legal Framework and
Methodologies

                  Eligibility to participate in TANF restored

Minnesotag

Wisconsin

                           Following a waiting period

Louisianaa, h

h

                                 Massachusetts

i

North Carolina Conditional upon not violating conditions of parole or probation
                            supervision Connecticut

Idaho Michigang

          For a 12-month maximum if felon participates in a drug court

Indiana

Source: GAO analysis of selected states' statutes as of January 2005, and
information from state corrections and court offices.

aState also requires drug testing (or demonstration of no drug use in
Nevada).

bState requires individuals to take action toward rehabilitation, such as,
but not limited to, drug treatment.

cState requires either a 2-year waiting period or participation in drug
treatment, and also restricts the exemption to drug offenders who are not
class X or class 1 felons.

dState restricts exemption to drug offenders who are not class A felons.

eState also requires there be no more than one conviction in a 3-year
period.

fFor current recipients, state also imposes a 1-year ban from date of drug
conviction.

gState also requires payments must be made to vendors or authorized
representatives.

hState imposes a 1-year waiting period.

iState imposes a 6-month waiting period and also requires drug treatment.

Appendix II: Denial of TANF and Food Stamps Benefits: Legal Framework and
Methodologies

Table 7: Types of State Exemptions or Modifications of the Federal Ban on
Food Stamps, by States within the Types of Modifications, as of 2004

               Eligibility to participate in food stamps restored

To drug users but not to drug traffickers

Arkansas

Californiaa

                                 Delaware a, b

Florida

Nebraskaa

              Conditional upon undergoing required drug treatment

Hawaii

Iowa Kentucky

b

Nevada

b

                                   New Jersey

                                  Pennsylvania

c

          Tennessee Conditional upon not failing required drug testing

Marylandd

Minnesotae Wisconsin Following a waiting period

f

Louisiana North Carolinag

Conditional upon not violating conditions of probation or parole
supervision Connecticut Idaho

e

Michigan

For a maximum of 12 months if felon participates in a drug court;
available only through July 1, 2005, unless extended

Indiana

Source: GAO analysis of selected states' laws and officials from state
corrections departments court offices.

aState also requires drug treatment.

bState also requires drug testing (or demonstration of no drug use in
Nevada).

cState restricts exemption to drug offenders who are not class A felons.

dFor current recipients, state also imposes a 1-year waiting period from
the date of the conviction.

eState will also impose a 1-year ban from date of drug convictions for
current recipients

Appendix II: Denial of TANF and Food Stamps Benefits: Legal Framework and
Methodologies

fState imposes a 1-year waiting period and also requires a negative drug
test result.

gState imposes a 6-month waiting period and also requires drug treatment.

Estimating the Percentage of Drug Arrests within States That Implement,
Modify, or Opt Out of the Bans on TANF and Food Stamps

To obtain a general assessment of the degree to which state decisions to
modify or opt out of the federal bans on TANF and food stamps exempt drug
felons from the federal ban, we estimated the percentage of drug arrests
that occurred within three groupings of states: (1) those that fully
implement the bans, (2) those that have modified them, and (3) those that
have completely opted out of the bans. We used drug arrests as a proxy for
drug convictions, as state-level data on the number of drug felony
convictions are not available. We analyzed data from the 2002 Crime in the
United States: Uniform Crime Reports on the number of persons arrested for
drug offenses in each of the 50 states. Table 8 reports the relative
distributions of drug arrests for the states falling into each category
for the TANF and food stamp bans.

Table 8: Number of States in Each TANF and Food Stamp Ban Status Category
and Percentage of Drug Offense Arrests in 2002 Occurring in the States
within Each Ban Status Category

                            TANF ban Food stamp ban

                  TANF/food                                     
                  stamp ban  Number of  Percentage of Number of Percentage of 
            status category    statesa drug arrestsa    statesa drug arrestsa 
              Implement the         18          43.8%        15         22.0% 
                  ban fully                                     
           Modified the ban         23          46.6%        20         57.1% 
           Opted out of the          9           9.6%        15         20.9% 
                        ban                                     
                      Total         50         100.0%        50        100.0% 

Source: GAO analysis of Uniform Crime Report data for 2002.

aExcludes the District of Columbia.

Appendix II: Denial of TANF and Food Stamps Benefits: Legal Framework and
Methodologies

Methods for Estimating the Percentage of Drug Felons Potentially Affected
by the TANF and Food Stamp Bans

To assess the potential impacts of the bans on TANF and food stamps, we
estimated the percentage of a population of drug felons released from
prison that would be eligible to receive TANF, and but for their drug
offense conviction could receive the benefit. By potentially affected, we
refer to convicted drug felons that we estimated met selected eligibility
criteria to participate in these benefit programs. According to our use of
the term "impact," only those drug felons who were otherwise eligible to
receive benefits actually stood to lose benefits as a result of the bans,
and could therefore be affected by the bans.

To determine the percentage of drug felons that met selected eligibility
criteria, we used data from the Bureau of Justice Statistics' Survey of
Inmates of State Correctional Facilities in 1997. This survey is based
upon a nationally representative sample of persons in state prisons during
July 1997. The 1997 data represent the most recently available data from
this recurrent survey, which BJS conducts about every 5 years. We used
information from the survey about prisoners' parental status, employment,
and income prior to incarceration in developing our estimates of the
percentages of drug offenders that were custodial parents and had incomes
within allowable maximums to qualify for the benefits.

For both benefits, we provide estimates that are based on drug offenders
released from prison during 2001 in the subset of states that fully
implemented either the ban on TANF or food stamps. To the extent possible,
we limited the data on drug offenders released from prison to those who
entered prison during 1997 or thereafter. This allowed for a period of
time between the possible date that a drug felony offense was committed
and the date that an offender entered prison, and in this way, we took
into account the implementation date of the ban, which was August 22,
1996.

Because of data limitations, we did not attempt to develop estimates for
states that modified the bans. For example, some states' exemptions to the
bans allow that convicted drug offenders may receive benefits (provided
that they are eligible for them) if they do not fail a drug test, if they
undergo required drug treatment, if they do not violate conditions of
probation or parole supervision, or if they meet certain other conditions.
The data that we used did not include this information; therefore, we
could not estimate the potential impacts of the bans in the states that
modified the bans.

We developed estimates of the potential impact of these bans on the
population of released prisoners for 1 year, 2001, the most recent year
for

Appendix II: Denial of TANF and Food Stamps Benefits: Legal Framework and
                                 Methodologies

which we obtained data. We did not attempt to develop estimates for all
persons potentially affected by the bans since they went into effect
during 1996. We discuss the problems associated with estimating all
persons potentially affected by the bans in a later section of this
appendix.

Data and Methods Used to Estimate the Potential Impacts of the TANF Ban

To estimate the potential impacts of the TANF ban, we obtained data from
states on drug felons released from prison, and using these data, we
applied estimates of the percentages that met selected TANF eligibility
requirements. These methods are described more fully below.

Methods to Obtain State-Specific Data on the Number of Drug Felons
Released from Prison

For 14 of the 18 states that fully implement the ban on TANF, we obtained
data on the number of drug offenders released from prison during 2001. We
used two sources of data: (1) the Bureau of Justice Statistics' National
Corrections Reporting Program (NCRP) and (2) data from selected other
states. From NCRP, we obtained counts of the number of drug felons
released from prison during 2001, given that they were committed into
prison in 1997 or thereafter for a new conviction that contained a drug
offense. We chose 1997 because the TANF ban went into effect on August 22,
1996, and data on the date that ex-prisoners committed their drug
offense-which is the factor that determines whether they are under the
ban-were not available in the data that we used. From the other states, we
obtained comparable data on the number of drug offenders released from
prison.

The 14 states for which we obtained data were Alabama, Arizona,
California, Georgia, Kansas, Mississippi, Missouri, Nebraska, North
Dakota, South Carolina, South Dakota, Texas, Virginia, and West Virginia.
The 14 states account for approximately 97 percent of the population in
the 18 states that maintain the ban on TANF for drug felons. For the 4
states that were excluded from our analysis--Alaska, Delaware, Montana,
and Wyoming-we were unable to obtain data on released prisoners. We also
excluded from our analysis states that may have implemented the ban in
2001 but as of January 2005 had modified or opted out of the ban.

Across the 14 states, about 96,000 drug offenders were released from
prison during 2001, given that they had been admitted during 1997 or
thereafter. This population of all drug felons released from prison
includes those who were sentenced to prison following their conviction for
a drug offense, and it also includes offenders who entered prison because
they

Appendix II: Denial of TANF and Food Stamps Benefits: Legal Framework and
Methodologies

had violated conditions of supervision. Among offenders who entered prison
for a violation of conditions of supervision, some may have committed
their offenses before the TANF ban went into effect, and they would not be
subject to the ban. However, some of the released prisoners who had
violated conditions of supervision may have been convicted after the ban
went into effect, but the available information reported only the date of
admission for the violation and not for the original sentence. These
offenders should be included among the population of drug felons that are
subject to the ban. Hence, the population of all released prisoners might
over estimate the number of drug offenders in these 14 states who
committed offenses after the TANF ban had gone into effect.

About 51,000 of the drug offenders released from prison during 2001 were
those who had been admitted into prison during 1997 or immediately after
their conviction. While this population of released drug offenders
includes those whose prison sentence occurred after the ban went into
effect, this number may under estimate the number of drug felons in these
states who were subject to the ban. It may do so because it will exclude
the parole violators who had initially been committed after 1997 but whose
most recent commitment was for a violation of parole that also occurred
after 1997.8

About 87 percent of all drug offenders released from prison during 2001 in
the 14 states for which we obtained data were males, as were about 86
percent of the first releases. Females constituted 13 percent of all
releases and 14 percent of first releases (table 9).

8Our estimates of the percentage of drug felons that meet selected
eligibility conditions are relatively unaffected by whether we base them
on all drug offenders released from prison or on first releases only.

Appendix II: Denial of TANF and Food Stamps Benefits: Legal Framework and
Methodologies

Table 9: Estimated Number of Drug Felons Released from Prison during 2001,
in 14 of 18 States That Currently Maintain the Ban on TANF

Percentage	Number of prisoner distribution of 	releases prisoner releases	

                      All drug felons released from prison

Total 95,940 100%

Males 83,528 87%

Females 12,412 13%

First releases of drug felons from a new court commitmenta

Total 51,332 100%

Males 44,230 86%

Females 7,092 14%

Source: GAO analysis of Bureau of Justice Statistics National Corrections
Reporting Program data for 2001 and data from selected states.

Note: The 14 states are Alabama, Arizona, California, Georgia, Kansas,
Mississippi, Missouri, Nebraska, North Dakota, South Carolina, South
Dakota, Texas, Virginia, and West Virginia. The 4 excluded states are
Alaska, Delaware, Montana, and Wyoming. We also excluded states that
implemented the ban in 2001 but have since modified or opted out of the
ban.

aNew court commitments include only those offenders who were sentenced to
prison on a new conviction for a drug offense.

Methods of Estimating the Percentage of Drug Felons That Meet TANF
Eligibility Requirements

To receive TANF assistance, an assistance unit (such as a household) must
meet the state-mandated definition of a needy family: It must either
contain at least one child living with an adult relative or consist of a
pregnant woman. The adult guardian must be related to the child by blood,
adoption, or marriage (or, if the state provides, the adult may stand in
for parents if none exist). Further, TANF recipients must in general be
either U.S. citizens or qualified aliens who entered the United States
prior to the passage of PRWORA on August 22, 1996, or who have lived in
the United States for a period of 5 years. States may also impose other
conditions for receipt of TANF benefits.

We used data from the 1997 version of the BJS Inmate Survey to estimate
the percentage of drug offenders who were custodial parents and who had
monthly incomes within state-determined earned income limits. For
estimation purposes, we defined a drug offender in the inmate survey as a
custodial parent if the offender met three conditions: (1) reported being
the parent of at least one minor child, (2) reported living with the child
prior to being incarcerated, and (3) reported that the child was not in
foster care or agency care while the offender was in prison. We computed
the number of prisoners who met these conditions, and from these counts,

Appendix II: Denial of TANF and Food Stamps Benefits: Legal Framework and
Methodologies

we estimated the percentages of drug offenders that met these conditions.
As the data were drawn from a sample, we used weighting factors provided
by BJS that were based on the original probabilities of being selected
into the sample that were adjusted for nonresponse and information about
the sex, race, age, prison security level, and type of offense of the
total prison population to produce national-level estimates. We estimated
the percentages separately by gender and region of the country. Table 10
shows our estimates of the percentage of convicted drug felons that were
reported to be parents and custodial parents (based on our definition) of
minor children.

Table 10: Estimated Parental Status of Drug Offenders, by Gender and
Region of the Country

Percentage of drug

Percentage of drug offenders that were Gender and region of offenders that
were custodial parents of minor drug offenders parents of minor children
childrena

                                Female offenders

                          Northeast        64.1%                        35.1% 
                            Midwest        69.9%                        48.1% 
                              South        66.4%                        44.6% 
                               West        62.5%                        29.9% 

Male offenders

                          Northeast        61.4%                        24.5% 
                            Midwest        66.6%                        31.0% 
                              South        65.0%                        29.1% 
                               West        65.3%                        30.8% 

Source: GAO analysis of BJS Survey of Inmates of State Correctional
Facilities, 1997.

aWe defined custodial parents as inmates that had minor children, lived
with them prior to their incarceration, and whose children were not in
foster care or agency care while they were incarcerated.

We also estimated the income distributions for drug offender parents who
reported living with their children. In the BJS survey, income is reported
as the offender's total income in the month prior to the arrest leading to
the incarceration. Monthly income can be from any source and may include
illegal income. We omitted from our analysis those offenders who reported
income from illegal sources, and we included only offenders who reported
earned income or who were unemployed prior to their imprisonment.
Offenders who were unemployed prior to their imprisonment received a value
of zero for earned income. We estimated

Appendix II: Denial of TANF and Food Stamps Benefits: Legal Framework and
Methodologies

the income distributions separately by gender and region to account for
differences in employment and earnings between male and female offenders,
and offenders in different states. We applied the regional income
distributions to all states within a region, as the BJS data did not
report the state in which the offender was incarcerated.

From the income distributions, we estimated the gender-specific
percentages of drug offenders who had incomes at or below statedetermined
earned income limits. The BJS inmate survey data report income in
intervals, and in many cases, the intervals do not correspond directly
with the state earned income limits. Therefore, we selected income
intervals that were as near to the state earned income limits as feasible.
We generally selected two income intervals for each state: one that
contained the state earned income limit level but whose lower bound was
less than the state level, and one that contained the state earned income
limit but whose upper bound was above the state level. In this way, we
obtained upper-and lower-bound estimates of the potential impacts of the
TANF ban.

To obtain estimates of the percentage of drug offenders released from
prison who were both custodial parents and were income eligible for TANF,
as defined above, we applied the gender-specific estimates of the
percentage of prisoners in each region of the country that met the
specific TANF eligibility criteria to state-specific counts of the number
of drug felons released from prison. We used the region of the country
within which a state was located to obtain estimates for a specific state.
The result of these operations was to obtain estimates of the percentage
of drug offenders released from prison who were both custodial parents and
were income eligible for TANF, as defined above. The estimated percentages
of drug offenders released from prison that met these conditions are shown
in table 11.

Appendix II: Denial of TANF and Food Stamps Benefits: Legal Framework and
                                 Methodologies

Table 11: Estimated Percentages of Drug Offenders Released from Prison in
2001 That Met Selected Guardianship and Earned Income Limits to Receive
TANF benefits

Estimated percentage of drug offenders released from prison that were custodial
                     parentsa and income eligible for TANF

                   Lower bound Upper bound Total 13.6% 16.4%

                               Males 11.7% 14.5%

                              Females 25.4% 28.4%

Source: GAO analysis of BJS inmate survey data and data for selected
states as provided by state corrections and court officials.

a"Custodial parent" is defined as one who lived with minor children prior
to incarceration and whose minor children were not in foster care while
the offender was incarcerated.

Limitations to the Estimates

We were unable to take into account all of the factors that determine
whether drug offenders met the eligibility criteria to receive TANF. Some
of these factors could contribute to reducing the estimated percentages of
drug offenders who were otherwise eligible; others could possibly
contribute to increasing the estimated percentages. In addition, our
estimates for drug offenders released from prison in a given year do not
apply to drug felons who were sentenced to probation. Finally, we are
unable to provide an estimate of the percentage of drug offenders
potentially affected by the ban for the entire period since it was
implemented.

Factors Contributing to a Decrease in the Percentage of Drug Felons That
Would Be Otherwise Eligible to Receive TANF

Data limitations preclude our explicitly taking into account all of the
factors that are related to TANF eligibility. Factors affecting TANF
eligibility for which we do not have data are the citizenship status and
length of residency of noncitizens, state-imposed work requirements to
receive TANF, and individual choices to participate in the program. While
we were unable to estimate the effect of these factors on our estimated
percentages that might have been eligible to receive TANF, these factors
would contribute to lowering our estimates of the percentage of drug
offenders released from prison that might have been eligible to receive
TANF.

Several of the states whose data we analyzed have relatively large
populations of noncitizens. In general, to qualify for TANF, aliens must
have at least 5 years of residence in the United States since August 22,

Appendix II: Denial of TANF and Food Stamps Benefits: Legal Framework and
Methodologies

1996. Given that our estimates are for 2001, it is unlikely that many
aliens among convicted drug felons would have qualified for TANF. Hence,
taking the alien qualification into account would lower our estimates of
the percentage of drug felons potentially affected by the TANF ban. For
2003, ACF reports that 8 percent of adult TANF recipients were qualified
aliens.

Individuals within needy families who do not participate in
statedetermined work requirements could lose their TANF eligibility.
Failing to comply with work requirements would reduce the percentage of
drug offenders that were otherwise eligible to receive TANF.

In the general population, adult males constitute comparatively small
numbers of TANF recipients. According to ACF, in 2001 adult males
constituted about 9 percent of all adult TANF recipients. If we applied
the general population adult male TANF recipiency rate to our estimates of
the percentage of all drug offenders released from prison, our estimated
impact of the TANF ban would be revised downward to about 4 percent of all
of drug offenders released from prison in 2001.

Change in Status after Release from Prison Could Increase the Percentage
of Drug Felons Affected by the Ban

One factor that could change the estimated percentage of convicted drug
felons eligible to receive TANF benefits and therefore potentially
affected by the ban is a change in a felon's eligibility to receive TANF.
Our estimates of the percentage of prisoners that may be eligible to
receive TANF are based on attributes existing at the time that offenders
were in prison. Upon release, these attributes may change, and an offender
might become otherwise eligible for TANF and therefore potentially be
affected by the ban. For example, if a drug offender was reunited with his
or her children after release and met other eligibility requirements, this
would contribute to increasing the percentage of released prisoners that
were eligible to receive TANF. Alternatively, imprisonment may be a factor
that reduces contact with children and therefore contributes to decreasing
the percentage of drug offenders released from prison that are eligible to
receive TANF.

Drug Felons Sentenced to In recent years, drug felons sentenced to
probation account for about one-

Probation 	third of all convicted and sentenced drug felons. We did not
apply the information about drug offenders in prison to the drug felons
sentenced to probation. This is because we do not have data on the
parental and income characteristics of drug felons sentenced to probation.
To the extent the drug felons sentenced to probation have characteristics
similar to those of

Appendix II: Denial of TANF and Food Stamps Benefits: Legal Framework and
Methodologies

drug felons released from prison, the estimated percentage of probationers
that may be eligible to receive benefits would be similar to those
estimated percentages among released prisoners. However, if income levels
and other factors differ between probationers and prisoners, this could
affect the estimates of the percentages that would be eligible to receive
benefits.

Drug Offenders Affected since the Ban Went into Effect

We do not provide an estimate of all drug offenders potentially affected
by the ban on TANF since it went into effect. We were unable to obtain
data on the number of persons convicted of drug felonies since the ban
went into effect in 1996, as only limited data are available. Over time,
an individual's attributes that are related to TANF eligibility may
change. Convicted drug felons who did not have characteristics that would
make them eligible to receive TANF at one point in time could develop
these attributes at a later point in time. Conversely, the circumstances
of convicted drug felons who at one point in time were otherwise eligible
to receive TANF could change so that they are no longer otherwise
eligible. To understand the long-term impacts of the ban therefore would
require data that track individuals over time and measure changes in their
characteristics that are related to TANF eligibility. We know of no such
national data on drug offenders.

Our estimates of the percentage of drug offenders released from prison in
a given year who are potentially affected by the ban represent lowerbound
estimates of the proportion of drug offenders released from prison during
that year that would ever be affected by the ban. If, among those released
from prison and estimated not to be eligible to receive TANF, any persons
became eligible at a later date, this would increase the percentage of
persons potentially affected by the ban. Consequently, the long-term
impacts of the ban would be greater than the impacts that we estimated for
the 1-year release cohort. Similarly, if the 1-year estimates of the
percentage potentially affected by the ban were to hold over time, then a
larger percentage of all convicted drug felons would be potentially
affected by the ban since its inception than the percentages that we
estimated for 1 year.

Data and Methods Used to Estimate the Potential Impact of the Ban on Food
Stamps

We focused our analysis of the potential impact of the ban on food stamps
on drug offenders that were reported to be custodial parents of minor
children. According to USDA, in fiscal year 2003, adult households with
children (containing either one or two adults) constituted 73 percent of

Appendix II: Denial of TANF and Food Stamps Benefits: Legal Framework and
Methodologies

food stamp recipients. Consequently, this is likely to be the largest
group of drug offenders that could be affected by the food stamp ban. We
were unable to develop a quantitative estimate of the percentage of
able-bodied adults without dependents (ABAWD) that could be affected by
the food stamp ban. ABAWDs, in general, may receive food stamps for 3
months within a given 3-year period or longer if they adhere to the work
requirements specifically laid out for ABAWDs. However, we were unable to
determine which drug offenders constituted the potential pool of ABAWDs.
We further gave the potential ABAWDs recipients separate consideration
because, according to USDA reports, in 2003, they constitute 2.5 percent
of food stamp recipients nationwide even though they form a large share of
the general population of such persons. We also did not attempt to develop
an estimate of the impact of the ban for elderly and disabled drug
offenders.

For 2003, USDA reported that adult households with children (containing
either one or two adults) constitute 73 percent of food stamp recipients.
In contrast, elderly individuals living alone constitute 6 percent of food
stamp recipients, and disabled nonelderly individuals living alone
constitute 5 percent of food stamp recipients. Single-adult
households-which according to USDA do not contain children, elderly
individuals, or disabled individuals-constitute 6 percent of food stamp
recipients. Therefore, adult households with children receive food stamps
at a rate greater than 12 times the rate at which single-adult households
receive food stamps. The percentage of single adult households receiving
food stamps is higher than the percentage of ABAWDs receiving food stamps
because an individual is not considered an ABAWD if the person is
pregnant, exempt from work registration, or over 50 years of age.9

For 12 of 15 states that maintain the full ban on food stamps, we obtained
data on the drug felons released from prison during 2001 (given that they
entered prison during 1997 or thereafter). The 12 states are Alabama,
Arizona, Georgia, Kansas, Mississippi, Missouri, North Dakota, South
Carolina, South Dakota, Texas, Virginia, and West Virginia. The 3 excluded
states for which we were unable to obtain data were Alaska, Montana, and
Wyoming. A total of 67,000 drug offenders were released in 2001 in the 12

9Since ABAWDs cannot be older than 50, and USDA defines an elderly person
as being 60 or older, someone who is over 50 and under 60 is considered
neither an ABAWD nor elderly.

Appendix II: Denial of TANF and Food Stamps Benefits: Legal Framework and
Methodologies

states, and of these, 30,000 were first releases from new court
commitments.

We used the BJS inmate survey data to estimate the percentage of drug
felony prisoners who were parents living with their minor children and
whose children were not in foster care while they were incarcerated. This
was our operational definition for a custodial parent. For these, we
estimated the percentage who had gross incomes within the poverty
thresholds, based on estimates of family size. Food stamp eligibility is
based on gross and net income tests. Because data on the deductions that
are used in determining whether households meet the net income tests were
not available, our estimates are at best gross income tests. We are unable
to determine how our use of the gross income test alone affects our
estimates of the percentage of drug felons released from prison that would
have been eligible to receive food stamps.

In general, ABAWDs may receive food stamp benefits for an extended
duration as long as they meet ABAWD-specific work requirements. This means
that a large percentage of drug felons could be eligible to receive, and
therefore potentially be denied, food stamps as long as they fell within
the income threshold to receive food stamps. However, among all food stamp
recipients, ABAWDs constitute only 2.5 percent of the total. Hence, while
we cannot estimate the percentage of ABAWDs within the drug offender pool
that would be otherwise eligible to receive food stamps, the ABAWD
participation rate in food stamps in general would suggest that relatively
few drug offenders who fall into this category would participate in the
program.

We assessed the impacts of the denial of TANF and food stamp benefits by
estimating the percentage of convicted drug felons released from prison
who were otherwise eligible to receive the benefits. To assess whether
impacts vary by race, we first assessed whether the percentage of drug
offenders who met the same eligibility requirements that we used to assess
the overall impacts of the TANF and food stamp bans varied according to
race. For example, if larger proportions of black than white drug
offenders were custodial parents of minor children and had earned income
that permitted them to qualify for TANF, then we would expect to find
larger percentages of black drug offenders to be affected by the TANF ban,
regardless of the racial composition of the group of all drug offenders
released from prison.

We used the BJS inmate survey data to compare the estimated percentages of
black and white drug offenders who were custodial parents (as we

Data on the Race of Drug Offenders Who Meet Selected TANF Eligibility
Criteria

Appendix II: Denial of TANF and Food Stamps Benefits: Legal Framework and
Methodologies

defined the term previously) and had earned incomes that could qualify
them to receive TANF. As before, we estimated these percentages by gender
and region. Our estimates indicated that in one region (the South), the
percentage of black female drug offenders who were otherwise eligible to
receive TANF differed from the percentage of otherwise eligible white
female drug offenders. A larger percentage of black female drug offenders
in that region were estimated to be eligible to receive TANF than white
female drug offenders in the region. Among male drug offenders, we
estimated differences in eligibility for TANF in two regions. For both
female and male drug offenders, the differences in estimated TANF
eligibility arose from differences in incomes, as there were no
differences in the percentage of black and white drug offenders that were
estimated to be custodial parents.

Appendix III: Denial of Federal Postsecondary Education Benefits to Drug
Offenders

Legal Framework for Denying Federal Higher Education Benefits

This appendix describes the legal framework for denying federal higher
education benefits to drug offenders, how the federal provision is
administered, our methods for estimating the number of students affected
by the federal provisions, and the impacts of the federal provision.

The Higher Education Act of 1965,1 as amended, provides for the suspension
of certain federal higher education benefits to students who have been
convicted for the possession or sale of a controlled substance under
federal or state law.2 The controlled substance offense may be either a
felony or a misdemeanor. Federal higher education benefits that are denied
to such individuals include student loans, Pell Grants, Supplemental
Educational Opportunity Grants, and the Federal Work-Study program.3

The Higher Education Act provision outlines different periods for which
such drug offenders are ineligible to receive certain federal higher
education benefits, depending upon the type and number of controlled
substance convictions. The period of ineligibility begins on the date of
conviction and ends after a specified interval. Table 12 illustrates the
period of ineligibility for the federal higher education benefits,
according to the type and number of convictions.

1Pub. L. No. 89-329, 79 Stat. 1219 (1965).

2See 20 U.S.C. S: 2091(r). This provision was added by the Higher
Education Amendments of 1998, Pub. L. No. 105-244, 112 Stat. 1581 (1998).

3Student loans include the Federal Family Education Loan (FFEL) Program
and the William D. Ford Federal Direct Loan Program. Higher education
institutions usually participate in one or the other program. Depending
upon the student's estimated financial need, students may qualify for
subsidized Stafford loans or, irrespective of financial need, unsubsidized
Stafford loans, through either program. Pell Grants are federal grants
that are awarded to students with financial need who have not received
their first bachelor's degree or who are enrolled in programs that lead to
teacher certification or licensure. Supplemental Educational Opportunity
Grants are awarded for the purpose of providing grants to needy
undergraduate students. The Federal Work-Study program provides part-time
employment to needy undergraduate and graduate students attending
participating institutions.

    Appendix III: Denial of Federal Postsecondary Education Benefits to Drug
                                   Offenders

Table 12: Periods of Ineligibility for Federal Higher Education Assistance
Based on Type and Number of Controlled Substance Offense Convictions

Type and number of drug convictions Period of Ineligibility

                      Possession of a controlled substance

First offense 1 year

                             Second offense 2 years

                            Third offense Indefinite

                         Sale of a controlled substance

First offense 2 years

Second offense Indefinite

Source: GAO analysis of Section 484 of the Higher Education Act of 1965,
as amended (see 20 U.S.C. S: 1091(r)).

This Higher Education Act provision allows for eligibility for federal
higher education to be restored prior to the end of the period of
ineligibility if either one of two conditions is met. First, a student
satisfactorily completes a drug rehabilitation program that includes two
unannounced drug tests and complies with criteria established by the
Secretary of Education. Second, a student has his or her drug conviction
reversed, set aside, or nullified.

Administration of the Denial of Federal Higher Education Benefits

The provisions of federal law mandating the denial of certain federal
higher education benefits were implemented beginning in July 2000 by
requiring students who applied for federal assistance to self-report
disqualifying drug convictions. Students must self-report disqualifying
drug convictions through the Department of Education's Free Application
for Federal Student Aid, a form that any student who wishes to receive
federal student aid must complete.4 The FAFSA is available online and is
free to use. ED uses the information that applicants provide on their
FAFSA to determine their eligibility for aid from the Federal Student Aid
(FSA) programs. Colleges and universities in 49 states also use
information from the FAFSA in making their financial aid determinations.
ED provides participating colleges and universities with a formula to use
when making decisions about financial assistance.

4The FAFSA provides a worksheet that guides students in determining
whether they have a disqualifying drug conviction.

Appendix III: Denial of Federal Postsecondary Education Benefits to Drug
Offenders

Applicants who either report that they have a drug conviction that affects
their eligibility or those applicants who do not answer the question about
drug convictions are automatically ineligible to receive federal higher
education assistance in the academic year for which they sought aid.5
(Below, we refer to this group as FAFSA ineligibles.) The drug conviction
worksheet of the FAFSA also notifies students that even though a drug
conviction may render them ineligible to receive federal higher education
assistance in the application year, individuals may still be eligible to
receive aid from their state or their academic institution.

Methods for Estimating the Numbers of Persons and Amounts of Federal
Assistance Affected by the Provisions

For several reasons, not all of the FAFSA applicants who self-report a
disqualifying drug conviction would otherwise have been eligible to
receive federal assistance; hence, the number of applications containing
self-reported disqualifying drug offenses overstates the number of persons
denied federal postsecondary education assistance because of a drug
offense conviction.

First, not all FAFSA applicants are eligible to receive all types of
federal postsecondary education assistance. For example, some applicants
may have incomes above the levels required to receive Pell Grants, and
even if they self-reported a disqualifying drug conviction, they would not
have been eligible to receive Pell Grants. Second, ED officials indicated
that not all FAFSA applicants become enrolled in postsecondary education
institutions, and these applicants are not eligible to receive federal
postsecondary education assistance. Third, some individuals may complete
the FAFSA application more than one time, and by counting only the number
of applications, some individuals may be double-counted.

To assess the impacts of the Higher Education Act's provisions that render
students with disqualifying controlled substances convictions ineligible
to receive federal postsecondary education assistance, we estimated the
number of students who self-reported a disqualifying drug offense and,
absent the controlled substances convictions provisions of the Higher
Education Act, would have been qualified to receive assistance but because
of the provisions would not have received assistance.

5This rule did not apply to applicants during the 2000-2001 academic year,
the first year that the drug conviction question appeared on the FAFSA.
During 2000-2001, students who did not answer the question were permitted
to receive federal aid.

Appendix III: Denial of Federal Postsecondary Education Benefits to Drug
Offenders

We developed estimates of the number of applicants for Pell Grants and
subsidized and unsubsidized Stafford loans (two of the best-funded federal
postsecondary education assistance programs) and the total amounts of
assistance lost, because of their self-reported controlled substances
convictions. Our methods for estimating these quantities are as follows:

o  	To estimate the number of students who were denied Pell Grants in a
given year, we use ED data on the number of FAFSA applicants that either
self-reported a disqualifying drug offense conviction or left this
question blank, the group that we labeled as FAFSA ineligibles.

o  	As applicants must meet needs-based criteria to make them eligible to
receive Pell Grants, we then use ED data on the percentage of FAFSA
applicants that were eligible to receive Pell Grants; we call this second
group Pell Grant eligibles.

o  	We use ED data on the percentage of Pell Grant eligibles that actually
received Pell Grants, as not all of the students who were eligible to
receive Pell Grants received them. By multiplying these quantities, we
obtained a rough estimate of the number of persons who, absent the
disqualifying drug offense conviction, would have received Pell Grants.

o  	To estimate the dollar amount of Pell Grants that these recipients
would have received, we multiplied the average amount of Pell Grants
(which we obtained from ED) by the estimated number of students denied
Pell Grants.

To estimate the number of student loan recipients who were denied
assistance because of disqualifying drug convictions, we followed a method
similar to the one that we used to estimate the numbers denied Pell
Grants.

o  	Specifically, beginning with the data on FAFSA ineligibles, we applied
to this number the percentage of all FAFSA applicants that received a
student loan. We could not obtain an estimate of the number of FAFSA
applicants that were eligible to receive student loans because, as ED
reports, unlike Pell Grants, where there are income limitations that can
be used to determine eligibility, with student loans, eligibility is
determined by both income and institution-specific factors (such as
tuition). Thus, our estimate is of the number of FAFSA ineligibles that
would have received a student loan but for their controlled substances
convictions.

o  	To estimate the amount of student loans denied, we multiplied our
estimate of the number denied student loans by the average amount of a
student loan.

    Appendix III: Denial of Federal Postsecondary Education Benefits to Drug
                                   Offenders

Critical Assumptions in Estimating the Number of Pell Grant and Student
Loan Recipients Denied Federal Benefits because of Disqualifying
Controlled Substances Convictions

In order to create our estimates for the number of individuals who would
have received a Pell Grant or a student loan if not for their drug
conviction, we assume that the characteristics of FAFSA eligibles are the
same as the characteristics of FAFSA ineligibles. This assumption means
that the percentage of FAFSA applicants who are eligible to receive
federal higher education assistance should be the same for FAFSA
ineligibles (apart from the drug conviction).

Income is an important determinant of eligibility for both Pell Grants and
student loans. Specifically, financial need is determined by ED using a
standard formula established by Congress to evaluate the applicant's FAFSA
and to determine the student's Expected Family Contribution (EFC). The EFC
calculation includes various data elements including income, number of
dependents, net assets, marital status, and other specified additional
expenses incurred. Different assessment rates are used for dependent
students, independent students without dependents, and independent
students with dependents. After filing the FAFSA, a student is notified if
he or she is eligible for a federal Pell Grant and of the student's EFC.

On the one hand, if FAFSA ineligibles on average have lower incomes than
FAFSA eligibles, then our estimates of the number of students denied
benefits are likely to be underestimates of the true number denied
benefits. This is because we rely on the information about eligibility for
Pell Grants and student loans from the persons who were eligible to
receive them, not from the population who are otherwise eligible but for
their disqualifying drug convictions. On the other hand, if FAFSA
ineligibles are less likely to be enrolled in postsecondary education
institutions, as compared with FAFSA eligibles, then our estimates of the
number denied benefits are likely to overestimate the true number denied
benefits.

Data Used to Estimate the Numbers and Amounts Denied Federal Assistance

Table 13 shows the data that we used to estimate the numbers and amounts
of federal postsecondary education assistance that was forgone to students
who, absent their controlled substances convictions, would have received
federal postsecondary education assistance. The data are provided annually
for academic years 2001-2002 through 2003-2004. The key data elements used
to estimate the numbers and amounts of federal assistance denied include

o  the number of FAFSA applicants and FAFSA ineligibles,

o  the percentage of Pell Grant eligibles among all FAFSA applicants,

    Appendix III: Denial of Federal Postsecondary Education Benefits to Drug
                                   Offenders

o  the percentage of Pell Grant recipients among Pell Grant eligibles,

o  the average amount of Pell Grant received,

o  the percentage of FAFSA applicants that received student loans, and

o  the average amount of student loan received.

The number of FAFSA ineligibles declined from 58,929 in academic year
2001-2002 to 41,061 in academic year 2003-2004.6 We note that FAFSA
ineligibles amount to less than 0.5 percent of all FAFSA applications.

Table 13: Data Used to Estimate the Number of Persons Denied Federal
Education Assistance because of Drug Convictions and Estimated Amount of
Federal Assistance Denied

Academic year 2001-2002 2002-2003 2003-2004

         Number of FAFSA applications        10,961,421 12,021,249 13,009,596 
               FAFSA ineligibles                 58,929     42,537     41,061 
         Self-reported drug conviction           48,642     37,451     34,914 
     Drug conviction question not answered       10,287      5,086      6,147 
FAFSA ineligibles as a percentage of all        0.5%       0.4%       0.3% 
               FAFSA applicants                                    

Pell Grant eligibles, recipients, and amounts Percentage of Pell Grant eligibles
                   among FAFSA applicants 51.5% 52.4% 58.0% a

Percentage of Pell Grant recipients among Pell Grant eligibles 76.9% 75.9% 77.0%
                                       a

    Average amount of Pell Grant $2,298 $2,436 $2,467b Stafford student loan
                            recipients and amountsc

    Percentage of FAFSA applicants receiving student   70.0% a 70.0%a 70.0% a 
                          loans                                       
             Average amount of student loan                           
                     Subsidized loan                    $3,378 $3,405 $3,469b 
                    Unsubsidized loan                   $3,970 $4,063 $4,162b 

Sources: GAO analysis of FAFSA data provided by the Department of
Education's FSA Central Processing System; Department of Education data on
Pell Grant and student loan recipients.

aEstimates provided by the Congressional Budget Office.

bEstimates provided by the Department of Education.

cThe amounts that may be borrowed are the same whether a student receives
a Stafford Loan through the Federal Family Education Loan (FFEL) Program
or through the William D. Ford Federal Direct Loan Program.

6We excluded from this analysis the data for academic year 2000-2001, in
which 9,605 persons self-reported a disqualifying drug offense, but more
than 278,000 left the question blank. Those who left the question blank in
2000-2001 were deemed eligible to receive federal assistance.

Appendix III: Denial of Federal Postsecondary Education Benefits to Drug
Offenders

Estimates of the Numbers and Amounts of Federal Assistance Affected by
Controlled Substances

In the academic years from 2001-2002 through 2003-2004, we estimated that
between 17,000 and 23,000 students were denied Pell Grants because of
their drug convictions, and that the total estimated amount of Pell Grants
that these students would have received ranged from $41 million to $54
million. See table 14.

Table 14: Estimated Number of Persons Denied Federal Education Assistance
because of Drug Convictions and Estimated Amount of Federal Assistance
Affected

Convictions Academic year

2001-2002 2002-2003 2003-2004

                Number of FAFSA ineligibles 58,929 42,537 41,061

Estimated number and amounts of Pell Grants denied

Estimated number of students denied Pell 23,000 17,000 18,000

Grants because of drug convictions, but who otherwise would have received
them

  Estimated amount of Pell Grant denied to $54 million $41 million $45 million
                                    students

Estimated number and amounts of student loans denied

Estimated number of students denied 41,000 30,000 29,000

student loans because of drug convictions, but who otherwise would have
received them

Estimated amount of student loans
denied to studentsa $139 million $101 million $100 million
Lower-bound estimate
Upper-bound estimate $164 million $121 million $120 million

Source: GAO analysis of Department of Education data.

aLower-bound estimates were calculated using subsidized loan amounts, and
upper-bound estimates were calculated using unsubsidized loan amounts.

We provide annual estimates of the numbers affected because the period of
benefit ineligibility can vary, and a student denied benefits in one year
may become eligible to receive benefits in a subsequent year. Thus, the
estimates for one year do not necessarily affect the estimates for another
year.

In academic year 2001-2002, there were 58,929 FAFSA ineligibles. During
that same year 51.5 percent of FAFSA applicants were eligible to receive
Pell Grants, and 76.9 percent of those who were eligible received Pell
Grants (as shown in table 13). Multiplying the 58,929 by the 51.5 percent
and then multiplying this result by the 76.9 percent results in the
estimate of 23,000 individuals denied Pell Grants who otherwise would have
received them. To obtain the amount of Pell Grant lost to these students

Appendix III: Denial of Federal Postsecondary Education Benefits to Drug
Offenders

during academic year 2001-2002, we multiplied our estimated number of
students denied Pell Grants (23,000) by the average amount of a Pell Grant
in academic year 2001-2002 ($2,298).

Table 14 also shows that between academic year 2001-2002 and academic year
2003-2004, an estimated 29,000 to 41,000 students per year would have
received student loans if not for their drug convictions. The estimated
total amount of student loans forgone by these students ranged between
$100 million and $164 million per year.

Estimates of the Impacts of the President's Proposed Changes to the Rules
for Denying Drug Offenders' Federal Postsecondary Education Assistance

The President's fiscal year 2005 budget contained a proposal that would
have changed the administration of the Higher Education Act provision
relating to eligibility for federal higher education benefits. Federal law
disqualifies students who have been convicted of controlled substance
offenses, in accordance with the period of ineligibility in table 12, from
receiving federal higher education assistance.7 As currently implemented
by the Department of Education, disqualifying convictions are those drug
convictions on a student's record at the time the student's eligibility is
being determined, using the rules on the FAFSA worksheet.8 Under the
President's proposal-which was supported by the Office of National Drug
Control and Prevention-students would be ineligible for federal higher
education assistance only if they committed a disqualifying drug-related
offense while they are enrolled in higher education. This proposed change
would make eligible all students whose controlled substance convictions
occurred prior to enrolling in higher education.

Because of data limitations, we are unable to provide reliable estimates
of the impacts of the proposed changes contained in the President's fiscal
year 2005 budget proposal. However, we expect that the proposal would
lower our estimates for the numbers of students denied benefits because
some individuals would regain their eligibility for benefits, and
relatively few students enrolled in postsecondary education institutions
would be expected to both use drugs and be convicted of drug crimes.

720 U.S.C. S: 1091(r).
864 Fed. Reg. 38,504, 38,505 - 38,506 (1999).

Appendix III: Denial of Federal Postsecondary Education Benefits to Drug
Offenders

Limited Data on the Long-Term Impacts of Denying Federal Postsecondary
Education Assistance

While studies consistently show that the economic returns to higher
education are positive, we cannot establish a direct link between the
denial of federal postsecondary aid to students and a reduction in the
amount of postsecondary education completed by those who were denied aid.
Moreover, officials at ONDCP also suggested that the provisions of the
Higher Education Act that provide for denying educational aid to drug
offenders might contribute a deterrent effect on drug use. Similarly, we
were unable to identify studies that assess whether provisions of the HEA
actually helped to deter drug use. Additionally, we are unable to address
the question of whether these provisions of the HEA that deny higher
education benefits to drug offenders result in net positive or negative
effects on society because we were unable to find research that
conclusively indicates whether these provisions of the HEA led individuals
to forgo postsecondary education or deterred individuals from engaging in
drug use and drug-related criminal activities.

Additional formal education-e.g., completing high school or attending or
completing postsecondary education-has been demonstrated to increase
annual and lifetime earnings. In its review of the returns to education,
the U.S. Census Bureau concluded that increases in formal education had a
positive impact on annual earnings. For example, the U.S. Census Bureau
reported that for full-time workers between the ages of 25 and 64 between
1997 and 1999 the average annual income for those who have not completed
high school is $23,400, for high school graduates it is $30,400, and for
those completing a bachelor's degree it is $52,200. Average annual income
rises higher yet for those who obtain advanced degrees.

This general pattern, that increases in formal education correlate with
increases in annual earnings, also holds true across an individual's
lifetime. The U.S. Census Bureau reported that the average lifetime
earnings, based upon 1997-1999 work experience, for those who have not
completed high school is approximately $1 million, for high school
graduates it is $1.2 million, and for those completing a bachelor's degree
it is $2.1 million. Again, the average lifetime earnings rise higher yet
for those who obtain advanced degrees. Hence, college graduates can
expect, on average, to earn nearly twice as much over a lifetime as those
persons who have only a high school diploma and more than twice as much as
those who have not completed high school.

Similarly, a study published by the congressional Joint Economic Committee
in January 2000 concluded that there is a strong consensus among
economists that formal education has a positive impact not only on
personal income but also on society. The study concluded that among the

Appendix III: Denial of Federal Postsecondary Education Benefits to Drug
Offenders

positive societal economic returns from increases in formal education are
the creation of new knowledge (translating into the development of new
processes and technologies) and the diffusion and transmission of
knowledge (translating into the expansion of innovative techniques such as
those found in the high-technology sector). Positive societal noneconomic
improvements are also associated with increased amounts of formal
education, which help Americans become better mothers, fathers, children,
voters, and citizens. These positive noneconomic improvements are
sometimes called positive neighborhood effects. Some of the positive
neighborhood effects may be (1) more informed and interested voters, (2)
decreases in crime, (3) decreased dependence upon certain types of public
assistance, and (4) decreased incidence of illegitimate pregnancies.

Although the census study and the study conducted by the Joint Economic
Committee show positive economic and societal impacts of increased levels
of education, the total net impacts of these benefits are difficult to
quantify.9 Moreover, these studies do not comment on whether the loss of
federal education assistance (as occurs for drug offenders through the
provisions of the HEA) contributes to individuals' not completing
postsecondary education, or whether those individuals who are denied
federal education assistance generate the necessary funding to attend
institutions of higher education in other ways.

Also at issue is whether the provisions of the HEA that deny postsecondary
education benefits to drug offenders contribute positively to society by
providing a deterrent to drug use. Research on the costs to society from
drug use, and drug-associated criminal involvement, demonstrated that
these costs to society are high. Therefore, if the denial of federal
higher education benefits deters people from engaging in drug crimes, then
the provisions might have positive economic and noneconomic impacts on
society. Some of the positive affects of deterrence may include reductions
in drug-related health care costs, reductions in drug-related crime and
associated criminal justice costs, and increased national economic
productivity. In addition, for many offenders and in particular for
first-time drug offenders, the denial of postsecondary education benefits
may delay entry into postsecondary education rather than prevent it.

9Joint Economic Committee, United States Congress, Investment in
Education: Private and Public Returns (Washington, D.C.: January 2000).

    Appendix III: Denial of Federal Postsecondary Education Benefits to Drug
                                   Offenders

Limited Data on Racial Minorities and the Denial of Federal Postsecondary
Education Benefits

With the available data, we were unable to determine whether the
provisions of the Higher Education Act that provide for denial of
postsecondary education benefits would affect relatively larger or smaller
numbers of minorities. The FAFSA does not request information about
applicants' race; therefore ED does not have data on the racial
distribution of applicants or FAFSA ineligibles. Without data on the race
of applicants for federal student aid, it is not possible to determine
whether minorities are denied aid at higher rates than whites.

The Bureau of Justice Statistics' Survey of Adults on Probation in 1995,
which is the only national survey of probationers that includes data on
the type of offense of conviction and educational attainment, indicates
that there may be racial differences in the levels of educational
attainment of drug offenders. The survey indicates that black and Hispanic
drug offenders on probation complete high school at a lower rate than
white drug offenders on probation. Specifically, while 68 percent of white
drug offenders on probation had completed high school, 51 percent of black
and 46 percent of Hispanic drug offenders on probation had completed high
school. As completing high school (or gaining a general equivalency
degree) is a prerequisite for enrollment in postsecondary education, these
data suggest that lower proportions of black and Hispanic drug offenders
(at least drug offenders on probation) would be eligible to enroll in
postsecondary educational institutions and would therefore be eligible for
federal higher education assistance.

Appendix IV: Denial of Federally Assisted Housing Benefits for Drug-Related
Criminal Activity

Legal Framework Relating to the Denial of Housing Benefits for
Drug-Related Criminal Activity

This appendix provides background on the legal and administrative
frameworks for denying federally assisted housing benefits to persons who
engage in drug-related criminal activities, our methods for estimating the
numbers of persons denied benefits, and how we assessed the available data
on racial minorities and the limited information on potential impacts.

Federal law contains a variety of provisions relating to the denial of
federally assisted housing benefits for certain types of drug-related
criminal activity. These provisions relate to, among other things, (1) who
may lose eligibility for federally assisted housing benefits because of
drugrelated criminal activity and (2) screening tools for the providers of
federally assisted housing to use to determine ineligibility for such
housing benefits. Motivation for prohibiting drug offenders from public
housing is reflected, in part, in congressional findings made in 1990 and
amended in 1998, about drug-related criminal activities in public housing;
these findings stated, in part, that (1) "drug dealers are increasingly
imposing a reign of terror on public and other federally assisted
low-income housing tenants," (2) "the increase in drug-related crime not
only leads to murders, muggings, and other forms of violence against
tenants, but also to a deterioration of the physical environment," and (3)
"the Federal government has a duty to provide public and other federally
assisted lowincome housing that is decent, safe, and free from illegal
drugs."1

Public housing agencies, which are typically local agencies created under
state law that, under Department of Housing and Urban Development
guidance, manage and develop public housing units for low-income families,
are required, for example, to utilize leases that provide that any
drug-related criminal activity on or off the premises by a public housing
tenant shall be cause for termination of the tenancy.2 This provision also
specifically applies to drug-related criminal activity by any member of
the tenant's household or any guest or other person under the tenant's
control. Similarly, federal law requires PHAs and owners of federally
assisted housing to establish standards or lease provisions that allow for
the

142 U.S.C. S:11901.

2The term "drug-related criminal activity" means the illegal manufacture,
sale, distribution, use, or possession with intent to manufacture, sell,
distribute, or use a controlled substance-as defined in section 102 of the
Controlled Substances Act (21 U.S.C. S: 802). 42 U.S.C. S: 1437d(l). With
respect to termination of tenancy, see 42 U.S.C. S: 1437d(l)(6).

Appendix IV: Denial of Federally Assisted Housing Benefits for
Drug-Related Criminal Activity

termination of the tenancy or assistance for any household with a member
who the PHA or owner determines is illegally using a controlled
substance.3 Federal law further specifies that tenants evicted from
federally assisted housing by reason of drug-related criminal activity are
to be ineligible for federally assisted housing for a 3-year period,
although evicted tenants that successfully complete an approved
rehabilitation program may regain their eligibility before the 3-year
period ends.

Under federal law and implementing regulations, PHAs have the discretion
to evict tenants for drug-related criminal activity but are not required
to evict such tenants. Rather, they are required to use leases that
provide that any drug-related criminal activity on or off the premises by
a public housing tenant shall be cause for termination of the tenancy.
Implementing regulations by the U.S. Department of Housing and Urban
Development relating to termination provide that a determination of such
criminal activity may be made regardless of whether a person has been
arrested or convicted of such activity and without satisfying a criminal
conviction standard of proof of the activity.4

With respect to methamphetamine convictions, PHAs are required under
federal law to establish standards to immediately and permanently
terminate a tenancy as well as permanently prohibit occupancy in public
housing for persons convicted of certain methamphetamine offenses
occurring on public housing premises. PHAs do not have discretion in
evicting these persons, and the standards also require that Housing Choice
Voucher Program (formerly Section 8 low-income housing) participation be
denied to such persons.

Federal law also provides various screening tools to assist with
determining possible ineligibility of tenants and applicants for federally
assisted housing benefits because of drug-related criminal activity. These
tools come primarily in the form of access to certain types of
information. For example, under federal law, housing assistance agencies
are authorized to request access to criminal conviction records from
police departments and other law enforcement agencies for the purposes of

342 U.S.C. S: 13662(a). This provision further provides for the
termination of tenancy or assistance to a household with a member whose
illegal use (or pattern of illegal use) of a controlled substance, among
other things, is determined by the PHA or owner to interfere with the
health, safety, or right to peaceful enjoyment of the premises by other
residents.

424 C.F.R. S: 5.861.

  Appendix IV: Denial of Federally Assisted Housing Benefits for Drug-Related
                               Criminal Activity

HUD and PHA Roles in Denying Federally Assisted Housing Benefits for
Reasons of Drug-Related Criminal Activities

applicant screening, lease enforcement, and eviction. PHAs have the
authority under certain conditions to request access to such information
with respect to tenants and applicants for the Housing Choice Voucher
Program. Public housing authorities are also authorized under federal law
to require that applicants provide written consent for the public housing
authorities to obtain certain types of records, such as criminal
conviction records and drug abuse treatment facility records.5

HUD is responsible for establishing the rules and providing guidance to
PHAs in their administration of federally assisted housing benefits. PHAs
can manage a single program or multiple HUD programs. HUD's Office of
Public and Indian Housing oversees the two key rental housing assistance
programs that we reviewed, namely the Low-Rent Public Housing Assistance
Program (also referred to as low-rent, or public housing) and the HCV
Program.

During the 1990s, PHAs gained broader latitude from HUD and Congress to
establish their own policies in areas such as selecting tenants.6 This
included increased latitude in taking actions to deny federally assisted
housing benefits to persons receiving housing benefits and to applicants
for benefits. HUD requires PHAs to submit for its review and approval
annual plans that include, among other things, their policies for
continuing occupancy and denying admission for drug-related criminal
activities.

Recent HUD guidance regarding denying federal housing benefits to persons
engaged in drug-related criminal activities was issued in its "Final
Rule," dated May 2001.7 The rule amended existing regulations regarding
implementing the federally assisted housing tenant eviction and applicant
screening provisions for drug-related criminal activities.

5See 42 U.S.C. S: 1437d, subsections (q), (s), and (t) for provisions
related to PHA authority to obtain criminal history information.

6GAO, Public Housing: Small and Larger Agencies Have Similar Views on Many
Recent Housing Reforms, GAO-04-19 (Washington, D.C.: October 30, 2003).

766 Fed. Reg. 28,776 (2001) (codified at 24 C.F.R. Part 5, among others).

Appendix IV: Denial of Federally Assisted Housing Benefits for
Drug-Related Criminal Activity

PHAs' Policies to Address Drug-Related Criminal Activities Vary

Termination and admission policies can vary substantially among PHAs
nationwide. In a baseline study (November 2000) of a stratified random
sample of the PHAs that were responsible for managing federally assisted
housing units in the HCV Program, HUD reviewed the discretionary authority
among PHAs.8 HUD reported that the variation among PHAs in conducting
criminal background checks could legitimately result in an applicant being
barred by one PHA even though the applicant could otherwise be admitted by
another PHA. Some of the variations reported in the study include
differences in (1) the sources used to obtain information about criminal
history and drug-related criminal activities (e.g., newspaper stories,
resident complaints, self-disclosure, official law enforcement
records-federal, state, local); (2) the costs (paid by the PHA) associated
with obtaining official law enforcement criminal background records; (3)
the time span covered by the criminal history search; and (4) whether
consideration is given to repeat offenses, only convictions, or arrests
and convictions.

We obtained and reviewed policies from seven of the largest PHAs having
combined programs-Public Housing and HCV.9 Our review of their policies
with respect to terminations and admissions for drug-related criminal
activities showed variations in the policies established to deny housing
benefits. For example, policies regarding terminations of leases (for
public housing tenants) or termination of assistance (for HCV recipients)
vary in how they implement the drug-related criminal activity

8See Devine, Deborah J., Barbara A. Haley, Lester Rubin, and Robert W.
Gray, The Uses of Discretionary Authority in the Tenant-Based Section 8
Program: A Baseline Inventory of Issues, Policy, and Practice, Washington,
D.C.: U.S. Department of Housing and Urban Development, November 2000. HUD
used a stratified random sample of 167 PHAs out of 1,786 that managed 100
or more units of Section 8 housing. These 1,786 represented 70 percent of
the 2,553 PHAs, but they also represented 98 percent of all units managed
by PHAs in the Section 8 program. The sample was stratified on PHA size
(four categories) and seven regions.

9We mailed a request for information to 40 of the nation's largest PHAs in
which we requested data about denying federally assisted housing for
reasons of drug-related criminal activities. We report on these data in
subsequent sections. We also requested that the PHAs provide us with
copies of policies and procedures used during 2003 for terminating leases
or assistance as well as screening applicants for admissions. Seven PHAs
provided copies of their 2003 or 2004 policies; listed alphabetically,
these were the Cincinnati Metropolitan Housing Authority, Cuyahoga
Metropolitan Housing Authority, Housing Authority of the City of Los
Angeles, Housing Authority of the County of San Diego, Jacksonville
Housing Authority, Memphis Housing Authority, and the Metropolitan
Development and Housing Agency of Nashville.

Appendix IV: Denial of Federally Assisted Housing Benefits for
Drug-Related Criminal Activity

provisions and in the scope of criminal background that can result in
terminations:

o  	Drug-related criminal activity provisions can range from certain types
of prohibited behaviors (e.g., those that threaten the health and safety
of other residents) to certain drug convictions (e.g., drug-related
criminal activity, and methamphetamine in particular).

o  	Scope of criminal background can vary by the period of prior criminal
history that can trigger termination of leases or assistance, the type of
prohibited drug-related criminal activities (e.g., personal use, felonious
distribution, etc.), or whether there was a conviction in the case.

Analogously, PHA policies on admissions into public housing or into HCV
can vary based on a number of factors, and these variations in policies
can result in differences among PHAs in the types of drug offenders that
are denied federally assisted housing.

o  	Applicant screenings for drug-related criminal activity can occur in
varying forms (such as an application, interview, or eligibility
verification) and at varying times-such as before or after placement on
the PHAs' waiting lists.

o  	Sources of criminal history information used can vary, so some PHAs
cast a wider net than others when searching for prohibited drug-related
criminal activity. Sources can range from using only local law enforcement
records to using Federal Bureau of Investigation/National Crime
Information Center data.

o  	Periods of ineligibility for prior evictions from federally assisted
housing can vary by time frame and criminal activity (e.g., drug-related
or violent). Ineligibility periods ranged from 3 to 5 years.

o  	The evidence standard for drug addiction can vary to include a
reasonable cause to believe illegal drug use exists or self-disclosure of
illegal use on the application itself.

We obtained data from HUD on the number of evictions from and applicants
denied admission into public housing during fiscal years 2002 and 2003 for
reasons of criminal activities. In each year, more than 98 percent of the
PHAs that manage public housing responded to HUD's request for data about
security within the units managed by the PHA, including information on
evictions and applicants denied admission. HUD's information pertains to
persons evicted or denied admission for reasons of criminal activity;
these data do not distinguish between criminal activity and drug-related
criminal activities. The HUD data also do not include measures of the
number of tenants or of the total number of applicants screened. To adjust
for differences in the size of the PHAs, we

Number of Denials of Federally Assisted Housing Benefits

Appendix IV: Denial of Federally Assisted Housing Benefits for
Drug-Related Criminal Activity

calculated a rate at which applicants were evicted or denied admission
into public housing because of criminal activities that was based on the
number of units maintained by all reporting PHAs. These data are reported
in table 15.

Table 15: Evictions from and Denials of Admission into Public Housing for
Reasons of Criminal Activity, Fiscal Years 2002 and 2003

Fiscal year 2002 2003

                   Number of PHAs reporting data 3,103 3,113

           Number of public housing units managed 1,164,089 1,156,735

Number of evictions from public housing for criminal activity 9,083 9,325 Number
of applications denied admission into public housing 49,207 48,853 for criminal
                                    activity

Evictions and denials as a percent of units managed by PHAs

Evictions as a percentage of unitsa 0.8% 0.8%

                                       a

Denials as a percentage of units 4.2% 4.2%

Source: GAO analysis of HUD Management Operations Certification Assessment
System (MASS) data.

aData on the total number of tenants or applicants were not reported in
the MASS data that we obtained.

During each of the fiscal years, 2002 and 2003, there were more than 9,000
evictions (amounting to less than 1 percent of all units managed) because
of criminal activities. There were about 49,000 applications for admission
into public housing that were denied for reasons of criminal activities
(amounting to about 4 percent of all units). As drug-related criminal
activities are a subset of criminal activities, these data suggest that
even if all of those evicted from public housing for reasons of criminal
activity had engaged in drug-related criminal activities, terminations
leading to evictions would amount to less than 1 percent of the public
housing units managed by PHAs.

  Appendix IV: Denial of Federally Assisted Housing Benefits for Drug-Related
                               Criminal Activity

Selected PHAs Mostly Report That Relatively Small Percentages of Tenants
Leases Are Terminated for Reasons of Criminal Activity or Drug-Related
Criminal Activities

To gauge the extent to which PHAs denied federally assisted housing by
terminating leases (leading to possible evictions) for drug-related
criminal activities, we contacted 40 of the largest PHAs in the country
and asked them to provide data on the number of actions10 that they took
to evict tenants by terminating leases and of these, the numbers that were
terminated for criminal activity and for drug-related criminal activity.11
Of the 40 PHAs that we contacted, we received data from 17. We assessed
the data that these PHAs provided for reliability.

As shown in table 16, 15 of 17 PHAs that responded to our request provided
data on the total number of public housing termination of leases. The rate
at which PHAs terminated leases for reasons of drug-related criminal
activities varied considerably, from 0 percent in Santa Clara County to
39.3 percent in Memphis. The Philadelphia PHA, which reported the largest
number of lease terminations (2,324), reported terminating 50 of these
leases (or 2.2 percent) for reasons of drug-related criminal activities.
The Santa Clara County PHA terminated the smallest number of leases (1).
Combined, the 13 PHAs that reported both the number of lease terminations
and the number of terminations for drug-related criminal activities,
reported ending a total of 9,249 leases, and 520 of the terminations (or
5.6 percent of the total) were for drug-related criminal activities.

10Actions taken by PHAs reflect decisions to terminate leases and
assistance, as discussed in this section, as well as denying admissions,
as discussed in a subsequent section of this appendix.

11HUD collects data on lease terminations and denial of admission into
public housing through its Management Operations Certification Assessment
System (MASS) data collection. This system reports the number of such
actions taken for reasons of criminal activities, but it does not report
data on the number of actions taken explicitly for reasons of drug-related
criminal activities.

  Appendix IV: Denial of Federally Assisted Housing Benefits for Drug-Related
                               Criminal Activity

Table 16: Public Housing Lease Terminations at Selected Public Housing
Agencies, by Reason for Termination (Criminal Activities and Drug-Related
Criminal Activities), 2003

Public housing leases terminated for reasons of criminal activities Public
housing leases terminated for reasons of drug-related criminal activities

       Percentage of all terminations of leases Percentage of terminations of
               leases for reasons of criminal activities Public housing lease
                                            terminations (leading to possible

eviction) Number

Percentage of all terminations of leases Number

Public housing agency

       Atlanta (GA)         320      78      24.4%       38  11.9%      48.7% 
      Cincinnati (OH)       721      21      2.9%        13   1.8%      61.9% 
      Cuyahoga County                                                
           (OH)            1,654     102     6.2%        84   5.1%      82.4% 
     Indianapolis (IN)      207      36      17.4%        a        a        a 
     Jacksonville (FL)     1,299     70      5.4%        47   3.6%      67.1% 
     Los Angeles (CA,                                                
         City of)           373      109     29.2%       51  13.7%      46.8% 
     Los Angeles (CA,                                                
        County of)          170      51      30.0%       26  15.3%      51.0% 
       Memphis (TN)         28       17      60.7%       11  39.3%      64.7% 
      Nashville (TN)        260      148     56.9%       90  34.6%      60.8% 
     New Orleans (LA)       233      133     57.1%       83  35.6%      62.4% 
        Newark (NJ)        1,720     27      1.6%        20   1.2%      74.1% 
       Oakland (CA)         146      35      24.0%        7   4.8%      20.0% 
     Philadelphia (PA)     2,324     133     5.7%        50   2.2%      37.6% 
       Portland (OR)        65        a        a          a        a        a 
      San Bernardino                                                 
        County (CA)          a        8        a          a        a        a 
     San Diego County                                                
           (CA)              a        a        a          a        a        a 

Santa Clara

b

County (CA) 1 0 0.0% 0 0.0% Source: GAO analysis of data from selected
public housing agencies. aData were missing or not reported. bNot
applicable.

Appendix IV: Denial of Federally Assisted Housing Benefits for
Drug-Related Criminal Activity

Further, although the data on lease terminations for reasons of
drugrelated criminal activities are not generalizable to all PHAs that
manage public housing program units, the information that they provided on
leases terminated for reasons of drug-related criminal activities, and our
calculation of these numbers as a percentage of terminations for criminal
activities, show wide variation in the extent to which drug-related
criminal activities predominate among all criminal activities that can
result in a termination of a lease. In Cuyahoga County, for example, 82.4
percent of lease terminations for criminal activity were terminations for
drug-related criminal activities, but in Oakland, 20 percent of the
terminations for criminal activity occurred as a result of drug-related
criminal activities. A majority of the PHAs that reported these data also
reported that the number of lease terminations and the reasons for them
(i.e., criminal or drug-related criminal activities) were similar to or
smaller than the numbers in the prior 3 years.

As shown in table 17, 16 of the 17 PHA respondents were able to provide
some (although often incomplete) data on actions taken to terminate HCV
assistance during 2003. Nine of the 16 PHA respondents were able to
provide data on the number of actions to terminate HCV assistance for
reasons of drug-related criminal activity or criminal activity. However,
only 5 of the 9 respondents were able to provide data on the number of
actions specifically taken to terminate HCV assistance for reasons of
drugrelated criminal activity. These 5 PHAs took 9,537 actions related to
terminating HCV assistance, of which 54 (or about 0.6 percent) of such
actions were for terminating assistance for reasons of drug-related
criminal activities.

Four of the 9 PHA respondents were able to provide data on the number of
actions taken to terminate HCV assistance for reasons of criminal
activity, and most, but not all, of them provided (at our request) broad
estimates for drug-related criminal activity based on the total number of
actions to terminate HCV assistance during 2003. These 4 PHAs took a total
of 3,166 actions related to terminating HCV assistance, of which 133
actions (or about 4.2 percent) were for terminating assistance because of
criminal activities. Three of the 4 PHAs estimated less than 25 percent of
their total actions could have been for reasons of drug-related criminal
activities, and 1 PHA did not provide an estimate. Applying the
upper-bound broad estimates (25 percent) to each PHA's total actions would
be overstating terminations for reasons of drug-related criminal activity
because the resulting number is most likely to be equal to or be a subset
of terminations for reasons of criminal activities. From a conservative
perspective, it is conceivable that the 133 actions also represent

Appendix IV: Denial of Federally Assisted Housing Benefits for
Drug-Related Criminal Activity

terminations of assistance for reasons of drug-related criminal activity,
thereby establishing a maximum rate of denial at 4.2 percent for reasons
of drug-related criminal activity.

Seven of the 16 PHA respondents provided only the total number of actions
taken to terminate HCV assistance, along with a broad estimate of the
percentages of terminations that could have been for reasons of
drugrelated criminal activity. Six of the 7 PHAs reported less than 25
percent of their total actions could have been for reasons of drug-related
criminal activities, and one reported 51 to 75 percent. Table 17 provides
the data on terminations of assistance from the HCV program.

Table 17: Housing Choice Voucher Program Terminations of Assistance at
Selected Public Housing Agencies, by Reason for Termination (Criminal
Activities and Drug-Related Criminal Activities) of Assistance, 2003

              HCV terminations for reasons of criminal activities

HCV terminations for reasons of drugrelated criminal activities

     Percentage of all terminations of assistance Estimated percentage of all
                                                   terminations of assistance

HCV terminations of assistance Number

Percentage of all terminations of assistance Number

Public housing agency

a a

                         Atlanta (GA) 855 11 1.3% < 25%

b

Cincinnati (OH) 98 14 14.3% 6 6.1%

Cuyahoga

b

County (OH) 1,214 14 1.2% 6 0.5%

    Indianapolis (IN)     247      a  25% to 50% c     a       a        < 25% 
    Jacksonville (FL)    5,770    26      0.5%         4   0.07%            b 
       Los Angeles       4,468     a    < 25% c        a       a        < 25% 
      (CA, City of)                                              
       Los Angeles       1,080     a    < 25% c        a       a        < 25% 
     (CA, County of)                                             
       Memphis (TN)        3       a    < 25% c        a       a   51% to 75% 
      Nashville (TN)      720      a    < 25% c        a       a        < 25% 
       New Orleans       2,241     a    < 25% c        a       a        < 25% 
           (LA)                                                  
       Newark (NJ)         8       a    < 25% c        a       a        < 25% 
       Oakland (CA)        a       a  51% to 75% c     a       a        > 75% 
    Philadelphia (PA)    1,445    54      3.7%        11    0.8%            b 
      Portland (OR)       380     71     18.7%         a       a            a 

San Bernardino

a a

                        County (CA) 1,009 47 4.7% < 25%

Appendix IV: Denial of Federally Assisted Housing Benefits for
Drug-Related Criminal Activity

HCV terminations for reasons of criminal activities

HCV terminations for reasons of drugrelated criminal activities

     Percentage of all terminations of assistance Estimated percentage of all
                                                   terminations of assistance

HCV terminations of assistance Number

Percentage of all terminations of assistance Number

Public housing agency

San Diego

b

County (CA) 1,010 42 4.2% 27 2.7%

Santa Clara

a a

                          County (CA) 922 4 0.4% < 25%

Source: GAO analysis of data from selected public housing agencies.

aData were missing or not reported.

bData reported in previous column.

cIn response to GAO's request, the PHA estimated the percentage (within
certain ranges) of all action initiated to terminate HCV assistance that
could have been taken for reasons of criminal activity.

The majority of PHAs that reported data on terminations from the HCV
program also reported that the number and types of actions that they took
during 2003 were similar to the numbers in the prior 3 years.

Selected PHAs Report That Relatively Small Percentages of Applicants Are
Denied Admission for Reasons of Criminal Activity or Drug-Related Criminal
Activities

As shown in table 18, 15 of 17 PHAs that responded to our request provided
data on the number of actions taken on applications for public housing.
However, only 6 of the 15 respondents provided data on the number of
actions specifically taken to deny admission into public housing for
reasons of drug-related criminal activity. Collectively, these six PHAs
took action on 11,538 applications, of which 330 (or about 2.9 percent of)
such actions were for denying admissions for reasons of drug-related
criminal activities.

Nine of the 15 PHAs did not provide counts of the number of denials for
reasons of drug-related criminal activity but provided data on the number
of actions taken to deny admission for reasons of criminal activity. In
completing our request, 4 PHAs provided broad estimates of denials for
drug-related criminal activity based on the total number of actions taken
on applications for public housing, and 5 PHAs did not provide estimates.
Collectively, these 9 PHAs reported a total of 17,921 actions related to
applications for admission into public housing, of which 1,081 actions (or
6 percent) were for denying admission for reasons of criminal activities.
On the basis of our assumption that admission denials for reasons of
drugrelated criminal activity are most likely to be either a subset of or

  Appendix IV: Denial of Federally Assisted Housing Benefits for Drug-Related
                               Criminal Activity

equivalent to the admission denials for reasons of criminal activities, we
estimate that the maximum rate of denial for reasons of drug-related
criminal activity for these PHAs is 6 percent.

Table 18: Selected Public Housing Agencies' Actions Reported on
Applications for Admission and Denials of Admission into Public Housing
for Reasons of Criminal Activities and Drug-Related Criminal Activities,
2003

Public housing applications denied because of criminal activities

 Public housing applications denied because of drug-related criminal activities

     Percentage of all actions taken on applications for assistance Estimated
     percentage of all actions taken on applications for assistance Number of
                                            actions taken on applications for

public housing Number

                          Percentage of all actions taken on applications for

assistance Number

Public housing agency

a a a

Atlanta (GA) 608 56 9.2%

b

Cincinnati (OH) 5,785 359 6.2% 24 0.4%

Cuyahoga

a a

                       County (OH) 2,307 282 12.2% < 25%

a a a

Indianapolis (IN) 506 75 14.8%

a a a

Jacksonville (FL) 970 22 2.3%

Los Angeles

a a

                       (CA, City of) 3,920 295 7.5% < 25%

Los Angeles

a a

                      (CA, County of) 1,217 44 3.6% > 75%

b

Memphis (TN) 420 157 37.4% 4 1.0%

b

Nashville (TN) 2,093 137 6.5% 127 6.1%

New Orleans

b

(LA) 365 12 3.3% 12 3.3%

       Newark (NJ)       2,359    600    25.4%    163      6.9%             b 
       Oakland (CA)       674      59     8.8%        a       a    51% to 75% 
    Philadelphia (PA)    5,400    226     4.2%        a       a             a 
      Portland (OR)      2,319     22     0.9%        a       a             a 
      San Bernardino                                            
       County (CA)         a      147      a          a       a         < 25% 

San Diego

a a a a a a

County (CA) Santa Clara

b

County (CA) 516 0 0.0% 0 0.0%

Source: GAO analysis of data from selected public housing agencies.

aData were missing or not reported. bData reported in previous column.

Appendix IV: Denial of Federally Assisted Housing Benefits for
Drug-Related Criminal Activity

As with the other outcomes, the PHAs varied in the extent to which they
reported that applicants were denied admission into public housing for
reasons of drug-related criminal activities, and the majority of PHAs that
provided data for 2003 reported that their activities related to actions
and denials in 2003 were similar to the numbers in the prior 3 years.

As shown in table 19, 14 of the 17 PHA respondents provided some (although
mostly incomplete) data on actions taken on applications for the HCV
program. Nine of the 14 PHAs provided data on the number of denials of
admission into the program for reasons of drug-related criminal activities
or criminal activity. Of the 2 PHAs that provided data on the number of
denials for reasons of drug-related criminal activity, 1 PHA reported no
denials, and the other PHA reported 10 denials out of 1,483 actions taken
on applications, or 0.7 percent. Seven PHAs provided data on the number of
denials for reasons of criminal activity. Among these 7 PHAs, there were a
total of 20,513 reported actions taken on applicants. Of these, 303 were
denied admission for reasons of criminal activities (or about 1.5
percent). On the basis of our assumption that admission denials for
reasons of drug-related criminal activity are most likely to be either a
subset of or equivalent to the admission denials for reasons of criminal
activities, we estimate that the maximum for admission denials for reasons
of drug-related criminal activity is 1.5 percent. We could not provide
reliable estimates for the remaining 5 PHAs that reported incomplete data.

  Appendix IV: Denial of Federally Assisted Housing Benefits for Drug-Related
                               Criminal Activity

Table 19: Housing Choice Voucher Program, Selected Public Housing
Agencies' Actions Reported on Applications for Assistance and Denials of
Assistance for Reasons of Criminal Activities and Drug-Related Criminal
Activities, 2003

          HCV denials of assistance for reasons of criminal activities

HCV denials of assistance for reasons of drug-related criminal activities

           Percentage of all actions taken on HCV applications for assistance
            Estimated percentage of all actions taken on HCV applications for
                                                                   assistance

Public housing agency

                              Number of actions taken on HCV applications for

assistance Number

              Percentage of all actions taken on HCV applications

                             for assistance Number

         Atlanta (GA)         3,886    57    1.5%       a      a        < 25% 
       Cincinnati (OH)        3,138    79    2.5%       a      a   25% to 50% 
     Cuyahoga County (OH)       a      49      a        a      a            a 
      Indianapolis (IN)        390     12    3.1%       a      a   25% to 50% 
      Jacksonville (FL)       4,191    91    2.2%       a      a            a 
    Los Angeles (CA, City    53,426    a    < 25% d     a      a        < 25% 
             of)                                                 
       Los Angeles (CA,       1,711    a    < 25% d     a      a        < 25% 
          County of)                                             
         Memphis (TN)           0      0       c        0      c 
        Nashville (TN)         990     a    < 25% d     a      a        < 25% 
       New Orleans (LA)       1,321    8     0.6%       a      a        < 25% 
         Newark (NJ)           472     a    < 25% d     a      a        < 25% 
         Oakland (CA)           a      a       a        a      a            a 
      Philadelphia (PA)       1,483    22    1.5%      10   0.7%            b 
        Portland (OR)         1,387    53    3.8%       a      a            a 
    San Bernardino County                                        
             (CA)               a      a       a        a      a            a 
    San Diego County (CA)     4,940    a       a        a      a            a 

Santa Clara County

a a

(CA) 6,200 3 0.05% < 25%

Source: GAO analysis of data from selected public housing agencies.

aData were missing or not reported.

bData reported in previous column.

cNot applicable.

dIn response to GAO's request, the PHA estimated the percentage (within
certain ranges) of all action taken on HCV applications for assistance
that could have been for reasons of criminal activity.

Appendix IV: Denial of Federally Assisted Housing Benefits for
Drug-Related Criminal Activity

Several Factors Can Contribute to the Reported Low Number of Denials
because of Drug-Related Criminal Activities, but Potential Impacts Are
Unclear

Our review of limited data and interviews with those involved in federally
assisted housing suggest a number of factors can contribute to the
relatively low percentages of denials being reported for reasons of
drugrelated criminal activities. As noted in a HUD baseline study,
variation among PHAs in conducting HCV criminal background checks could
legitimately result in an applicant being barred by one PHA who would
otherwise be admitted by another PHA. In addition, a HUD official
suggested that the percentages of denials that were reported to us by
selected PHAs can be influenced by whether (1) the PHAs place drug users
at the bottom of their waiting lists, (2) PHAs differ in the treatment of
applicants if a household member rather than the applicant is the subject
of the drug-related criminal activity, and (3) local courts presiding over
eviction proceedings view the PHAs as the housing provider of last resort.
In the last instance, the PHA's decision to terminate a lease for reason
of drug-related criminal activity may not be upheld.

Moreover, comments made during interviews with selected officials on
matters related to housing were consistent with our analysis of HUD data
on PHA denials for criminal activity, and the relatively low number of
denials for drug-related criminal activity provided to us by selected
PHAs. Regarding the relatively small number of persons whose housing
benefits were reported as terminated or persons denied program
participation for reasons of criminal or drug-related criminal activities,
a representative from the National Association of Housing and
Redevelopment Officials stated that PHAs are not looking to turn away
minor offenders (e.g., "the type of people that may have only stolen a
candy bar") but rather hardened criminals. On the variation in denials of
federally assisted housing for drug-related criminal activities, the
Project Coordinator for the Re-Entry Policy Council at the Council of
State Governments suggested that the barriers to housing ex-drug offenders
revolve around the discretion afforded PHAs, and that these barriers can
best be dealt with at the local level by making states more aware of the
issue, the applicability of the local rules, and the need for building
collegial relationships with PHAs to develop options for housing
ex-felons.

More generally, assessing the impacts of the denial of federal housing
benefits on the housing communities or on the individuals and families
that have lost benefits was beyond the scope of our review, given the
limited data that are available. Officials at HUD reported that they have
not studied this issue, and our review of the literature did not return
any comprehensive studies of impacts. In our opinion, any full assessment
of the impacts of denial of housing benefits to drug offenders would have
to consider a wide range of possible impacts, such as improvements in
public

Appendix IV: Denial of Federally Assisted Housing Benefits for
Drug-Related Criminal Activity

Data on the Race of Persons Denied Federally Assisted Housing Were
Infrequently Provided

safety that result from terminating leases of drug offenders; displacement
of crime from one area to another with perhaps no overall (or area-wide)
improvements in crime reduction; as well as the impacts on individuals and
families, to name a few. Any such impact assessments would be complicated
by the market conditions (limited quantity and high demand) for federally
assisted housing and the variation in PHAs' policies and practices that
would also need to be considered.

We requested the PHAs to provide us with data on the race of persons who
were denied federally assisted housing for reasons of drug-related
criminal activities. Of the 17 PHAs that responded to our request, few
provided data by race on (1) the total number of actions taken and (2)
those actions that were specifically for drug-related criminal activity.
Only 4 PHAs provided data by race on public housing terminations, and 3
PHAs provided data by race on public housing admission denials. Four PHAs
provided data by race on HCV terminations of assistance. Only 1 PHA
provided data by race on HCV admission denials.

From these limited data, we were unable to develop reliable estimates of
racial differences in the frequency of terminations and denials of
admission into federally assisted housing. In some cases, the number
reported as terminated for drug-related criminal activities was too small
to provide stable estimates, and because of the small numbers, the
estimates of racial differences could exhibit large changes with the
addition of a few more cases. For example, only 4 PHAs provided data by
race on the number of leases terminated for reasons of drug-related
criminal activities. In 1 PHA, slightly more than 3 percent of all lease
terminations of blacks were for drug-related criminal activities, while
almost 6 percent of all lease terminations of whites were for drug-related
criminal activities. In this PHA, whites were about one and one half times
more likely than blacks to have their leases terminated for reasons of
drug-related criminal activities. In this PHA, 110 whites had leases
terminated during 2003, and 6 of these terminations were for drug-related
criminal activities. In a second PHA, blacks were three times as likely as
whites to have their leases terminated for reasons of drug-related
criminal activities, as 18 percent of blacks and 5 percent of whites had
leases terminated for reasons of drug related criminal activities. But in
this second PHA, 19 whites had leases terminated, and 1 of these was for
drug-related criminal activities. An addition of 2 whites to the number
that had leases terminated for drugrelated criminal activities would have
almost eliminated the racial difference.

Appendix V: Legal and Administrative Framework for the Denial of Federal
Benefits Program

About the Law 	The Denial of Federal Benefits Program established under
section 5301 of the Anti-Drug Abuse Act of 1988,1 in general, provides
federal and state court judges with a sentencing option to deny selected
federal benefits to individuals convicted of federal or state offenses for
the distribution or possession of controlled substances. The federal
benefits that can be denied include grants, contracts, loans, and
professional or commercial licenses provided by an agency of the United
States.2 Certain benefits are excluded from deniability under this
provision of the law; these include benefits such as social security,
federally assisted housing, welfare, veterans' benefits, and benefits for
which payments or services are required for eligibility.3 Federally
assisted housing, TANF, and food stamp benefits may be denied to drug
offenders under other provisions of federal law. (See app. II for more
information on the denial of TANF and food stamp benefits, and see app. IV
for more information on the denial of federally assisted housing
benefits.)

Federal and state court sentencing judges generally have discretion to
deny any of the deniable benefits for any length up to those prescribed by
the law, with the exception of the mandatory denial of benefits required
for a third drug trafficking conviction. More specifically, depending upon
the type of offense, and conviction and the number of prior convictions,
the law provides for different periods of ineligibility for which benefits
can or must be denied. As the number of convictions for a particular type
of drug offense increases, so does the period of ineligibility for which
benefits can or must be denied. Table 20 shows these periods.

1Pub. L. No. 100-690, 102 Stat. 4310 (1988) (codified at 21 U.S.C. S:
862).

2The term "Federal benefit" is defined at 21 U.S.C. S: 862(d)(1) to mean
"any grant, contract, loan, professional license, or commercial license
provided by an agency of the United States or by appropriated funds of the
United States."

3The term "Federal benefit," under 21 U.S.C. S: 862(d)(1), specifically
excludes "any retirement, welfare, Social Security, health, disability,
veterans benefit, public housing, or other similar benefit, or any other
benefit for which payments or services are required for eligibility."

Appendix V: Legal and Administrative Framework for the Denial of Federal
Benefits Program

Table 20: Periods of Ineligibility for Which Benefits Can or Must Be
Denied under Section 5301 of the Anti-Drug Abuse Act, as Amended

                           Type of conviction offense

              Number of convictions  Drug possession     Drug trafficking     
                   First conviction     Up to 1 year      Up to 5 years       
                  Second conviction    Up to 5 years      Up to 10 years      
                   Third conviction                a  Permanent ineligibility 

Source: GAO analysis of provisions of section 5301 of the Anti-Drug Abuse
Act of 1988, as amended.

aNo period of ineligibility is listed for drug possession offenses with
three or more convictions.

With respect to first-time drug possession convictions, a court may impose
certain conditions, such as the completion of an approved drug treatment
program, as a requirement for the reinstatement of benefits. In addition,
the sentencing court continues to have the discretion to impose other
penalties and conditions apart from section 5301 of the Anti-Drug Abuse
Act of 1988.

Section 5301 of the Anti-Drug Abuse Act of 1988, as amended, also provides
that under certain circumstances, the denial of benefits penalties may be
waived or suspended with respect to certain offenders. For example, the
denial of benefits penalties are not applicable to individuals who
cooperate with or testify for the government in the prosecution of a
federal or state offense or are in a government witness protection
program.4 In addition, with respect to individuals convicted of drug
possession offenses, the denial of benefits penalties are to be "waived in
the case of a person who, if there is a reasonable body of evidence to
substantiate such declaration, declares himself to be an addict and
submits himself to a long-term treatment program for addiction, or is
deemed to be rehabilitated pursuant to rules established by the Secretary
of Health and Human Services."5 Also, the period of ineligibility for the
denial of benefits is to be suspended for individuals who have completed a
supervised drug rehabilitation program, have otherwise been rehabilitated,
or have made a good faith effort to gain admission to a supervised drug
rehabilitation program but have been unable to do so because of

421 U.S.C. S: 862(e). 521 U.S.C. S: 862(b)(2).

Appendix V: Legal and Administrative Framework for the Denial of Federal
Benefits Program

Administration of the Law

inaccessibility or unavailability of such a program or the inability of
such individuals to pay for such a program.6

State and federal sentencing judges generally have discretion to impose
denial of federal benefits, under section 5301 of the Anti-Drug Abuse Act
of 1988, as a sanction. This sanction can be imposed in combination with
other sanctions, and courts have the option of denying all or some of the
specified federal benefits and determining the length of the denial period
within certain statutorily set ranges. When denial of benefits under
section 5301 of the Anti-Drug Abuse Act of 1988 is part of a sentence, the
sentencing court is to notify the Bureau of Justice Assistance, which
maintains a database (the Denial of Federal Benefits Program
Clearinghouse) of the names of persons who have been convicted and the
benefits that they have been denied. BJA passes this information on to the
U.S. General Services Administration (GSA), which maintains the debarment
list for all agencies. GSA publishes the names of individuals who are
denied benefits in the Lists of Parties Excluded from Federal Procurement
or Nonprocurement Programs, commonly known as the Debarment List. The
Debarment List contains special codes that indicate whether all or
selected benefits have been denied for an individual and the expiration
date for the period of denial. Before making an award or conferring a
pertinent federal benefit, federal agencies are required to consult the
Debarment List to determine if the individual is eligible for benefits.

The Department of Justice also has data-sharing agreements with the
Department of Education and the Federal Communications Commission. The
purpose of these agreements is to provide these agencies with access to
information about persons currently denied the federal benefits
administered by them. For example, as described in this report, students
who are convicted of offenses involving the sale or possession of a
controlled substance are ineligible to receive certain federal
postsecondary education benefits. In order to ensure that student
financial assistance is not awarded to individuals subject to denial of
benefits under court orders issued pursuant to section 5301, DOJ and the
Department of Education implemented a computer matching program. The
Department of Education attempts to identify persons who have applied for
federal higher education assistance by matching records from applicants
against

621 U.S.C. S: 862(c ).

    Appendix V: Legal and Administrative Framework for the Denial of Federal
                                Benefits Program

the BJA database list of persons who have been denied benefits. Officials
at the Department of Education report that the department has matched only
a few records of applicants for federal higher education assistance to the
DOJ list of persons denied federal benefits. The individuals whose names
appear on the DOJ list may differ from those individuals who selfcertify
to a drug offense conviction on their applications for federal
postsecondary education assistance. (See app. III for more information on
this.)

The Administrative Office of United States Courts is responsible for
administrative matters for the federal courts. Shortly after the passage
of the Anti-Drug Abuse Act of 1988, AOUSC added the Denial of Federal
Benefits sentence enhancement to the Pre-Sentence Report Monograph, which
provided information to probation officers about the availability of the
DFB as sanction along with its requirements. AOUSC also developed a
standard form for federal judges to use in reporting the imposition of the
Denial of Federal Benefits sanctions; the form is part of the Judgment and
Commitment Order that is completed by the court upon sentencing.

The United States Sentencing Commission promulgates federal sentencing
guidelines and collects data on all persons sentenced pursuant to the
federal sentencing guidelines. After the passage of the Anti-Drug Abuse
Act of 1988, the USSC prepared a guideline for this sanction and included
it in the Sentencing Guidelines Manual. Annually, USSC distributes the
Sentencing Guidelines Manual to federal court officials.

Use of the Denial of Federal Benefits Program Provisions

Bureau of Justice Assistance data show that between 1990 and the second
quarter of 2004, 8,298 offenders were reported as having been denied
federal benefits by judges who imposed sanctions under the Denial of
Federal Benefits Program. About 38 percent (or 3,128) of these offenders
were denied benefits in state courts, and about 62 percent (or 5,170) were
denied benefits in federal courts. About an average 635 persons per year
were denied benefits under the program over the 1992 through 2003 period,
and the number denied in any given year ranged from about 428 to 833. The
number denied a benefit under the program decreased to 428 in 2002 and
increased to 595 in 2003.7

7The second quarter of 2004 was the most recent BJA data on the number of
offenders denied benefits under the DFB.

Appendix V: Legal and Administrative Framework for the Denial of Federal
Benefits Program

According to BJA data, judges in comparatively few courts used the denial
of federal benefits provisions. State court judges in 7 states and federal
judges in judicial districts in 26 states were reported to have imposed
the sanction. Among state courts, judges in Texas accounted for 39 percent
of the state court totals, while judges in Oregon and Rhode Island
accounted for 30 percent and 13 percent, respectively. Among the federal
courts, judges in judicial districts in Texas accounted for 21 percent of
the federal totals, while judges in North Carolina, Mississippi, Georgia,
Florida, Nevada, and Kansas accounted for between 8 percent and 15 percent
of the totals. Federal judges in each of the remaining 19 states accounted
for less than 3 percent of the federal totals.

Not all of the 8,298 offenders recorded as having been denied federal
benefits between 1990 and 2004 under the program are currently denied
benefits. For about 75 percent of these offenders, the period of denial
has expired. Officials at BJA report that as of April 2004, they
maintained about 2,000 active records of persons currently under a period
of denial.

Relative to the total number of felony drug convictions, the provisions of
the Denial of Federal Benefits (DFB) Program are reportedly used in a
relatively small percentage of drug cases. For example, biannually between
1992 and 2000, there were a minimum of 274,000 and a maximum of 348,000
convictions for drug offenses in state courts, or about 307,000 per year.
In federal courts over this same period, there were between 15,000 and
24,000 drug offenders convicted, or about 19,000 per year. As the average
annual number of drug defendants in state courts denied benefits under the
DFB was 223, the rate of use of the DFB in state courts averaged about
0.07 percent. Among federal drug defendants, the annual average number
reported as having received a sanction under the program was about 369,
while the average annual number of drug defendants sentenced federally was
about 19,000; hence, the percentage of all federal drug defendants
receiving a sanction under the program was about 2 percent.

Appendix V: Legal and Administrative Framework for the Denial of Federal
Benefits Program

Assessing the Impacts of the Denial of Federal Benefits Program

Throughout the history of the program, questions have been raised about
its apparently limited impacts. In 1992, we reported on the difficulties
in denying federal benefits to convicted drug offenders8 and suggested
that there would not be widespread withholding of federal benefits from
drug offenders.9 Officials at BJA also reported that the sanction has not
been widely used by judges. In 2001, BJA program managers met with some
U.S. attorneys in an attempt to provide them with information about the
potential benefits of the program. According to BJA officials, the U.S.
attorneys responded that they typically used other statutes for
sanctioning and sentencing drug offenders, rather than the sanctions under
the Denial of Federal Benefits Program.

The benefits that can be denied under the program-federal contracts,
grants, loans, and professional or commercial licenses-suggest some
reasons as to its relatively infrequent use. Persons engaged in federal
contracting, for example, are generally engaged in business activities,
and such persons compose small percentages of federal defendants sentenced
for drug offenses. Hence, relatively few defendants may qualify to use
these federal benefits, and therefore relatively few may be denied the
benefits.

8GAO, Drug Control: Denying Federal Benefits to Convicted Drug Offenders,
GAO/GGD-92-56 (Washington, D.C.: April 21, 1992).

9GAO's opinion was based on the following factors: (1) the views held by
those who may affect the imposition of the sentence-judges and other
criminal justice officials-that the sentence would not have much impact on
many of the offenders convicted in federal and state courts; (2) accepted
court sentencing practices such as excluding many first-time drug
offenders-such as those charged with drug possession-from receiving such a
sentence, and (3) federal benefit administration policies and practices
such as those that preclude the interruption or termination of ongoing
benefits.

Appendix V: Legal and Administrative Framework for the Denial of Federal
Benefits Program

Reliable Data by Race	None of the data sources that we reviewed provided
reliable data on the race and ethnicity of persons denied federal benefits
under the Denial of

of Persons Receiving Federal Benefits Program.

the DFB Are Not

Available

Appendix VI: GAO Contact and Staff Acknowledgments

GAO Contact Acknowledgments

(440270)

Laurie E. Ekstrand, (202) 512-8777

In addition to the contact named above, William J. Sabol, Clarence Tull,
Brian Sklar, DuEwa Kamara, Geoffrey Hamilton, David Makoto Hudson, Michele
Fejfar, David Alexander, Amy Bernstein, Anne Laffoon, Julian King and
Andrea P. Smith made key contributions to this report.

GAO's Mission

Obtaining Copies of GAO Reports and Testimony

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.

The fastest and easiest way to obtain copies of GAO documents at no cost
is through GAO's Web site (www.gao.gov). Each weekday, GAO posts newly
released reports, testimony, and correspondence on its Web site. To have
GAO e-mail you a list of newly posted products every afternoon, go to
www.gao.gov and select "Subscribe to Updates."

Order by Mail or Phone 	The first copy of each printed report is free.
Additional copies are $2 each. A check or money order should be made out
to the Superintendent of Documents. GAO also accepts VISA and Mastercard.
Orders for 100 or more copies mailed to a single address are discounted 25
percent. Orders should be sent to:

U.S. Government Accountability Office 441 G Street NW, Room LM Washington,
D.C. 20548

To order by Phone: 	Voice: (202) 512-6000 TDD: (202) 512-2537 Fax: (202)
512-6061

To Report Fraud, Contact:

Waste, and Abuse in Web site: www.gao.gov/fraudnet/fraudnet.htm

E-mail: [email protected] Programs Automated answering system: (800)
424-5454 or (202) 512-7470

Gloria Jarmon, Managing Director, [email protected] (202)
512-4400Congressional U.S. Government Accountability Office, 441 G Street
NW, Room 7125 Relations Washington, D.C. 20548

Public Affairs 	Paul Anderson, Managing Director, [email protected] (202)
512-4800 U.S. Government Accountability Office, 441 G Street NW, Room 7149
Washington, D.C. 20548

                           PRINTED ON RECYCLED PAPER
*** End of document. ***