Welfare Reform: Improving State Automated Systems Requires Coordinated
Federal Effort (Chapter Report, 04/27/2000, GAO/HEHS-00-48).

GAO examined the ability of the states' automated systems to provide
information needed for state and local officials to help low-income
individuals with children obtain employment and become economically
independent. GAO found that better systems would help ensure that the
intended goals and requirements of Temporary Assistance for Needy
Families (TANF), such as promoting work and enforcing the five-year time
limit on aid, are met. Many of the obstacles states face may be best
addressed at the state level, such as the need for collaboration among
state and local agencies. However, coordinated federal action would help
address several of the obstacles the states encounter as they take steps
to improve their automated systems for social programs. However, no
group or organization formally brings together the array of federal
agencies involved in welfare reform to help devise solutions to the
issues the states face in improving their welfare and welfare-related
automated systems. Health and Human Services could play a pivotal role
in orchestrating such a broad-based collaborative approach by bringing
representatives of key federal agencies and other organizations together
to work on these issues.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  HEHS-00-48
     TITLE:  Welfare Reform: Improving State Automated Systems Requires
	     Coordinated Federal Effort
      DATE:  04/27/2000
   SUBJECT:  Public assistance programs
	     Workfare
	     Federal/state relations
	     State-administered programs
	     Program graduation
	     Disadvantaged persons
	     Management information systems
	     Accountability
IDENTIFIER:  Aid to Families with Dependent Children Program
	     HHS Temporary Assistance for Needy Families Program
	     AFDC
	     Food Stamp Program
	     Medicaid Program
	     Georgia
	     New Jersey
	     Ohio
	     Texas
	     Washington
	     Wisconsin

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Testimony.                                               **
**                                                              **
** 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/HEHS-00-48
8

16

Federal Funding and Oversight Have Played Key Roles in Developing
States' Automated Welfare Systems 18

Welfare Reform Legislation Established Numerous Requirements for
Those Receiving TANF and for States 21

New and Evolving Welfare Environment Is Transforming States'
Automated Systems Needs 24

Objectives, Scope, and Methodology 27

State and Local Efforts to Help TANF Recipients Move to Employment

31

Insufficient Links Among Automated Systems Constrain Case
Managers' Ability to Coordinate Services and Monitor Recipients' Progress 31

Difficulties in Querying Automated Systems to Obtain Needed
Information Limit Capabilities for Service Planning 37

Automated Systems Vary in Capabilities to Support Program
Oversight 41

Ability to Obtain and Analyze Data From Multiple Sources

47

States Seek to Improve Capabilities for Coordinated Service
Delivery by Expanding Automated System Links 47

States Are Developing Databases and Query Tools to Improve
Managers' Capabilities to Analyze Data Across Programs 50

in Improving Automated Systems

53

States Face a Range of Obstacles to Improving Their Automated
Systems 53

Federal Actions Could Help States Improve Automated Systems 60

Conclusions 63

Recommendation to the Secretary of Health and Human Services 63

Agency Comments and Our Response 64

Appendix I: The GAO/Rockefeller Institute Working Seminar on
Social Program Information Systems: Objectives, History, and Membership

66

Appendix II: Links Among Automated Systems for Different Social Programs

69

Appendix III: Comments From the Department of Health and
Human Services

71

Appendix IV: Comments From the Department of Agriculture

76

Appendix V: Comments From the Department of Labor

78

Appendix VI: GAO Contacts and Staff Acknowledgments

79

80

82

85

Table 1: Increasing Participation Requirements for One-Parent and Two-Parent
Families Under Federal Law 22

Table 2: Information That States Need to Meet Federal Reporting Requirements
and to Participate in the Optional High-Performance
Bonus Program 23

Table 3: Key Information Users and Functions They Perform to Help Low-Income
Families Become Employed 26

Table 4: Case Study States and Local Sites 29

Table 5: Respondents to the Automated Systems Questionnaire 30

Table 6: Capabilities of Automated Systems in Six States to Support Selected
TANF Case Management Activities 36

Table 7: Capabilities of Automated Systems in Six States to Provide Program
Managers With Information on Selected Caseload Characteristics 40

Table 8: Automated Systems Projects Being Designed to Help
Frontline Workers Coordinate Service Delivery 48

Table 9: Automated Systems Projects Being Designed to Enhance Analytical
Capabilities of Program Managers 51

Figure 1: Broad Array of Programs and Services That Case Managers
Can Draw Upon to Serve Their Clients 25

Figure 2: Proportion of Information Needed for Case Management
That Is Provided by Automated Systems 32

Figure 3: Automated Systems That TANF Case Managers Can Access
From Their Desktop Computers 35

Figure 4: Proportion of Information Needed for Service Planning
That Is Provided by Automated Systems 38

Figure 5: Proportion of Information Needed for Monitoring Program
Performance With Respect to Employment Progress of TANF
Recipients That Is Provided by Automated Systems 42

Figure 6: Programs With an Automated System That Either Is the
Same as, or Shares Data With, the System Used for Determining
TANF Eligibility 70

ACF Administration for Children and Families

AFDC Aid to Families With Dependent Children

APD advanced planning document

APHSA American Public Human Services Association

FAMIS Family Assistance Management Information System

FNS Food and Nutrition Service

HCFA Health Care Financing Administration

HHS Department of Health and Human Services

JOBS Job Opportunities and Basic Skills Training

JTPA Job Training and Partnership Act

PRWORA Personal Responsibility and Work Opportunity Reconciliation Act of
1996

TANF Temporary Assistance for Needy Families

UI unemployment insurance

USDA Department of Agriculture

Health, Education, and
Human Services Division

B-282551

April 27, 2000

Congressional Committees

The sweeping welfare reforms enacted by the Personal Responsibility and Work
Opportunity Reconciliation Act of 1996 (P.L. 104-193) have profound
implications for the information needs of states and the automated systems
designed to meet these needs. This report, Welfare Reform: Improving State
Automated Systems Requires Coordinated Federal Effort (GAO/HEHS-00-48),
based on a research and development effort by GAO, examines the capabilities
of states' automated systems to provide information needed for state and
local officials to help low-income individuals with children obtain
employment and become economically independent. In addition, the report
provides information on approaches states are using to improve their
automated systems, obstacles they have encountered in this process, and the
potential role of the federal government in helping overcome these
obstacles. We are making a recommendation to the Secretary of Health and
Human Services. (The complete list of committees to which this report is
addressed appears at the end of this letter.)

We are sending copies of this report to the Honorable Donna E. Shalala,
Secretary of Health and Human Services; the Honorable Dan Glickman,
Secretary of Agriculture; the Honorable Alexis M. Herman, Secretary of
Labor; and other interested parties. We will also make copies available to
others on request.

If you or your staff have any questions concerning this report, please call
me at (202) 512-7215 or Gale C. Harris at (202) 512-7235. Other GAO contacts
and staff acknowledgments for this report are listed in appendix VI.

Cynthia M. Fagnoni
Director, Education, Workforce, and
Income Security Issues

List of Congressional Committees

The Honorable Richard G. Lugar, Chairman
The Honorable Tom Harkin, Ranking Minority Member
Committee on Agriculture, Nutrition, and Forestry
United States Senate

The Honorable William V. Roth, Jr., Chairman
The Honorable Daniel Patrick Moynihan, Ranking Minority Member
Committee on Finance
United States Senate

The Honorable Fred Thompson, Chairman
The Honorable Joseph I. Lieberman, Ranking Minority Member
Committee on Governmental Affairs
United States Senate

The Honorable James M. Jeffords, Chairman
The Honorable Edward M. Kennedy, Ranking Minority Member
Committee on Health, Education, Labor, and Pensions
United States Senate

The Honorable Larry Combest, Chairman
The Honorable Charles W. Stenholm, Ranking Minority Member
Committee on Agriculture
House of Representatives

The Honorable Tom Bliley, Chairman
The Honorable John D. Dingell, Ranking Minority Member
Committee on Commerce
House of Representatives

The Honorable William F. Goodling, Chairman
The Honorable William L. Clay, Ranking Minority Member
Committee on Education and the Workforce
House of Representatives

The Honorable Dan Burton, Chairman
The Honorable Henry A. Waxman, Ranking Minority Member
Committee on Government Reform
House of Representatives

The Honorable Bill Archer, Chairman
The Honorable Charles B. Rangel, Ranking Minority Member
Committee on Ways and Means
House of Representatives

Executive Summary

In the wake of the Personal Responsibility and Work Opportunity
Reconciliation Act of 1996 (PRWORA), states' assistance programs for needy
families with children have undergone dramatic shifts in objectives,
policies, and operations. PRWORA replaced the Aid to Families With Dependent
Children (AFDC) program with a block grant to states to provide Temporary
Assistance for Needy Families (TANF). PRWORA gave states greater flexibility
in designing their programs under TANF but also established new
accountability measures for states and a 5-year lifetime limit on TANF
assistance. These measures heighten the importance of helping TANF
recipients find work quickly and retain employment. Moreover, as welfare
agencies focus on moving needy families toward economic independence, these
agencies are drawing on numerous federal and state programs--often
administered by separate agencies--to provide a wide array of services, such
as child care, food stamps, and employment and training services.1 These
sweeping changes have profound implications for the information needs of
states and the automated systems designed to meet those needs.

To provide information for congressional oversight, GAO reviewed states'
efforts to meet the information needs associated with welfare reform, with a
focus on TANF. GAO (1) assessed the extent to which current automated
systems in selected states meet key information needs of programs that help
low-income individuals with children obtain employment and become
economically independent, (2) identified the approaches states are using to
develop or modify their automated systems to better meet these information
needs, and (3) identified the major obstacles states have encountered in
working to improve their automated systems as well as the potential role of
the federal government in helping overcome these obstacles. In collaboration
with field researchers from the Nelson A. Rockefeller Institute of
Government's State Capacity Project, GAO conducted case studies at the state
and local levels in six states and collected supplemental information
through a survey of these six and nine additional states and some of their
localities. In addition, GAO jointly established the GAO/Rockefeller
Institute Working Seminar on Social Program Information Systems to provide
an ongoing source of expertise in this area.2

Although automated systems in the states GAO examined support welfare reform
in many ways, a number of these systems have major limitations in one or
more of three key areas. With respect to information needs for case
management, the major shortcoming--which exists to varying degrees across
the states--is an inability to obtain data on individual TANF recipients
from some of the agencies serving them, including job assistance agencies.
This situation makes it difficult for TANF case managers to arrange needed
services; ensure that the services are provided; and respond quickly when
problems arise, such as when a recipient does not attend a scheduled work
activity. Second, officials in the states, especially those at the local
level, said that it is sometimes difficult or impossible to query automated
systems to obtain information for planning service strategies for their
overall TANF caseloads, such as information on the number of adults with no
prior work experience. Finally, automated systems have shortcomings for
program oversight purposes; specifically, they do not provide enough
information to support enforcement of the 5-year TANF time limit and to
monitor the employment progress of TANF recipients overall in some
instances.

States' automated systems projects embody a range of approaches to expanding
the ability of system users to obtain and analyze data from multiple
sources. Some projects are designed primarily to support TANF case managers
and other frontline workers in providing more coordinated delivery of
services. Ohio and Texas, for example, are developing "common front ends,"
which are intended to provide a single point of access to multiple systems
for frontline workers and reduce duplicate data entry by automatically
transferring data across systems. These common front ends can reduce data
entry errors and free up workers to spend more time working directly with
families. Other projects, geared more to improving the ability of program
managers to collect and analyze data from different programs, involve
developing new query tools and databases that are expected to help program
managers with key tasks, such as determining program results and assessing
the performance of service providers. Modifying and developing automated
systems that better support welfare reform is a long-term and evolving
process.

States face a number of obstacles to improving their automated systems, such
as the magnitude of changes in the mission and operations of welfare
agencies due to welfare reform, the inherent difficulties associated with
successfully managing information technology projects, competition with the
private sector to recruit and retain information technology staff, and the
complexity of obtaining federal funding for systems projects that involve
multiple agencies. The federal government could take various actions to help
overcome such obstacles, such as providing more information on best
practices for managing information technology. In this way, the federal
government could serve a facilitative role, in addition to its regulatory
role, in helping states improve automated systems for social programs.
However, no group or organization that brings together key federal agencies
involved in welfare reform has been formally charged with devising solutions
to the range of obstacles confronting states. GAO is thus recommending that
the Secretary of Health and Human Services (HHS) establish an interagency
group that would play such a role.

Local Efforts to Help TANF Recipients Become Employed

Of the 15 localities responding to GAO's survey on the extent to which
automated systems support case management, 5 indicated that their systems
provide all or most of the information that TANF case managers need, whereas
10 said that systems provide about half or less of the information needed. A
major shortcoming cited by officials in GAO's case study states is that some
of the automated systems used by agencies providing services to TANF
recipients do not share information about these recipients. For example,
local officials in New Jersey told us that some TANF recipients had received
sanctions in error because data on their attendance at work activities are
not shared between the automated systems used by the welfare and labor
departments.3

Automated systems in some cases also have shortcomings that limit the
ability of program managers to obtain information needed for service
planning. While state officials from 8 of the 15 states surveyed indicated
that automated systems provide all or most of the caseload information that
program managers need for service planning, local officials from 6 of the 15
localities responded similarly. For example, local officials at one site
commented that data on the characteristics of TANF recipients contained in
the state's automated systems are often not available in a format that can
be easily manipulated, and, as a result, accessing data depends on the
technical expertise of the user. Difficulties in accessing such data can
limit the ability of managers to identify and meet the service needs of
their caseloads.

GAO also identified a gap in the ability of automated systems to support
enforcement of the 5-year TANF time limit and to provide program managers
with information to monitor TANF recipients' employment progress. Officials
in five of the six case study states said that their automated systems
provide data on the number of months countable toward a TANF recipient's
time limit in their state. However, these states generally reported that
they do not collect data on recipients' prior receipt of TANF in other
states or that they rely on TANF recipients to disclose this information,
which is an unreliable method that could lead to payments to ineligible
individuals. In addition, local officials in GAO's case study states
generally reported that they can obtain information on the number of TANF
recipients that enter employment. However, local officials vary in their
capabilities to obtain aggregate information on the job retention rates of
these recipients and generally cannot obtain information on the extent to
which recipients have increased their wage levels.

Obtain and Analyze Data From Multiple Sources

States' automated systems projects are using various approaches to support
more coordinated delivery of services to low-income families, including
developing new links among separate automated systems, replacing existing
systems with new integrated systems, and constructing electronic networks to
link agencies and service providers. New Jersey's One Ease-E Link project
aims to establish local electronic networks that both expand access to
client data among agencies serving TANF recipients and provide such
capabilities as automated appointment scheduling for recipients. In
addition, states have projects under way to expand program managers'
abilities to obtain and analyze information that covers a broader range of
programs and is more comprehensive than that contained in preprogrammed
reports. Several states are developing large databases that are designed to
extract data from different programs' automated systems and enable users to
generate customized reports to meet their information needs, such as
determining the extent to which former TANF recipients are receiving food
stamps or Medicaid.

Improving Automated Systems

Some of the obstacles cited by states pertain to the large-scale changes in
the mission and operations of welfare agencies that complicate states'
efforts to manage information technology projects, including states' need to
define the functions a system must support, translate these functions into
specific requirements, and then identify the applications and hardware that
will be used to meet these needs. In addition, the need to collaborate with
new partners in carrying out these tasks and to obtain cross-state
information to enforce the 5-year TANF time limit has compounded the
difficulties involved. Other obstacles involve obtaining sufficient staff
resources, such as difficulties in competing with the private sector to
recruit and retain qualified information technology staff. In addition,
states maintained that the federal review process for systems procurement is
too cumbersome when a project requires approvals from multiple federal
agencies.

While some of these obstacles may be best addressed at the state level, GAO
identified four key areas in which federal actions could better facilitate
states' efforts to improve their automated systems:

ï¿½ Disseminate information on best practices for managing information
technology generally as well as best practices specific to automated systems
that support welfare reform.

ï¿½ Review and modify as needed the federal process for systems procurement to
ensure that it meets federal needs for state accountability without
unnecessarily hindering state development efforts.

ï¿½ Facilitate links among the automated systems used by different state and
local agencies through such means as supporting demonstrations designed to
promote better partnerships between state and local agencies and
coordinating data reporting requirements for different federal programs.

ï¿½ Address the need for states to have access to cross-state information on
individuals' TANF receipt to enable enforcement of the 5-year TANF time
limit.

The issues cross various federal programs and agencies. HHS could play a
pivotal role in orchestrating a collaborative approach among the agencies by
bringing representatives from each of them together to work on these issues.
Increased federal attention to the ongoing and evolving process of improving
states' automated systems for social programs could ultimately help bring
about more effective and efficient service delivery for low-income families.

GAO recommends that the Secretary of HHS establish an interagency group to
identify, and develop implementation plans for, federal actions that would
facilitate states' efforts to improve their automated systems for federal
programs that serve low-income families. The group should include high-level
federal officials from HHS, the Department of Agriculture (USDA), the
Department of Labor, and other federal agencies as appropriate. The
interagency group should obtain input from state and local social program
and information technology managers; the group could also seek input from
others, such as state organizations, professional organizations of social
program officials, and welfare researchers. In addition to any actions that
the interagency group identifies on its own, the group should consider
actions in the four key areas GAO identified for a federal facilitative
role.

GAO obtained comments on a draft of this report from HHS, USDA, and Labor
(these comments are included in apps. III, IV, and V). USDA and Labor agreed
with GAO's recommendation. HHS said that it would review, along with other
involved federal agencies, the federal process for approving state systems
acquisitions. However, HHS did not say that it would take responsibility for
establishing an interagency group to conduct such a review. GAO believes
that HHS is the appropriate agency to establish the group because HHS
oversees many of the affected programs. Furthermore, GAO's recommendation
envisioned more than just a review of systems procurement requirements.
Identifying, and developing implementation plans for, needed actions in the
other recommended areas would further assist states in improving the
effectiveness of their automated systems.

Introduction

The Personal Responsibility and Work Opportunity Reconciliation Act of 1996
(PRWORA) significantly changed federal policy for aiding low-income families
with children. It ended the 61-year-old Aid to Families With Dependent
Children (AFDC) program, under which eligible families were entitled to
ongoing monthly cash assistance, and replaced it with the Temporary
Assistance for Needy Families (TANF) block grant for states. The TANF block
grant, administered by the Department of Health and Human Services (HHS),
provides a total of up to $16.5 billion to the states each year through
2002. As specified in PRWORA, the goals of TANF include ending the
dependence of needy parents on government benefits by promoting job
preparation and work. Under TANF, states have increased flexibility to
design their programs and are therefore able to build upon initiatives they
had begun experimenting with before federal reform. However, among other
provisions, the law also requires that, to avoid financial penalties, states
impose work requirements for adults, meet steadily rising requirements for
the percentage of adults who participate in work activities, and enforce a
5-year lifetime limit on receiving TANF. To ensure they meet these and other
requirements, states must also meet new federal program and fiscal reporting
requirements.

The new emphasis on work, job placement, and the temporary nature of aid
requires a fundamental shift in how welfare offices do business and has
implications beyond the TANF program and its workers. The goal of welfare
workers and program managers is to help families become self-supporting.
This new work-focused welfare affects other key programs that can help needy
families in their transition to employment, most prominently the Food Stamp
program, Medicaid, and the employment and training programs that are a part
of each state's workforce development system.4 Work-focused welfare also has
an effect on the multitude of service providers--public, nonprofit, and
private--that are involved in helping low-income families move toward
economic independence.

Because of the new ways of doing business, the increased number of agencies
and service providers involved, and the expanded federal reporting
requirements, welfare reform has placed increased demands on state
information systems that have traditionally been critical to the successful
operation of states' welfare programs. Under the AFDC program, information
systems were primarily used to determine eligibility and to provide some
data for oversight, typically aggregate data on caseloads and expenditures,
to state and federal program managers.5 In the new welfare environment,
however, information systems must support an expanded set of functions in
three key areas: case management, including eligibility determination;
service planning; and program oversight, with a new emphasis on outcomes and
results. In addition, the increased devolution of responsibility for program
operation and performance to states and localities increases the need for
systems that can respond to the multiple needs of users at all levels of
government. State and local information systems used in operating and
overseeing welfare programs will be important to the effective
implementation of the new welfare program.

States' Automated Welfare Systems

Three of the federal government's major programs for needy
families--AFDC/TANF, Medicaid, and the Food Stamp program--have historically
relied heavily on state-run automated computer systems to help determine
applicants' eligibility and the amount of assistance each participant should
receive. Recognizing the importance of automated systems in efficiently and
accurately determining eligibility, over the years the Congress has acted to
encourage states to develop automated systems for these programs.
Specifically, the Congress authorized HHS' Administration for Children and
Families (ACF), which oversaw the AFDC program, and HHS' Health Care
Financing Administration (HCFA), which oversees Medicaid, to reimburse
states for a significant proportion of their total costs to develop and
operate automated eligibility determination systems.6 In 1980 the Congress
also authorized the Department of Agriculture's (USDA) Food and Nutrition
Service (FNS), which oversees the Food Stamp program, to reimburse states
for 75 percent of their costs for planning, designing, developing, and
installing automated eligibility systems and 50 percent of the costs to
operate these systems.7 The goals of these automated eligibility systems
were to minimize mistakes and the amount of time needed to determine
eligibility, and to lower administrative costs.

To obtain federal funding for a portion of the cost of automated welfare
systems development and implementation for AFDC, Medicaid, and food stamps,
states have been required to follow the advanced planning document (APD)
process. Under this process, states submit APDs to the specific oversight
agencies detailing state plans to develop and implement automated systems.
The federal agencies then make funding decisions on the basis of these APD
submissions. If a system is to be used for more than one program, APDs that
meet the separate requirements of each program must be submitted and
approved, and development costs are allocated to the various programs. After
approving funding to states, the federal agencies monitor development and
operation of the state systems to ensure all federal requirements are met.
To obtain enhanced funding for AFDC automated systems, states had to meet
the requirements for a Family Assistance Management Information System
(FAMIS), a general system design developed by HHS to improve state
administration of the AFDC program. Because eligibility for Medicaid and
food stamps was linked to eligibility for AFDC, most of the AFDC systems
also covered Medicaid and food stamps. To the extent that Medicaid and food
stamp eligibility systems were integrated with AFDC systems, ACF, HCFA, and
FNS used a cost allocation method to determine the portion of the system
costs to be borne by each agency. At the time PRWORA was enacted in August
1996, 38 states operated state systems that complied with FAMIS
requirements.8

The APD process remains in place for Medicaid and food stamps. However,
under TANF, states are no longer required to submit APDs, and ACF is no
longer required to play a role in establishing system requirements or in
monitoring system development. States may use whatever portion of their
capped federal TANF block grant funds they wish for developing and operating
systems. However, no legislatively mandated ceilings exist to specifically
limit federal funds for states' development and operation of automated
systems costs related to Medicaid and food stamps. According to data
reported by states to HHS, states had expended a total of about $456
million, or about 2 percent, of federal TANF funds on systems from fiscal
year 1997 through the second quarter of fiscal year 1999.9

HHS also provides states with a proportion of the costs of automated systems
for other federal welfare-related programs, including the child support
enforcement program, child care subsidy program, child welfare services, and
foster care/adoption assistance. We reported in 199410 that 52 of the 54
states and territories were operating multiple systems--ranging from 2 to 12
systems each--to provide welfare program support for AFDC, Medicaid, food
stamps, child support enforcement, child care, child welfare services,
foster care/adoption assistance, and the now-repealed Job Opportunities and
Basic Skills Training (JOBS) program.11 We also reported that HHS and USDA
had contributed over $6.8 billion from 1984 to 1992, and over $1.8 billion
prior to 1984, to help fund development and operation of these automated
systems.12 The 1994 report also asserted that the federal government might
pay the largest share of an estimated $10.7 billion in additional automated
systems costs from fiscal year 1993 through the end of 1999. These cost
figures do not include federal funding for other automated systems used by
programs involved in helping welfare recipients move toward economic
independence, such as systems related to state workforce development
systems.

Recognizing the importance of information systems in implementing key
aspects of welfare reform, the Congress in PRWORA directed HHS to prepare a
report that addressed (1) the status of the automated systems operated by
states to help program managers administer their TANF programs and (2) what
would be required to establish a system capable of tracking participants in
public programs over time and checking case records of the states to
determine whether individuals were participating illegally in public
programs in two or more states at a time. In collecting information intended
for use in this report, HHS worked with several organizations representing
states, information technology specialists, and welfare administrators to
survey states on the likely effects on information systems of legislatively
mandated changes to welfare programs. The 1997 report projected on the basis
of state estimates that adapting current information systems to meet the
demands of welfare reform would require spending about $1 billion.13

Those Receiving TANF and for States

While PRWORA limited the federal government's role in approving and
overseeing states' development of TANF automated information systems and
provided states much greater flexibility than before to design and implement
their own programs, it also emphasized states' accountability for meeting
program requirements and for program performance. This accountability focus
is in keeping with the increased emphasis governmentwide on program
performance fostered by the Government Performance and Results Act of 1993.
Under TANF, to avoid financial penalties, states must ensure that a certain
minimum percentage of their caseloads are participating in work or
work-related activities each year. These percentages are referred to as
"minimum mandated participation rates." To count toward states' mandated
rates, adult recipients in families must participate a certain minimum
number of hours in work or a work-related activity as prescribed in the law,
such as job readiness workshops; on-the-job training; and, under certain
circumstances, education. The required number of hours of participation and
the percentage of a state's caseload that must participate to meet mandated
rates generally increase over time, as shown in table 1.

Table 1: Increasing Participation Requirements for One-Parent and Two-Parent
Families Under Federal Law

                                          Fiscal year
                                          1997 1998  1999  2000  2001 2002
 Minimum weekly average hours of
 participation
 One-parent families                      20   20    25    30    30   30
 Two-parent families                      35   35    35    35    35   35
 Minimum mandated participation rates
 (percentages)
 All families                             25   30    35    40    45   50
 Two-parent families                      75   75    90    90    90   90

Source: 42 U.S.C. sec. 607.

States must also track the length of time that individuals receive TANF to
implement the new time limits. States must enforce the PRWORA prohibition on
provision of TANF funds to families with adults who have received TANF for a
total of 5 years over their lifetime. Families with no adult receiving
assistance (commonly referred to as "child-only" cases) are not subject to
this limit, and up to 20 percent of a state's average monthly caseload may
be exempt on the basis of hardship or having been subjected to domestic
violence.14 Also, states may opt to continue to provide assistance beyond
the 5-year limit using state funds.

In addition to holding states accountable for tracking and enforcing
participation requirements and time limits, PRWORA authorized HHS to award
federal dollars to states with exceptional performance in achieving the
goals of TANF. PRWORA established the broad parameters of a high-performance
bonus system and required that HHS work with organizations representing
states and welfare administrators to determine the details of the system.
For the first year that the bonus was to be awarded--fiscal year
1999--states were to compete in the areas of TANF recipients' rates of job
entry, earnings gains, and job retention as well as increases in these rates
over time.

The new focus on work as well as the new high-performance bonus system
resulted in an expanded set of reporting requirements and options for
states, as shown in table 2. HHS issued final regulations on data collection
and reporting requirements in April 1999 and published information on the
first year of the high-performance bonus system in March 1998.15

Table 2: Information That States Need to Meet Federal Reporting Requirements
and to Participate in the Optional High-Performance Bonus Program

 Reports to HHS                     Information contained in the reports
                                    Case-level data on the characteristics
                                    and circumstances of families receiving
                                    TANF

 TANF Data Report                   Case-level data on families at the time
                                    their case is closed

                                    Aggregate information on families at
                                    application for TANF, while receiving
                                    it, and at case closure
                                    Data on expenditures of federal TANF
 TANF Financial Report              funds and state maintenance-of-effort
                                    funds used for TANF or separate state
                                    programsa
                                    Descriptive information on TANF program
                                    and data on child care subsidies
 Report on TANF and state
 maintenance-of-effort programs     Descriptive information and expenditure
                                    data for programs and services funded
                                    with state funds

 Separate state program             Information on families receiving
 maintenance-of-effort data reportb assistance under separate state
                                    programs and at case closure
                                    Information on job entries, job
 TANF high-performance bonus reportcretention, and earnings gains of TANF
                                    recipients and former recipients

aPRWORA requires that states maintain a specified level of their historical
spending for welfare and welfare-related programs to receive federal TANF
funds. The state funds used to meet this maintenance-of-effort requirement
may be used in the TANF program or in separate state programs.

bThis report is required only if a state wishes to compete for a TANF
high-performance bonus or to have its participation rate standards reduced
because of reductions in its caseload, as authorized by PRWORA.

cThis report is required only if a state wishes to be considered for a TANF
high-performance bonus.

Source: TANF Final Rule, Federal Register, Vol. 64, No. 69, Apr. 12, 1999,
and HHS Memorandum No. TANF-ACF-PI-98-05, Aug. 13, 1998.

Automated Systems Needs

Our work and other studies show that states and localities have made
progress in implementing key aspects of TANF. States are refocusing their
programs on moving people into employment rather than signing them up for
monthly cash assistance. To better support this new work focus, many states
are changing how their welfare offices and workers do business; converting
their offices into job placement centers; and helping clients address and
solve problems that interfere with employment, such as lack of child care or
transportation as well as more complex mental and physical health problems.

As welfare agencies focus on moving needy families toward economic
independence, frontline workers are drawing on other federal and state
programs, often administered by separate agencies, to provide a wide array
of services. These programs range from those designed to meet families'
basic needs for food and shelter to those that provide employment and
training services and support services, such as subsidies for child care.
While local welfare agencies typically administer eligibility determination
for TANF, food stamps, and Medicaid, other programs that provide key
services to TANF clients may be administered by separate entities, such as
housing authorities or education agencies. Most notably, because TANF has
focused welfare agencies on employment, a focus that has long been the
province of state and local workforce development systems, welfare agencies
need to work more closely than before with workforce development systems,
particularly to administer services funded by the new welfare-to-work grants
created by the Congress in 1997 and administered by the Department of
Labor.16

In addition, some states now rely on state unemployment insurance (UI)
systems to obtain data on the employment and earnings of former welfare
recipients to help monitor both clients' progress and the performance of
various state programs. Figure 1 shows many of the programs and services
that frontline workers may be responsible for arranging and monitoring to
help clients move toward economic independence through employment.

Figure 1: Broad Array of Programs and Services That Case Managers Can Draw
Upon to Serve Their Clients
The new environment in which state and local agencies provide services to
low-income families has profound implications for the information needs of
states and the automated systems designed to meet those needs. Typically,
the programs to which welfare workers and managers turn to provide services
have automated information systems that were designed to meet the particular
needs of each program rather than the cross-program needs of the clients
they serve. Yet, successful implementation of welfare reform calls for
information from a range of programs in three key areas: case management,
service planning, and program oversight. As shown in table 3, these key
functions are performed at different levels of program administration.

Table 3: Key Information Users and Functions They Perform to Help Low-Income
Families Become Employed

 Key areas of                                 Functions performed by
 information needs Users of the information   information users
                                              Helping TANF clients obtain
                   Case managers, employment  jobs
                   service specialists, and
 Case management   other frontline workers    Providing training and
                   serving individual TANF    support services as needed
                   clients
                                              Monitoring clients' progress
                                              toward economic independence
                                              Collecting aggregate
                   Local and state program    information on the
 Service planning  managers administering     characteristics and service
                   programs                   needs of the caseload to
                                              determine appropriate
                                              services
                                              Monitoring program
                                              performance using information
                                              on

                                              ilocal or state performance
                                              measures

                                              iclients diverted from
                                              receipt of monthly cash
                                              assistance
                   Local and state program
 Program oversight managers overseeing        ifamilies returning to
                   programs                   welfare
                                              ifamilies' use of other
                                              programs, such as food
                                              stamps, Medicaid, and child
                                              care subsidies

                                              ijob entry, job retention,
                                              and wage progression

                                              Meeting federal reporting
                                              requirements

Source: GAO analysis of relevant literature.

While the information needs of the users differ, these needs also share some
key characteristics that distinguish them from the information needs of the
previous welfare environment. First, information is now needed that allows
workers and managers to monitor people's activities and progress over time
toward employment and economic independence. Second, more than ever before,
case and program managers must focus on ensuring the delivery of an
integrated package of services for families seeking aid. While the
importance of service integration--creating methods to unite or link the
services provided by different programs and agencies to serve the same
population--has been acknowledged as a critical need of human services
programs, including welfare, for at least 2 decades,17 the need for
providing integrated services has heightened with the new emphasis on moving
welfare recipients into employment and placing a time limit on the receipt
of aid. The goal of service integration has been reinforced by the Workforce
Investment Act of 1998 (P.L. 105-220), which requires that all states and
localities use one-stop career centers to deliver most employment and
training services (except TANF-related services). These centers seek to
bring together all workforce development programs--each with its own target
population--in a single system that serves all individuals, regardless of
their eligibility for any specific program. In developing one-stop systems,
Labor has encouraged states to involve human services agencies in the
planning and delivery of services.

To better understand and assess the role of information systems in states'
welfare reforms, we collaborated with the Nelson A. Rockefeller Institute of
Government to establish the GAO/Rockefeller Institute Working Seminar on
Social Program Information Systems. The working seminar has about 30
members, including congressional staff, federal and state managers of
information technology, and welfare researchers and has met six times over a
period of almost 2 years. (See app. I for a list of members and summary of
working seminar activities.) The working seminar has been meeting regularly
to determine what changes are needed to facilitate states' efforts to
improve their automated systems to meet the information needs for welfare
reform. To provide information for congressional oversight, we undertook a
review of states' efforts to meet the information needs for welfare reform,
with a focus on the TANF program.

For this study, we (1) assessed the extent to which current automated
systems in selected states meet key information needs of programs that help
low-income individuals with children obtain employment and become
economically independent, (2) identified the approaches states are using to
develop or modify their automated systems to better meet these information
needs, and (3) identified the major obstacles states have encountered in
working to improve their automated systems and the potential role of the
federal government in helping overcome these obstacles. With respect to the
first objective, we focused on three broad types of information needs: those
for case management, service planning, and program oversight. In
collaboration with field researchers from the Nelson A. Rockefeller
Institute's State Capacity Project, we conducted case studies at the state
and local levels in 6 states and collected supplemental information from a
survey of these 6 and 9 additional states and some of their localities, so
that some comparable information was available for a total of 15 states.

In selecting the six states for our in-depth fieldwork, we sought to include
states (1) in varying stages of revamping their automated systems, (2) with
both state- and locally administered TANF programs, and (3) that were
geographically diverse. We interviewed officials in state welfare and labor
departments, including program and information technology managers. In
addition, we interviewed officials at two local sites in each state.
Officials interviewed at local sites included welfare and labor program
managers, information technology managers, TANF case managers, and
supervisors of TANF case managers. At state and local site visits, we
observed demonstrations of automated systems and obtained relevant
documents, such as systems printouts and manuals. (See table 4.)

Table 4: Case Study States and Local Sites

 States     Local sitesa
 Georgia    Fulton County (Atlanta) and Cherokee County (Canton)
 New Jersey Essex County (Newark) and Middlesex County (New Brunswick)
 Ohio       Franklin County (Columbus) and Hamilton County (Cincinnati)
 Texas      Travis County (Austin) and Harris County (Houston)
 Washington Region 4 (Seattle) and Region 5 (Tacoma)b
 Wisconsin  Milwaukee County (Milwaukee) and Dane County (Madison)

aWe visited local site offices located in the cities enclosed in
parentheses.

bThese are regions of the Washington Department of Social and Health
Services.

The state and local interviews were administered using a semistructured
interview guide that we developed through a review of relevant literature
and discussions with members of the GAO/Rockefeller Institute Working
Seminar on Social Program Information Systems. We conducted the interviews
in five of the six states and were accompanied at some interviews by field
researchers from the Rockefeller Institute's State Capacity Project. The
interviews in Georgia were conducted entirely by the Rockefeller Institute's
field researcher for Georgia, an associate professor of political science at
Emory University.

On the basis of our fieldwork in the six states, we identified several key
topics for follow-up work on a larger geographic scale. We developed a
questionnaire to obtain information on four topics: (1) the extent to which
automated systems for different programs share data; (2) the accessibility
of data from different automated systems using the desktop computers of TANF
case managers; (3) the overall extent to which information needs for case
management, service planning, and monitoring program performance are met by
current automated systems; and (4) the most helpful actions that the federal
government could take to help resolve the major obstacles encountered by
states and localities in developing or modifying automated systems. The
questionnaires were administered by our evaluators and field researchers
from the Rockefeller Institute's State Capacity Project. (See table 5.)

Table 5: Respondents to the Automated Systems Questionnaire

 State respondents Local respondents
 Arizona           Phoenix
 Georgia           Fulton County
 Kansas            Kansas City
 Minnesota         Itasca County
 Mississippi       a
 Missouri          Kansas City
 New Jersey        Essex County
 New York          a
 Ohio              Athens County,b Franklin County,b and Hamilton County
 Texas             Travis County and Harris County
 Utah              Salt Lake City
 Washington        Region 4 and Region 6c
 West Virginia     Mercer County
 Wisconsin         Milwaukee County and Dane County
 Wyoming           a

aWe did not obtain local responses.

bWe did not obtain a response to the questions about the extent of automated
capabilities to support case management, service planning, and monitoring
program performance.

cThese are regions of the Washington Department of Social and Health
Services.

We provided a draft of this report to HHS, USDA, Labor, and the six case
study states. We conducted our work from April 1999 to March 2000 in
accordance with generally accepted government auditing standards.

Current Automated Systems Do Not Always Fully Support State and Local
Efforts to Help TANF Recipients Move to Employment

Current automated systems in the states we studied provide support for
implementing and overseeing welfare reform in many critical areas. However,
a number of these systems have shortcomings that limit their usefulness in
helping to move TANF recipients toward employment and economic independence.
One shortcoming is that these automated systems sometimes do not share
information about TANF recipients that is needed by the different agencies
that serve them. This constrains the ability of case managers to coordinate
services and monitor recipients' progress. A second shortcoming is that the
systems' limited capabilities to query and manipulate data sometimes prevent
users from readily obtaining aggregate information on caseload
characteristics. This shortcoming limits capabilities for service planning
and was reported to be especially pervasive at the local level. With regard
to program oversight, we identified some system limitations that negatively
affected the ability of program managers to monitor TANF recipients'
employment progress and enforce the 5-year time limit on federal TANF
assistance.

Ability to Coordinate Services and Monitor Recipients' Progress

Automated systems do not always fully meet the information needs of TANF
case managers--the frontline workers with primary responsibility for
managing services for TANF clients and monitoring their progress toward
employment. Figure 2 presents the assessments of officials from 15
localities of the overall extent to which current automated systems support
case management. While 5 of the localities indicated that automated systems
provide all or most of the information that TANF case managers need to
support their clients' movement to employment and economic independence, the
other 10 localities said that automated systems provide about half or less
of the information needed.

Figure 2: Proportion of Information Needed for Case Management That Is
Provided by Automated Systems
Source: Responses to GAO's questionnaire on automated systems from officials
at local sites.

A major shortcoming of current automated systems, cited to varying degrees
in the six states in which we did in-depth fieldwork, is that some of the
systems used by the agencies providing services to TANF recipients do not
share data on these recipients. This shortcoming constrains the ability of
case managers to arrange needed services; ensure that these services are
provided; and respond quickly when problems arise, such as when a recipient
does not attend a scheduled work activity. For example, local officials in
New Jersey told us that some TANF clients have received sanctions in error
because data are not transferred electronically between the automated system
used by staff at the labor department, who receive data on recipients'
attendance at work activities, and the system used by staff at the welfare
department, who issue sanctions for failure to meet work requirements. Local
officials in Ohio and New Jersey said that TANF case managers are unable to
determine which of the children receiving TANF are also involved with the
child welfare system, a situation that can hinder the ability of case
managers to devise appropriate service strategies for such families.18
Appendix II provides more detailed information from our survey of 15 states
on the extent to which the automated systems for different programs share
data.

In the absence of links between automated systems, local officials in our
study states generally said that paper forms or telephone contacts are used
to refer recipients to, or obtain information on their use of, vocational
education, secondary education, substance abuse services, and mental health
services. The reliance on paper forms was cited as a major burden for case
managers because of the substantial amount of time involved in collecting
all the needed forms from service providers and keying data from these forms
into automated systems. At many local sites, the problem was reported to be
compounded by the need for double or even triple data entry for some items:
case managers or other frontline workers must separately input the same
data, such as a recipient's entry into employment, into different automated
systems because the data are not automatically transferred and updated from
one system to the other. Local officials told us that multiple entry of the
same data not only reduces the time available to work directly with TANF
clients but also increases the risk of introducing errors into the data
contained in automated systems.

An important dimension of automated support for case management is the
extent to which TANF case managers are able to access data from the
automated systems for different programs from their desktop computers. As
shown in figure 3, case managers in the localities we surveyed generally
have desktop access to data in the automated systems for food stamps, TANF
work activities, child support enforcement, Medicaid eligibility
determination, and transportation subsidies for TANF recipients. However,
case managers in most localities do not have desktop access to data from
automated systems for welfare-to-work grants, the Job Training Partnership
Act (JTPA), job listings, child welfare programs, vocational rehabilitation,
and subsidized housing. Over half the localities said that case managers
have desktop access to data from automated systems for child care subsidies
and UI. As a result, case managers do not always have ready access to
information that could help them coordinate services for TANF recipients and
monitor their use of these services. In such cases, the relevant information
either is not available to case managers or it takes longer to obtain it
(for example, from hard copy reports from a specific program).

Figure 3: Automated Systems That TANF Case Managers Can Access From Their
Desktop Computers
Notes: We asked respondents whether TANF case managers had access to any of
the data in the automated systems for each of these programs from their
desktop computers. In a few instances, respondents indicated that access
varies within the locality; we categorized each of these responses as a "no"
because at least some case managers in the locality do not have desktop
access to data for a particular program.

Source: Responses to GAO's questionnaire on automated systems from officials
at local sites. In four states, we obtained responses from more than one
locality: only the responses from the locality in the most populated city in
each of these states are presented here.

Our work in the six case study states also uncovered some specific case
management tasks that are not always well supported by automated systems, as
shown in table 6.

Table 6: Capabilities of Automated Systems in Six States to Support Selected
TANF Case Management Activities

 Activity                                     Do automated systems support
                                              the activity?a
 Identifying suitable jobs for TANF clients
 Help TANF clients identify job openings (or
 provide them tools to do this themselves)    Yes
 Help TANF clients identify job openings
 located near public transportation routes
 (or provide them tools to do this            Vary
 themselves)
 Calculate the change in a TANF client's
 total income (including TANF grant and food
 stamps) under different scenarios of wage    Generally do
 levels and hours worked, to show that "work
 pays"
 Monitoring the activities and progress of
 TANF clients
 Determine the number of active TANF cases in
 the case manager's caseload                  Yes
 Determine the number of active cases
 assigned to different types of activities,
 such as education, training, or subsidized   Generally do
 employment
 Determine which activities in a client's
 individual responsibility plan have been     Vary
 completedb
 Identify clients in the caseload who are
 within some specified number of months of
 reaching their lifetime limit for TANF       Vary
 assistance

a"Generally do" means that at least 70 percent (but less than 100 percent)
of respondents said that the activity can be performed using their current
automated systems. "Vary" means that from 31 to 69 percent of respondents
said that the activity can be performed using their current automated
systems. In calculating percentages, we excluded the small number of
instances in which either we did not obtain a response or respondents said
that they did not know.

bAn individual responsibility plan specifies the employment goal and
required activities for a TANF recipient, as well as the services to be
provided to help the recipient achieve the employment goal.

Source: Interviews with officials at local sites in Georgia, New Jersey,
Ohio, Texas, Washington, and Wisconsin.

As a result of separating cash assistance from Medicaid,19 local officials
in five of the six states cited automated system glitches that sometimes
occur in enrolling families in Medicaid or ensuring their continued
enrollment. For example, local officials said that frontline workers on
occasion have to intentionally enter inaccurate data to enable a Medicaid
case to be opened correctly. Local officials also told us that automated
systems sometimes close Medicaid cases that should not be closed or fail to
correctly process cases for transitional Medicaid,20 which requires
frontline workers to do "workarounds" to correct these problems. 21 PRWORA
authorized $500 million to aid states in maintaining Medicaid coverage for
individuals affected by welfare reform, some of which may be used to pay the
costs of modifying Medicaid eligibility systems. However, as of June 1999,
states had submitted claims for only 10 percent of the available funds.22

Information Limit Capabilities for Service Planning

In addition to supporting the activities of TANF case managers, automated
systems can provide aggregate information on the characteristics and service
needs of TANF recipients to help program managers determine the appropriate
services to provide for their TANF caseloads. However, as shown in figure 4,
the 15 states we surveyed vary considerably in their assessments of the
level of automated support available for service planning, and local
officials tended to assess the level of support lower than state officials
did. While state officials from 8 of the 15 states indicated that automated
systems provide all or most of the information on the TANF caseload that
state program managers need for service planning, officials from 6 of the 15
localities responded that systems provide all or most of the information
needed by local program managers.

Figure 4: Proportion of Information Needed for Service Planning That Is
Provided by Automated Systems
Source: Responses to GAO's questionnaire on automated systems.

Some gaps in information on caseload characteristics occur because the
desired data are not contained in automated systems. Other gaps arise
because even though the data are contained in automated systems, these data
are difficult or impossible to extract in a way that answers the particular
question of concern to the program manager. For example, local officials at
one site said that the locality does not have adequate access to data it
enters into the state welfare system and that writing the computer program
needed to extract data generally takes an entire day. Officials at a
locality in another state said that data on the characteristics of TANF
recipients contained in the state's automated systems are often not
available in a format that can be easily manipulated, so obtaining data
depends on the technical expertise of the user.

As shown in table 7, automated systems in our case study states can provide
information on some characteristics of TANF caseloads but not others. The
ability to identify long-term welfare recipients can be useful because they
may possess characteristics that make them harder to serve. When we asked
whether officials could identify their current "hard-to-serve" TANF cases by
identifying cases that include adults who either are long-term welfare
recipients or have multiple barriers to employment, states and localities
generally said that their automated systems provide this capability. For
example, state officials in Georgia and Washington told us that their
automated systems had been programmed to identify recipients who have
received cash assistance for 30 months or more. In comparison with long-term
welfare recipients, adults who have repeatedly cycled on and off welfare may
differ somewhat in their service needs.23 Local officials generally said
that they are unable to identify such cases in their caseloads. Responses
from state officials varied: while some said that they cannot identify such
cases, others said that they can do so with some difficulty.24

Table 7: Capabilities of Automated Systems in Six States to Provide Program
Managers With Information on Selected Caseload Characteristics

                             Can state program       Can local program
 Caseload characteristic     managers obtain this    managers obtain this
                             information from        information from
                             automated systems?a     automated systems?a
 Number of adults in the
 state/local TANF
 caseload…
 With no prior work
 experience                  Generally can           Generally cannot
 Assessed as having substance
 abuse problems              Cannot                  Generally cannot
 Assessed as having mental
 health problems             Cannot                  Generally cannot
 Current state/local TANF
 cases that…
 Include adults who are
 long-term welfare recipients
 or have multiple barriers toGenerally can           Generally can
 employment
 Have cycled on and off
 AFDC/TANF in the state in   Vary                    Generally cannot
 the last 5 years

a"Generally can" means that at least 70 percent (but less than 100 percent)
of respondents said that the activity can be performed using their current
automated systems, whereas "generally cannot" means that 30 percent or less
(but more than 0 percent) responded in this way. "Vary" means that from 31
to 69 percent of respondents said that the activity can be performed using
their current automated systems. In calculating percentages, we excluded the
small number of instances in which either we did not obtain a response or
respondents said that they did not know.

Source: Interviews with state and local officials in Georgia, New Jersey,
Ohio, Texas, Washington, and Wisconsin.

Officials at some localities cited examples of other information gaps they
had encountered with respect to characteristics of their TANF caseloads. For
example, officials at one site explained that while the state's automated
welfare system contains data on whether individual TANF clients have
received high school diplomas, it does not contain data on the highest grade
completed.25 Officials explained that the lack of such information
constrains their ability to purchase appropriate education and training
services for the local TANF caseload.

Another way in which automated systems can support welfare reform is by
providing information for program oversight and, in particular, information
for monitoring measures of program performance and meeting federal TANF
reporting requirements. As shown in figure 5, the 15 states and localities
we surveyed varied considerably in their assessments of the level of
automated support available for monitoring performance measures related to
helping TANF recipients find jobs and become economically independent. Local
officials tended to assess the level of automated support lower than state
officials did. State officials from 10 of the 15 states indicated that
automated systems provide all or most of the information needed by state
program managers, and officials from 6 of the 15 localities responded that
systems provide all or most of the information needed by local program
managers.

Figure 5: Proportion of Information Needed for Monitoring Program
Performance With Respect to Employment Progress of TANF Recipients That Is
Provided by Automated Systems
Source: Responses to GAO's questionnaire on automated systems.

Limited at the Local Than the State Level

With regard to monitoring measures of the employment progress of TANF
clients, state officials in all six of our case study states said that their
states have data on job entries, job retention, and wage progression and
have used these data to apply for the TANF high-performance bonus. The
states vary in the source(s) they use to obtain these state-level data:
state UI data, TANF administrative data, or both.

In contrast, local officials in the six states reported having more limited
capabilities to monitor the employment progress of TANF clients. Local
officials generally can obtain data from automated systems on the number of
TANF clients in a locality who have entered employment in some specified
time period. However, local officials vary in their automated capabilities
to obtain aggregate information on the job retention of these recipients and
generally cannot obtain aggregate information on recipients' wage
progression. Local officials in one state cited the need to have more
detailed information on the employment progress of TANF clients than is
available from state UI records. For example, they noted that while UI
records can provide information on the amount of a person's total earnings
for a quarter, these records cannot provide information on the person's
hourly wage and number of hours worked per week. These officials said that
having access to such information could help localities better target the
appropriate job retention and advancement services for current or former
TANF clients.26

Officials at several local sites also reported problems obtaining
information on the TANF work participation rates for their caseloads. The
problems cited by local officials in this area included not being able to
obtain timely data on work participation rates for their area from the
state, receiving state-calculated rates that do not match the locality's own
calculated figures, and not being able to interpret the state-provided
reports on local participation rates. If local program managers and
contracted service providers are held accountable for meeting specified
target levels for work participation rates and other performance measures,
it becomes critical that automated systems be able to provide data that are
timely, accurate, and clearly presented so that program managers and service
providers can make needed adjustments in their service strategies.

Recipients

Diversion is a central component of many states' welfare reform programs.
The objective of diversion strategies is to meet the needs of potential TANF
recipients in ways other than through monthly cash assistance, such as by
having them engage in immediate job search to obtain employment quickly;
providing one-time cash payments; or providing support services, such as
child care and medical assistance. The ability to obtain aggregate
information in this area can facilitate program oversight by highlighting
the frequency with which potential recipients are diverted for various
reasons.

State officials in the case study states generally said that they have
automated capabilities to determine the number of families that have
received one-time cash diversion payments. Officials in some of these states
reported that their automated systems also provide information on other
types of diversions. For example, Texas' system tracks both the number of
people who leave welfare offices without applying for TANF and the number
who obtain employment through the state's welfare-to-work program but do not
receive cash assistance. While Wisconsin does not have a one-time cash
payment diversion program, the state's automated welfare system tracks the
number of people diverted from TANF for a range of reasons and uses separate
codes to track people who have been screened for TANF eligibility and those
who have not.

Leaving TANF Are Limited in Some Cases

Information on the use of social service programs by families after they
exit TANF can help program managers determine whether families are receiving
services such as Medicaid and food stamps, which can facilitate families'
efforts to retain employment and increase their wages. The recent declines
in the size of the national Medicaid and food stamp caseloads have generated
concerns that some families who leave TANF are not receiving the Medicaid
and food stamps for which they are eligible and that this, in turn, may
jeopardize their employment progress.27 Conversely, information on the use
of social service programs by families after they exit TANF can also help
program managers determine whether families are reducing their dependence on
government-provided benefits over time.

State officials in the case study states generally said that they have
automated capabilities to determine what percentage of families that have
left TANF within some time period are receiving Medicaid or food stamps some
specified number of months after leaving TANF. Capabilities vary at the
local level: some officials said that their automated systems can perform
this task, whereas others said that their systems lack this capability. In
contrast, both state and local officials generally said that they do not
have the capability to determine how many children are placed in foster care
within some specified time period after their families leave TANF. This
information gap limits the ability of program managers to monitor the extent
to which TANF case closures are associated with subsequent financial
hardship that leads to child abuse and neglect.

Ability to Meet Federal TANF Reporting Requirements

At the time of our interviews, our case study states were in the process of
modifying their automated systems to generate data for some of the required
elements in the quarterly TANF Data Report. Data on recipients' education
levels, reasons for case closure, and noncustodial parents were among the
elements cited by states as especially challenging to obtain. Some state
officials commented that a major source of the problem in meeting the
reporting requirements is that considerably more information on individuals
is now needed than was collected under AFDC.

One of the required data elements in the quarterly TANF Data Report, the
number of months countable toward a TANF adult's federal 5-year time limit,
is a primary source of data for enforcing PRWORA's limit on federal TANF
assistance for individuals. We identified some limitations of automated
systems that affect the states' ability to enforce the federal time limit.
State officials in five of the six case study states said that their
automated systems provide data on the number of months countable toward a
TANF recipient's time limit in the state.28 However, when we asked about the
mechanisms states are using to obtain data on prior receipt of TANF in other
states by their TANF recipients, we learned that our states generally either
do not collect such data or rely on TANF clients to disclose this
information. Data that rely on client self-reporting are not likely to be as
reliable as those derived from a database of TANF recipients and could
result in payments to ineligible individuals.

As discussed earlier, PRWORA directed HHS to prepare a report that would
identify requirements for a system capable of tracking participants in
public assistance programs over time and checking case records of the states
to determine whether individuals are participating in the public assistance
programs of more than one state. In its December 1997 report to the
Congress, HHS noted that the tracking provisions present interstate issues
and would require extensive coordination and interstate agreements. HHS
identified five alternative system architectures for meeting the participant
tracking requirements of the law. HHS analyzed these alternatives and
concluded that while some clear distinctions existed in the functional
capabilities of these system architectures, no system was clearly the best
because of uncertainty regarding how some provisions of the law would be
implemented.29 HHS included three options for congressional consideration in
its report.30

USDA has recently completed a study of the feasibility of a national system
for tracking participation in public assistance programs. In August 1998, we
issued a report in which we recommended the development of a national client
database to reduce improper food stamp payments resulting from including
individuals as members of recipient households in more than one state during
the same period.31 In a federally mandated report issued in November 1999,32
USDA concluded that two of the five alternatives identified by HHS in its
1997 report would provide feasible models for a database of public
assistance recipients that could be used for the Food Stamp program and to
track TANF time limits. Nonetheless, the issue of how best to establish such
a database remains unresolved.

State Automated Systems Projects Seek to Expand the Ability to Obtain and
Analyze Data From Multiple Sources

The states we visited are planning or implementing a variety of automated
systems projects to better provide the information needed by frontline
workers and program managers in the new welfare environment. Because links
among a range of program data systems are key to meeting these information
needs, some projects focus on increasing frontline workers' access on their
desktop computers to data on individual recipients across programs. Other
projects, geared more to improving the capabilities of program managers to
obtain and analyze data from different programs, involve developing new
databases and query tools. Modifying and developing automated systems that
better support welfare reform is a long-term and evolving process.

by Expanding Automated System Links

States are using various approaches to make data on individual recipients
stored in multiple systems more readily available to frontline workers to
help them coordinate service delivery. These approaches include developing
links among separate systems, replacing existing systems with new integrated
systems, and constructing electronic networks to link agencies and service
providers. In some states we reviewed, such projects represent the first
time that data previously stored in separate automated systems operated by
different agencies, such as human services and employment services, are
being shared by the agencies and their affiliated service providers. Table 8
provides an overview of the objectives, scope, and implementation status of
the case study states' projects.

Table 8: Automated Systems Projects Being Designed to Help Frontline Workers
Coordinate Service Delivery

Continued

                      Objectives       Scope of program    Implementation
     Project                                data               status
 Georgia
                  Create an on-line,
                  real-time, desktop
                  system for case
                  and program
                  management.

                  Provide automated                      Planning began in
 System for the   support for                            1992 and
 Uniform          determining                            implementation, in
 Calculation and  eligibility,       TANF, JOBS, food    1998. Project
 Consolidation of issuing benefits,  stamps, and         staff are
 Economic Support determining        Medicaid            currently working
 Services         claims, and                            to achieve full
                  generating                             reporting
                  management                             capability.
                  reports.

                  Support caseworker
                  follow-up of
                  former TANF
                  recipients.
                  Create an
                  Internet-based
                  employer database
                  that provides
                  information on
                  jobs, wage levels,
 Employer         benefits, and so
 Information      on.                Employer and job    Being piloted in
 System                              information         five offices
                  Provide automated
                  support for
                  coordinating
                  services for TANF
                  and
                  welfare-to-work
                  recipients.
 New Jersey
                  Create local,
                  Internet-based
                  electronic
                  communication and
                  data-sharing
                  networks to help
                  coordinate
                  services delivered
                  to TANF recipients
                  by public,
                  private, and
                  not-for-profit     Access to, and
                  providers.         linking of,         Being used in 2
                                     multiple statewide  counties; 15 other
 One Ease-E Link  Provide automated  systems in support  counties are in
                  capability for     of TANF, food       various stages of
                  determining        stamps, and child   implementation.
                  eligibility,       support
                  identifying
                  potential service
                  providers, and
                  scheduling
                  recipients'
                  appointments on a
                  real-time basis.

                  Reduce duplicate
                  collection of
                  data.
                  Provide a single,
                  on-line,                               Automated
                  simultaneous                           application
                  "front-end"                            software installed
                                                         statewide in
 Universal        application for    TANF, AFDC,         support of TANF,
 Application      multiple state     Medicaid, food      child support, and
 Process          systems that       stamps, and child   food stamps.
                  provides           support
                  eligibility,                           Additional program
                  calculation, and                       support and
                  benefit delivery                       enhancements are
                  functions.                             being developed.
 Ohio
                  Provide frontline
                  workers with a
                  single entry point
                  (a "common front
                  end") to data on
                  TANF clients in
                  various automated
                  systems.

                  Through Internet
                  link allow
                  matching of
                  clients'
                  qualifications to                      Directory of
                                     TANF, food stamps,
 Integrated       on-line job banks  Medicaid, state     service providers
 Client           and employment     general             being piloted in
 Management       services.          assistance, child   four counties;
 System                              support, and child  other components
                  Eliminate                              of project are
                  redundant data     care                being developed.
                  entry and provide
                  on-line reporting
                  capabilities.

                  Provide tools to
                  support case
                  management, such
                  as a directory of
                  service providers
                  and a geographic
                  information system
                  to provide mapping
                  capability.
 Texas
                  Help coordinate
                  provision of
                  services to all
                  job seekers by
                  providing a single                     Several phases of
                  point of access                        the project have
                  for multiple                           been implemented.
                  programs through                       The next phase
                  use of a "common   TANF, Employment    will incorporate
                  front end."        Services, Food      employment
                                     Stamp Employment
                                                         services
 The Workforce    Provide frontline  and Training        information
 Information      workers with       program, JOBS,      pertaining to job
 System of Texas  history of all     JTPA, child care,   orders,
                  services provided  UI, Supplemental    applicants, and
                  to a client.       Security Income,    job-matching
                                     food stamps, and
                                                         procedures.
                  Provide            child support       Subsequent
                  information for                        enhancements will
                  meeting reporting                      incorporate child
                  requirements.                          care data.

                  Provide standard
                  and customized
                  reports.
                  Improve access to
                  health and human
                  services by
                  creating an
                  integrated system
                  for determining                        In the early
                  eligibility for                        planning stages of
 Texas Integrated various programs.  Health and human    a projected 8-year
 Enrollment                          services, and all   project; the
 Redesign System  Provide a          medical assistance  initial plan was
                  directory of       programs            issued in June
                  community                              1999.
                  services.

                  Provide ad hoc
                  reporting
                  capability.
 Washington
                  Allow ad hoc
                  generation of data
 Jobs Automated   reports by user
 System Ad Hoc    specifications,    TANF work           Implemented in
                  such as cases by
                                     activities          fall 1999
                  demographics,
                  activity status,
                  or program
                  participation.
 Wisconsin
                                     TANF, child care,
                  Provide case       UI, job training,
                  managers with      Medicaid, Food
 Partnership for  desktop access to  Stamp Employment
 Full Employment  data from various  and Training
 Case Management  systems on         program,            Being tested
 System           recipients'        vocational
                  involvement with   rehabilitation,
                  various programs.  new hires, and
                                     transportation.

Source: GAO interviews in case study states and analysis of written
materials about the projects.

These projects are intended to enhance the capabilities of frontline workers
in such ways as the following.

ï¿½ Ohio's Integrated Client Management System is expected to enable frontline
workers in the state to use their desktop computers to obtain a holistic
view of TANF recipients' and their families' use of social services, their
service needs, and their strengths. The project is also expected to enable
frontline workers to input data only once in a system and then have them
automatically transferred to other appropriate systems, thereby helping
reduce errors that arise from duplicate data entry.

ï¿½ Wisconsin's Partnership for Full Employment Case Management System is
intended to provide frontline workers with read-only access to data in
various state systems to enable them to determine a TANF recipient's
involvement with different programs. For example, if a recipient has a
record on the state's database of new hires, the system should display the
hiring date and the address of the employer.

ï¿½ Georgia is using its System for the Uniform Calculation and Consolidation
of Economic Support Services to facilitate having a single frontline worker
handle an applicant's involvement with TANF, food stamps, and Medicaid,
whereas in the past the applicant would have been referred to three
different eligibility workers.

ï¿½ The New Jersey Departments of Health and Senior Services, Human Services,
and Labor are providing leadership to counties that want to develop an
on-line local electronic communication network. A frontline worker in one
county piloting the One Ease-E Link network told a state official that she
is now referring recipients to services "that she never even knew existed
before," such as assistance in paying for needed medicines.

Capabilities to Analyze Data Across Programs

In the wake of welfare reform, program managers have a greater need to
obtain and analyze data from multiple programs to support their
responsibilities for service planning and program oversight. To meet this
need, states are developing data query tools and databases of current and
historical data from multiple programs. States are working to expand their
capabilities for analysis beyond the information contained in preprogrammed
monthly reports by developing query tools that enable users to perform
on-line queries and generate customized reports that meet their particular
information needs. In addition, states are extracting and consolidating data
from multiple systems in data warehouses, data marts, and other specialized
databases to which more sophisticated query tools can be applied. Table 9
provides an overview of projects in the states we reviewed.

Table 9: Automated Systems Projects Being Designed to Enhance Analytical
Capabilities of Program Managers

                       Objectives       Scope of program    Implementation
     Project                                  data             status
 Ohio
                  Provide a single
                  repository of all
                  transactions data
                  from the Department
                  of Human Services
                  and 5 years of
                  historical data.     TANF, Medicaid,
                                       food stamps, child
 Data warehousea  Provide              care, child        Under development
                  instantaneous        support, and child
                  responses to         welfare
                  queries.

                  Provide data needed
                  to meet federal TANF
                  reporting
                  requirements.
 Texas
                                       TANF, AFDC,
                  Support management   employment,
                  in program and       training, child
                  performance          care, labor market
                  evaluation.          information, food
                                       stamps, JOBS,
                  Develop a repository human resources,
 Data warehouse   of data from eight   North American     In use
                  data marts that is   Free Trade
                  accessible from the  Agreement,b UI,
                  Internet.            wages, child
                                       support, and Food
                  Provide 8 years of   Stamp Employment
                  labor market data.   and Training
                                       program
 Washington
                  Provide
                  up-to-the-minute
                  data on TANF
                  recipients.          TANF eligibility
 Data warehouse                        and work           Implemented in
                  Support performance  activities         1997
                  monitoring and ad
                  hoc report
                  generation.
                                                          Constructed and
                                                          available to
                  Support                                 state
                  user-specified data                     headquarters
 Data martc       queries and provide  Employment         staff who are
                  capability to view   services           being trained
                  reports on-line.                        now; rollout to
                                                          field staff will
                                                          follow.
                                       TANF eligibility
 Bar Code         Provide capability   and work           Implemented
                  for ad hoc queries.  activities, and
                                       child care
 Wisconsin
                  Provide a more
                  flexible reporting
                  environment and
                  generate faster
                  responses to
                  requests for data
                  and reports through
 Wisconsin Data   Internet access.     TANF, child care,  Under
 for                                   child support,     development; some
 Organizational   Provide ability to   Medicaid, and food components in
 Management       display data by      stamps             use.
                  individual case
                  managers in local
                  TANF agencies.

                  Decrease drain on
                  mainframe caused by
                  data queries.
                  Create an
                  integrated,
                  longitudinal         TANF, Medicaid,
                  database to support  food stamps,
 Wisconsin Policy management,          wages, criminal
 and              evaluation, and      justice, UI,
 Administrative   research.            vocational         Under development
 Data                                  rehabilitation,
                  Provide information  child support, and
                  for monitoring the   child welfare
                  status of former
                  TANF recipients.

aA data warehouse is a massive database that integrates information
collected from disparate sources. Data warehouses are separate from the
systems used for daily business operations and are usually dedicated to
management decision support.

bThe Department of Labor implemented a program in 1994 to assist workers who
lose their jobs as a result of the North American Free Trade Agreement.

cA data mart is a smaller version of a data warehouse that usually
incorporates data from fewer sources and is designed to meet a specific
business need or problem.

Source: GAO interviews in case study states and analysis of written
materials about the projects.

The following examples illustrate how states expect their projects to help
managers.

ï¿½ Texas' data warehouse is expected to help managers assess program results
and the performance of service providers by providing the capability to
answer various questions, such as how many participants in different
training programs entered employment within 2 months and what their average
wage levels were.

ï¿½ The Wisconsin Policy and Administrative Data project seeks to help track
former TANF recipients' progress toward economic independence by providing
information such as their post-TANF earnings as compared with the earnings
of other low-wage workers, their job entry and retention rates, and their
use of other support programs.

The extent to which the large databases under development will effectively
and efficiently meet information needs for welfare reform remains to be
determined. According to a May 1998 report of the National Association of
State Information Resource Executives, "Practically every state has
developed, is designing or is planning to build a data warehouse to
implement welfare reform."33 However, establishing a data warehouse is a
complex and potentially costly undertaking. Data must be reformatted to use
standard definitions and conventions across programs and checked for missing
or erroneous entries.34 Nonetheless, these automated systems projects hold
promise for expanding the amount of information available to program
managers.

Federal Action Could Help Overcome Obstacles States Face in Improving
Automated Systems

States face a number of obstacles to improving their automated systems, such
as the magnitude of changes due to welfare reform, the inherent difficulties
involved in successfully managing information technology projects,
competition with the private sector to recruit and retain information
technology staff, and the complexity of obtaining federal approval and
funding for systems projects that involve multiple agencies. The federal
government could take actions to facilitate states' automation efforts by
providing more information to states on best practices for managing
information technology, reviewing and modifying as needed the federal
approval and funding process for automated systems, facilitating links among
state automated systems, and addressing the need for information to track
TANF time limits across state lines. Currently, no group or organization
that brings together key federal agencies involved in welfare reform has
been formally charged with developing solutions to the range of obstacles
states face in improving their automated systems. To encourage such federal
efforts, we recommend that HHS establish an interagency group that would
meet such a challenge.

Systems

Experience shows that developing new automated systems or modifying existing
systems to meet current needs can be a complex and difficult undertaking.
This is certainly true in the current environment, in which states face
significant obstacles to effectively and efficiently managing complex
information systems projects.

and Modification

The large-scale changes in the mission and operations of welfare agencies
have complicated states' efforts to define the functional and data
requirements for automated systems projects.35 As the roles of frontline
workers and program managers have expanded beyond eligibility determination
to include a focus on employment, welfare agencies have taken and continue
to take steps to change the way they do business. Washington state officials
cited the large-scale changes in welfare reform as a major obstacle to
systems development. One state official commented that planning systems
modifications under these conditions is like "building the plane while you
are flying"--trying to be responsive to customer needs before the customer's
needs are fully understood.

Many state and local agencies have established new collaborative efforts
with other agencies and community organizations to provide services to TANF
clients, which creates new needs for information sharing among these
partners. The added complexity of working with many partners can exacerbate
systems development efforts, as has occurred in Georgia and Washington.
Georgia's welfare reform program involves three primary partners at the
state level: the labor, welfare, and adult education agencies. The ability
of these agencies to develop automated systems to meet their overlapping
information needs for welfare reform has been hampered by significant
differences in agencies' program priorities, systems architectures, and
technological capabilities. For example, the labor agency has its own
mainframe system, whereas the welfare agency uses a mainframe system
maintained by the Department of Administrative Services and is therefore
dependent on the cooperation of this department to make any needed systems
changes.

Officials in Washington also reported obstacles in this area, noting that
the complexity of automated systems projects typically increases
substantially as more agencies are involved, and the project scope tends to
expand to satisfy the needs of all the agencies. Some states have made
organizational realignments that helped to ease such systems development
issues. For example, the governor of Wisconsin merged the state labor and
welfare agencies into a single department in 1998, and Wisconsin state
officials told us that the merger had facilitated their systems development
efforts.36

Furthermore, some states have devolved substantial authority to their
localities for designing TANF programs, which means that state automated
systems will be called upon to support a potentially more diverse range of
local program goals and operations. For example, Ohio cited the devolution
of its TANF program as a major obstacle in designing its Integrated Client
Management System. Ohio state officials told us that while the state's focus
in developing this system is to support local needs, it has been very
challenging to actually do so because of the substantial diversity of
operations among the state's 88 counties.37

In addition to the large-scale changes in the mission and operation of
welfare agencies, the expanded federal reporting requirements have
complicated for states the task of defining their own systems needs. Several
states we visited or surveyed noted that the substantial investment in
resources required to modify systems to meet the expanded federal
requirements constrained their own efforts to use or improve their systems
to better implement and oversee their welfare reform programs. In our
earlier work on states' automated systems for JOBS, the previous
welfare-to-work program for AFDC recipients, we found that states focused
system design solely on meeting federal data collection and reporting
requirements rather than on other program objectives, such as providing
information for the use of frontline workers in helping welfare recipients
find employment.38

In addition, welfare reform has created a need for states to have access to
cross-state information on individuals' TANF receipt to enable enforcement
of the 5-year TANF limit. As previously discussed, the states we reviewed
generally do not have access to such information.

Finally, a long-standing obstacle to developing automated systems that
support service delivery to low-income families is the multitude of programs
operated by several different federal agencies that provide services to this
population. We reported in 1995 that the complexity of the system of aid for
low-income families, including diverse and sometimes contradictory program
requirements and separate funding provisions, has made it difficult for
states to develop integrated, streamlined automated information systems,
often leading states to develop essentially separate automated systems for
each program.39 States continue to cite these conflicting program
requirements as obstacles to developing systems that meet information needs
that cross program and agency boundaries, which has become more critical
than ever as a result of the broad objectives of welfare reform. For
example, Texas officials said that the variety and complexity of eligibility
requirements for different federal programs have presented a major challenge
for the state's plan to design an integrated system intended to improve the
efficient completion of eligibility processes.40 In addition, Wisconsin
officials commented that federal TANF and welfare-to-work grant reporting
requirements use different definitions for some key terms, such as "closed
case," which complicates systems development.

Successfully managing information technology projects is difficult, and many
projects fail. Our fieldwork produced examples of difficulties states
encountered in effectively managing information technology projects. New
Jersey officials said that in planning for the development of systems to
support their welfare reforms, they had to decide whether to maintain and
modify existing systems or replace them using new technologies, which would
require identifying the most appropriate technological alternatives. As a
result of such challenges, it took about 4 years for the state to develop
its plans for automated systems initiatives. State officials in Ohio decided
to delay the rollout of its new system in order to shift from a client
server technology, as originally planned, to an Internet-based technology,
in large part to take advantage of the greater opportunities for information
access provided by the latter.

The HHS study of state automated systems mandated by PRWORA found that the
automated systems that were being used for TANF in 66 percent of the states
in 1997 first became operational in the 1970s or 1980s.41 New types of
systems hardware have been developed since the initial state welfare
automated systems were developed using mainframe architectures in the 1970s.
Client server architectures were developed to provide end users with greater
capabilities to manipulate data in a system. Additionally, in the last
several years, Internet-based architectures have been developed that allow
frontline workers and managers to access data from several separate systems
at one time. Through this new Internet capability, states can build systems
that better promote integrated service delivery while still drawing on the
capacities of existing systems. The need to update both hardware and
software is especially pronounced in the welfare arena because many states
are using older automated systems to manage their welfare programs.

Y2K Preparations and Competition With the Private Sector for Staff

Obtaining sufficient staff resources for automated systems projects is
another area in which states have encountered obstacles. Two principal
obstacles cited by states were the substantial staff resources diverted to
address the Y2K issue and difficulties recruiting and retaining qualified
information technology staff. The Y2K issue had a sharply defined deadline
and the potential for major service disruptions, which helps account for the
high priority states placed on modifying systems to address this issue.
States told us that this priority constrained their abilities to obtain
information technology staff to undertake new automated systems projects for
welfare reform.

In addition, states have encountered long-standing problems in recruiting
and retaining information technology staff. States reported that these
problems are due to factors such as the relatively lower pay and benefits of
information technology jobs in the public versus the private sector and
fewer opportunities in the public sector to work with the newest hardware
and software. Several states noted that the loss of some information
technology staff to the private sector had significantly diminished their
organizational expertise in the area of information technology. Child
support enforcement officials in Texas told us that about 80 percent of
information technology personnel, such as systems analysts and programmers,
left state government jobs to join various firms that contract with the
child support enforcement program and other program areas.42 According to
state officials, the loss of these personnel has resulted in poor or reduced
service to the public, because without timely upgrades to automated systems,
program personnel cannot easily access case information, update files, or
respond to customer inquiries.

Obstacles

States reported that the federal procedure that they must follow for systems
procurement is a major obstacle in obtaining federal funding for their
automated systems projects. As discussed earlier, to promote accountability
for the use of federal funds, the federal government requires that states
develop an APD to obtain federal funding for automated systems projects
related to food stamps, Medicaid, or child support enforcement.43 While
these acquisition procedures applied to AFDC, they do not apply to systems
supported with TANF funds only. This reduced federal role in TANF-funded
systems is in keeping with the general devolution of responsibility for TANF
programs and operations to states. PRWORA stated that HHS may not regulate
state-managed TANF programs except in those areas specifically authorized by
PRWORA. PRWORA also specified reduced HHS staffing levels for the
administration of TANF. Policy devolution and reduced staffing have altered
HHS' oversight roles and responsibilities, particularly in the area of
automated systems.

States criticized the APD process as being too cumbersome with respect to
designing systems that require approvals from multiple federal agencies. For
example, Ohio state officials told us that they decided to develop their new
case management system using TANF funds only--even though they would have
preferred to use additional sources of federal funds--because of a desire to
avoid the APD process for such reasons. New Jersey state officials commented
that even though federal agencies encourage the development of integrated
systems, the narrow programmatic focus of federal agencies in the APD cost
allocation process makes this difficult. The American Public Human Services
Association (APHSA) has echoed such concerns, maintaining that federal
procurement procedures, including the APD approval process, are not
structurally capable of quickly and effectively responding to the rapidly
changing business and technological environments confronting welfare
reform.44

Previous collaborative efforts of federal and state officials have had some
success in improving the APD process. In response to our 1992 recommendation
that ACF, HCFA, and USDA collaborate when approving states' system
proposals, an Information Technology Partnership Project was established
that year to address identified problems with the approval process.45 The
project involved a series of meetings of representatives from HHS, USDA, the
American Public Welfare Association,46 and the National Association of State
Information Resource Executives. In 1994, the project developed a list of
action items for improving the APD process. USDA officials told us that the
collaborative efforts of USDA and HHS contributed to improvements in the APD
approval process after the project ended. For example, they noted that
regulations issued jointly by USDA and HHS in 1996 provided regulatory
relief by raising the thresholds that trigger APD reviews. USDA officials
also pointed out that all federal partner agencies agreed upon an expedited
approval process for state systems approvals related to Y2K. While all of
the short-term action items identified by the Information Technology
Partnership Project have been implemented, several items that were
designated for long-term action have not been implemented and remain open
issues, according to HHS officials.47

The federal government's primary role in the area of state development of
automated systems for social programs has historically been that of a
regulator, focusing on ensuring compliance with applicable federal statutes
and regulations. While PRWORA reduced the role of HHS in overseeing systems
funded solely with TANF funds, HHS still plays a key role, along with USDA
and Labor, in funding and overseeing states' information systems for social
programs. Because of this key role, the federal government could further
enhance the progress of welfare reform by also functioning as a facilitator
of states' automated system initiatives. On the basis of our visits to six
states, surveys of an additional nine states, a review of numerous reports
on these issues, and consultation with experts participating in the
GAO/Rockefeller Institute Working Seminar on Social Program Information
Systems, we identified four key areas in which federal actions could
facilitate states' efforts.

ï¿½ Disseminate information on best practices for managing information
technology, particularly in the area of welfare reform. Without assuming
greater responsibility for managing states' information technology projects,
the federal government could do more to provide information on and increase
states' access to expertise in that area.48 For example, the government
could serve as a clearinghouse for information on best practices in public
and private organizations and sponsor conferences that disseminate
information on successfully managing information technology projects. In
addition, it could identify and disseminate information on state best
practices in developing or modifying automated systems to support the new
welfare reforms.

ï¿½ Review, and modify as needed, the APD process to ensure that it meets
federal needs for state accountability without unnecessarily hindering state
development efforts. Although the APD is no longer required for systems
developed solely with TANF funds, the process is still key to states'
efforts to integrate or connect TANF, food stamps, and Medicaid systems. The
federal process for APD could be reviewed to ensure that it supports what is
currently known about best practices in systems development. In addition,
the key departments and agencies involved--USDA and HHS' ACF and HCFA--could
work more closely in the APD review and approval process, as is warranted by
the need to approve plans for systems projects that involve multiple
programs, and in keeping with our recommendation about this in our 1992
report.

ï¿½ Play a stronger role in facilitating links among the automated systems
used by different state and local agencies through such means as the
following.

ï¿½ Serve as a clearinghouse for information on strategies being used to
develop connections among automated systems for different programs,
particularly in new areas such as vocational education and criminal justice.

ï¿½ Support demonstrations designed specifically to promote better
partnerships between state and local agencies, which lead to more effective
and efficient performance of automated systems.

ï¿½ Coordinate existing and new data collection and reporting requirements
developed by federal agencies for different programs to encourage common
definitions and standards where feasible.

ï¿½ Identify conflicting eligibility requirements for federal programs that
serve similar populations. Federal agencies could work together to identify
and revise those eligibility requirements determined at the agency level
rather than in legislation. In addition, agencies could present to the
Congress proposals for changing selected conflicting eligibility
requirements in order to better facilitate integrated service delivery while
still meeting congressional objectives.

ï¿½ Address the need for states to have access to cross-state information on
individuals' TANF receipt so the states can enforce the federal 5-year TANF
time limit. In its December 1997 report to the Congress, HHS presented some
options for a national tracking system for congressional consideration and
also noted that more information about how states implemented welfare
reforms would be needed before considering the most appropriate options for
tracking the time limit nationwide. A November 1999 report by USDA studied
options for a national system to detect certain types of food stamp fraud
and time limits that had been exceeded as well as to track recipients' time
on TANF. TANF recipients could begin to reach the 5-year time limit
established by PRWORA as early as 2001. In addition to establishing this
time limit, PRWORA required HHS to impose a financial penalty on states that
provide federal TANF aid to families no longer eligible because of the time
limit, and HHS will need accurate information to enforce this penalty.49 HHS
could take the lead in addressing the need for a national system and work
with the Congress to the extent that legislation and resources are needed to
accomplish that goal. HHS could also work with USDA as that agency considers
a national system to help it identify individuals who are illegally
receiving food stamps.

The current environment provides a window of opportunity for more
coordinated federal attention to automated systems for social programs.
Because states have reported that their automated systems have successfully
made the transition to the new century, many state and local staff resources
will no longer be diverted to Y2K-related work. Increased attention to the
ongoing and evolving process of improving states' automated systems for
social programs could ultimately help bring about more effective and
efficient service delivery for low-income families. More specifically,
improved systems would help ensure that the intended goals and requirements
of TANF, such as promoting work and enforcing the 5-year time limit on aid,
are met.

Many of the obstacles states face may be best addressed at the state level,
such as the need for collaboration across state and local agencies. However,
coordinated action at the federal level would help address several of the
obstacles encountered by states as they take steps to improve their
automated systems for social programs. Because the issues that need to be
addressed involve several federal programs and agencies, any actions
undertaken must involve key federal agencies. However, currently, no group
or organization formally brings together the array of federal agencies
involved in welfare reform to help devise solutions to the issues facing
states in improving their welfare and welfare-related automated systems. HHS
could play a pivotal role in orchestrating such a broad-based collaborative
approach by bringing representatives of key federal agencies and other
organizations together to work on these issues.

We recommend that the Secretary of HHS establish an interagency group to
identify, and develop implementation plans for, federal actions that would
facilitate states' efforts to improve their automated systems for federal
programs that serve low-income families. The group should include high-level
federal officials from HHS, USDA, Labor, and other federal agencies as
appropriate. The interagency group should obtain input from state and local
social program and information technology managers; the group could also
seek input from others, such as state organizations, professional
organizations of social program officials, and welfare researchers. In
addition to any actions that the interagency group identifies on its own, it
should consider actions in the following areas.

ï¿½ Disseminating information on best practices for managing information
technology generally and best practices specific to automated systems that
support welfare reform.

ï¿½ Reviewing, and modifying as needed, the federal process for systems
procurement to ensure that it meets federal needs for state accountability
without unnecessarily hindering state development efforts.

ï¿½ Facilitating links among the automated systems used by different state and
local agencies through such means as supporting demonstrations designed to
promote better partnerships between state and local agencies and
coordinating data reporting requirements for different federal programs.

ï¿½ Addressing the need for states to have access to cross-state information
on individuals' TANF receipt to enforce the 5-year TANF time limit.

We obtained comments on a draft of this report from HHS, USDA, and Labor.
USDA and Labor agreed with our recommendation. HHS said that it would, along
with other involved federal agencies, undertake a review of, and modify as
needed, the federal process for systems procurement. HHS said that the
objective of the review would be to ensure that federal needs for state
accountability are met without unnecessarily hindering state development
efforts. HHS did not specifically say that it would take responsibility for
establishing an interagency group to accomplish this review, as we
recommended. We continue to believe that HHS, as the agency responsible for
many of the programs, should take the initiative in establishing an
interagency group. In agreeing with the report recommendation, Labor
suggested that it might make sense to place the recommended interagency
group under the direction of the Domestic Policy Council because of the
cross-cutting nature of providing services to low-income families. This is
another way that our recommendation could be implemented.

In its comments, HHS emphasized the initiatives it has taken or has under
way to help states achieve effective information systems that support the
transition of families from welfare to work. HHS also emphasized the
information systems procurement process when discussing what it would review
with other agencies. We believe that the interagency group should, as
outlined in the recommendation, address multiple issues, not just the
computer systems procurement process.

USDA commented that the report did not sufficiently recognize
accomplishments since 1992 with respect to federal streamlining of the APD
approval process. We added information on these accomplishments to the
report. USDA also commented that effective implementation of the report's
recommendation would require additional funding and staff resources and that
USDA's Food and Nutrition Service has not been successful in its efforts to
obtain additional resources, including federal employees and contractor
support, to bolster the existing APD process. If USDA believes that
additional resources would be needed to implement components of our
recommendation, it should request and justify additional resources in its
budget proposal.

Comments from HHS, USDA, and Labor appear in appendixes III, IV, and V,
respectively. We also obtained technical comments from USDA and the six case
study states and incorporated them in the report as appropriate.

The GAO/Rockefeller Institute Working Seminar on Social Program Information
Systems: Objectives, History, and Membership

The capability to obtain the information needed to implement and oversee
welfare reform is a critical element for its success. To assist
congressional oversight and provide expertise to inform our work in the area
of automated information systems, GAO jointly established the working
seminar with the Nelson A. Rockefeller Institute of Government in March
1998. The federal government, as a major funder of automated systems for
social programs, has a major stake in the development and operation of these
systems. The primary objectives of the working seminar are to (1) determine
the overall directions in which state automated systems must move to better
meet the information needs of welfare reform, (2) identify changes needed at
the federal level to facilitate movement in these directions, and (3)
disseminate these findings to stimulate action to implement such changes.

The working seminar met six times between March 1998 and November 1999. In
the initial three meetings, members worked with GAO staff to articulate some
conceptual categories and methodological approaches to help structure our
study of the subject matter. The fourth meeting, in March 1999, featured
views from officials from Ohio, Texas, and Wisconsin on the capabilities of
their automated systems, their efforts to improve these systems, and the
obstacles they have encountered. The June 1999 meeting focused on federal
obstacles that can impede states' progress in developing integrated
information systems for social programs. The meeting sought to identify the
sources of some of these obstacles and explore how they might be addressed.
At the November 1999 meeting, we presented preliminary findings based on the
fieldwork we had completed on this job in collaboration with field
researchers from the Rockefeller Institute's State Capacity Project and
obtained feedback from members of the working seminar.

Program Information Systems

Patrick Babcock, W.K. Kellogg Foundation

Rebecca Blank, University of Michigan

Rachel Block, Health Care Financing Administration, U.S. Department of
Health and Human Services

Catherine Born, University of Maryland

Henry Brady, University of California, Berkeley

Brett Brown, Child Trends

John Cuddy, Oregon Department of Human Services

Cynthia Fagnoni, GAO (cochair)

John Thomas Flynn, Litton PRC

Thomas Gais, Rockefeller Institute of Government

Susan Golonka, National Governors' Association

Richard Hardin, Employment and Training Administration, U.S. Department of
Labor

Ronald Haskins, Committee on Ways and Means, U.S. House of Representatives

John Hurd, Ohio Department of Human Services

Anil Kakani, U.S. Office of Management and Budget

Andrea Kane, White House Domestic Council

Joseph Leo, U.S. Department of Agriculture

Jan Lilja, Food and Nutrition Service, U.S. Department of Agriculture

Terrence Maxwell, Rockefeller Institute of Government

Lawrence Mead, New York University

Ronald Mincy, The Ford Foundation

Richard Nathan, Rockefeller Institute of Government (cochair)

Mark Ragan, Administration for Children and Families, U.S. Department of
Health and Human Services

Howard Rolston, Administration for Children and Families, U.S. Department of
Health and Human Services

Elaine Ryan, American Public Human Services Association

Stephanie Shipp, U.S. Census Bureau

Larry Singer, Public Interest Breakthroughs, Inc.

Douglas Steiger, Finance Committee, U.S. Senate

Sheri Steisel, National Conference of State Legislatures

William Waldman, American Public Human Services Association

Joel Willemssen, GAO

Michael Wiseman, The Urban Institute

Links Among Automated Systems for Different Social Programs

Links among automated systems can provide frontline workers and program
managers access to a broader range of information on TANF recipients'
involvement with different social programs and thereby facilitate the
implementation of welfare reform. The extent to which states have
established links among automated systems for different programs varies
substantially. In the 15 states we surveyed, the systems that support TANF
eligibility determination are, in almost all cases, linked with the
automated systems for food stamps, child support enforcement, TANF work
activities, Medicaid eligibility determination, and transportation subsidies
for TANF recipients, as shown in figure 6.50 These links reflect federal
mandates and enhanced federal funding for automated system links in these
programs.51 In contrast, automated systems for other services that TANF
recipients may need in order to facilitate their movement toward employment,
such as child care subsidies, job training through JTPA, welfare-to-work
grant services, vocational rehabilitation, job listings, and subsidized
housing, are generally not linked to systems for determining TANF
eligibility.52

Figure 6: Programs With an Automated System That Either Is the Same as, or
Shares Data With, the System Used for Determining TANF Eligibility

Note: We asked state officials to indicate whether the automated system used
for each of the specified programs is (1) the same system as used for
determining TANF eligibility, (2) a separate system that is linked to the
system used for TANF eligibility, or (3) a separate system that is not
linked to the system used for TANF eligibility (that is, the systems do not
share data). This figure shows the number of states that responded either
"1" or "2" for each of the programs; thus, lower bars indicate a larger
number of states in which a program is supported by a separate system that
does not share data with the system for determining TANF eligibility.

a One state did not respond to the question on unemployment insurance.

b One state did not respond, and another did not receive welfare-to-work
grants.

Source: Responses from state officials to GAO's questionnaire.

Comments From the Department of Health and Human Services

Comments From the Department of Agriculture

Comments From the Department of Labor

GAO Contacts and Staff Acknowledgments

Gale C. Harris, (202) 512-7235, [email protected]

Andrew Sherrill, (202) 512-7252, [email protected]

The following people also made important contributions to this report:
Suzanne M. Lofhjelm; Catherine V. Pardee; Mark E. Ward; and Michael J. Rich,
Emory University.

Glossary

A specific application that provides users a single access point to multiple
databases without making separate queries to them; instead, it appears to
users that there is only one database. With a common front end, data can be
entered once and updated across all systems. Common front ends are often
built to assist caseworkers and are used with systems containing data
warehouses or other arrangements of multiple integrated databases.

A database that integrates information collected from disparate sources and
is designed to meet a specific business need or problem.

A larger version of a data mart that usually incorporates data from more
sources. Data warehouses are subject-oriented and separate from databases
used for daily business operations. They are usually dedicated to management
decision support and are accessed through report writers, query tools, and
data access and retrieval tools. All data are in a consistent format, exist
for a particular point in time, and do not change. Some warehouses store
data for points in time spanning several years.

A computer-based system that enables users to identify social services
listed in resource directories on the basis of proximity to a client's
residence. The location of a client's residence, the service provider, and
local geographic information are mapped and displayed in a "point-and-click"
link.

An industry term for a large computer that is designed for the most
intensive computation. Mainframe computers are often shared by multiple
users connected to the computer via terminals.

A direct connection between the user's terminal and the computer containing
the data. An on-line system is one in which data to be input enter the
computer directly from the point of origin, and the output data are
transmitted directly to the location where they are to be used.

A system that processes transactions as they occur rather than batching
them; response to input is fast enough to affect subsequent inputs and guide
a process control system or a computer-assisted instruction system. For
processing to be real-time it must also be on-line.

Bibliography

American Public Human Services Association. Tools for Information Systems
Reform. Washington, D.C.: APHSA, distributed Dec. 1998.

American Public Welfare Association. Statewide Automated Child Welfare
Information Systems: Survey of State Progress. Washington, D.C.: APWA, July
1997.

____. A Shock to the Systems: Automating Welfare Reform in the States.
Washington, D.C.: APWA, Sept. 1997.

Brown, Amy. "Management Information Systems--Section 19" in Work First: How
to Implement an Employment-Focused Approach to Welfare Reform. New York,
N.Y.: Manpower Demonstration Research Corporation, Mar. 1997.

Center for Technology in Government. Tying a Sensible Knot: A Practical
Guide to State-Local Information Systems. Albany, N.Y.: State University of
New York, 1997. http://www.ctg.albany.edu/resources/pdfrpwp/iis1.pdf (cited
Feb. 28, 2000).

Elwood, Deborah A., and Donald J. Boyd. Changes in State Spending on Social
Services Since the Implementation of Welfare Reform: A Preliminary Report.
Albany, N.Y.: Nelson A. Rockefeller Institute of Government, Feb. 2000.

Gais, Thomas L. "Welfare, Information, and the New Federalism." The
Evaluation Exchange, Harvard Family Research Project, Vol. IV, No. 2 (1998).

Kaplan, April. "Linking the Systems: The Relationship Between Welfare and
Child Welfare Systems." Resources, Welfare Information Network, Vol. 2, No.
4 (July 1998).

Maxfield, Myles, James Ohls, and William S. Borden. Options for a National
Database to Track Participation in Federal Means-Tested Public Assistance
Programs: Report to Congress. Report prepared by Mathematica Policy
Research, Inc., for the U.S. Department of Agriculture, Food and Nutrition
Service. Washington, D.C.: USDA, Nov. 1999. http://www.fns.usda.gov/oane
/MENU/Published/FSP/FILES/dbaseopt.pdf (cited Feb. 28, 2000).

Maxwell, Terrence A. Welfare Reform and Information Management: Rewiring the
Human Service Systems. Albany, N.Y.: Nelson A. Rockefeller Institute of
Government, Dec. 1997.

_____. "Information Federalism: History of Welfare Information Systems."
Working Paper. Albany, N.Y.: Nelson A. Rockefeller Institute of Government,
1999.

Nathan, Richard P., and Thomas L. Gais. Implementing the Personal
Responsibility Act of 1996: A First Look. Albany, N.Y.: Nelson A.
Rockefeller Institute of Government, 1999.

National Association of State Information Resource Executives. Welfare
Reform and State Human Service Information Systems. Lexington, Ky.: NASIRE,
May 1998.

_____. State Human Service Information Systems: Measuring the Impact of
Welfare Reform. Lexington, Ky.: NASIRE, Feb. 2000.

Nelson A. Rockefeller Institute of Government. "Symposium: Implementing
Welfare Reform: A First Look." Rockefeller Institute Bulletin (1999).

Northwestern University/University of Chicago, Joint Center for Poverty,
Advisory Panel on Research Uses of Administrative Data. Administrative Data
for Policy-Relevant Research: Assessment of Current Utility and
Recommendations for Development. Chicago, Ill.: Northwestern
University/University of Chicago, Jan. 1998.

Public Interest Breakthroughs, Inc. California Welfare Reform Impact
Assessment Project Report. Vienna, Va.: Public Interest Breakthroughs, Inc.,
June 5, 1998.

Singer, Larry. "The Role of IT in Welfare Reform." Government Technology
(May 1997).

_____. "States Need Federal Partner." Government Technology (Apr. 1998).

UC DATA, University of California at Berkeley. An Inventory of Research Uses
of Administrative Data in Social Services Programs in the United
States--1998. Report to the Northwestern University/University of Chicago
Joint Center for Poverty Research. Berkeley, Calif.: University of
California, Feb. 1, 1999.

U.S. Department of Health and Human Services, Administration for Children
and Families, Office of State Systems. Report to Congress on Data Processing
and Case Tracking in the Temporary Assistance for Needy Families Program.
Washington, D.C.: HHS, Dec. 1997. http://www.acf.dhhs.
gov/programs/oss/report.htm (cited Feb. 28, 2000).

Wiseman, Michael. A Management Information Model for New-Style Public
Assistance (99-10). Washington, D.C.: The Urban Institute, Aug. 1999.
http://newfederalism.urban.org/html/discussion99-10.html (cited Feb. 28,
2000).

Yates, Jessica. "Information Systems and Tools to Support Goals of
Self-Sufficiency." Issue Notes, Welfare Information Network, Vol. 2, No. 11
(July 1998).

Related GAO Products

Food Stamp Program: Various Factors Have Led to Declining Participation
(GAO/RCED-99-185, July 2, 1999 ).

Welfare Reform: Information on Former Recipients' Status (GAO/HEHS-99-48,
Apr. 28, 1999 ).

Welfare Reform: States' Experiences in Providing Employment Assistance to
TANF Clients (GAO/HEHS-99-22, Feb. 26, 1999 ).

Welfare Reform: Status of Awards and Selected States' Use of Welfare-to-Work
Grants (GAO/HEHS-99-40, Feb. 5, 1999 ).

Year 2000 Computing Crisis: Readiness of State Automated Systems to Support
Federal Welfare Programs (GAO/AIMD-99-28, Nov. 6, 1998 ).

Welfare Reform: States Are Restructuring Programs to Reduce Welfare
Dependence (GAO/HEHS-98-109, June 18, 1998) .

Medicaid: Early Implications of Welfare Reform for Beneficiaries and States
(GAO/HEHS-98-62, Feb. 24, 1998 ).

Chief Information Officers: Ensuring Strong Leadership and Effective Council
(GAO/T-AIMD-98-22, Oct. 27, 1997 ).

Child Support Enforcement: Strong Leadership Required to Maximize Benefits
of Automated Systems (GAO/AIMD-97-72, June 30, 1997 ).

Managing Technology: Best Practices Can Improve Performance and Produce
Results (GAO/T-AIMD-97-38, Jan. 31, 1997 ).

Welfare to Work: Most AFDC Training Programs Not Emphasizing Job Placement
(GAO/HEHS-95-113, May 19, 1995 ).

Executive Guide: Improving Mission Performance Through Strategic Information
Management and Technology: Learning From Leading Organizations
(GAO/AIMD-94-115, May 1, 1994 ).

Automated Welfare Systems: Historical Costs and Projections
(GAO/AIMD-94-52FS, Feb. 25, 1994 ).

Strategic Information Planning: Framework for Designing and Developing
System Architectures (GAO/IMTEC-92-51, June 1992 ).

(116030)

Table 1: Increasing Participation Requirements for One-Parent and Two-Parent
Families Under Federal Law 22

Table 2: Information That States Need to Meet Federal Reporting Requirements
and to Participate in the Optional High-Performance
Bonus Program 23

Table 3: Key Information Users and Functions They Perform to Help Low-Income
Families Become Employed 26

Table 4: Case Study States and Local Sites 29

Table 5: Respondents to the Automated Systems Questionnaire 30

Table 6: Capabilities of Automated Systems in Six States to Support Selected
TANF Case Management Activities 36

Table 7: Capabilities of Automated Systems in Six States to Provide Program
Managers With Information on Selected Caseload Characteristics 40

Table 8: Automated Systems Projects Being Designed to Help
Frontline Workers Coordinate Service Delivery 48

Table 9: Automated Systems Projects Being Designed to Enhance Analytical
Capabilities of Program Managers 51

Figure 1: Broad Array of Programs and Services That Case Managers
Can Draw Upon to Serve Their Clients 25

Figure 2: Proportion of Information Needed for Case Management
That Is Provided by Automated Systems 32

Figure 3: Automated Systems That TANF Case Managers Can Access
From Their Desktop Computers 35

Figure 4: Proportion of Information Needed for Service Planning
That Is Provided by Automated Systems 38

Figure 5: Proportion of Information Needed for Monitoring Program
Performance With Respect to Employment Progress of TANF
Recipients That Is Provided by Automated Systems 42

Figure 6: Programs With an Automated System That Either Is the
Same as, or Shares Data With, the System Used for Determining
TANF Eligibility 70
  

1. The Department of Health and Human Services oversees programs such as
TANF, Medicaid, child care, and child support enforcement; the Department of
Agriculture oversees food stamps; and the Department of Labor oversees
employment and training programs.

2. The working seminar has about 30 members, including congressional staff,
federal and state managers of information technology for social programs,
and welfare researchers. See app. I for a list of members and an overview of
working seminar activities.

3. TANF recipients who do not comply with work requirements are subject to
reductions in or terminations of their TANF grants.

4. For this report "workforce development system" refers to the state or
local entity responsible for administering employment and training programs
that originate through the Department of Labor, such as the state Employment
Service or Job Training Partnership Act programs.

5. Under the Job Opportunities and Basic Skills Training (JOBS) program,
many states did have JOBS automated systems that included some capacity for
managing cases, such as tracking attendance in education, training, and
work-related activities. However, these systems did not cover the entire
AFDC caseload because most AFDC recipients did not participate in JOBS.

6. For more detailed information on the proportion of total costs paid by
the federal government for AFDC and Medicaid automated eligibility systems,
see Automated Welfare Systems: Historical Costs and Projections
(GAO/AIMD-94-52-FS, Feb. 25, 1994) . Subsequent revisions led to the 1989
reduction of HHS' funding rate for Medicaid eligibility determination
systems to 50 percent of states' development costs. The Social Security Act
and related amendments had authorized states to operate two types of
automated systems to help manage their Medicaid programs: Medicaid
eligibility systems and Medicaid management information systems. The
management information systems, as required by HCFA, are basically claims
processing and payment delivery systems and are not intended to perform
eligibility determination functions. All states operate their management
information systems separately from their eligibility systems.

7. Legislation in 1993 reduced the food stamp funding rate to 50 percent of
states' development costs, effective April 1, 1994.

8. After this time, HHS no longer certified systems for compliance with
FAMIS requirements.

9. States expended an additional $291 million of state funds for their TANF
programs on systems during the same period.

10. GAO/AIMD-94-52FS, Feb. 25, 1994.

11. JOBS was created in 1988 to ensure that AFDC families obtained the
education, training, and employment that would help them avoid long-term
welfare dependence. PRWORA ended the JOBS program along with the AFDC
program in 1996.

12. All figures are expressed in 1993 constant dollars. About two-thirds of
these costs were for Medicaid systems, including Medicaid management
information systems and Medicaid eligibility systems.

13. HHS, ACF, Office of State Systems, Report to Congress on Data Processing
and Case Tracking in the Temporary Assistance for Needy Families Program
(Washington, D.C.: HHS, Dec. 1997).

14. PRWORA allows states to exempt a family from the time limit because of
hardship, or if the family includes an individual who has been battered or
subjected to extreme cruelty, and to use the average monthly caseload for
either the current fiscal year or the preceding fiscal year as the basis for
determining the 20-percent limit.

15. States were to meet emergency data reporting requirements, issued in
July 1997, until the finalized data collection and reporting requirements,
published in the TANF final rule issued on April 12, 1999, became effective
on October 1, 1999.

16. Through these grants, the Department of Labor is authorized to provide
states and grantees up to $3 billion over 2 years to help the welfare
clients who are considered to be the hardest to employ to find jobs.

17. See Integrating Human Services: Linking At-Risk Families With Services
More Successful Than System Reform Efforts (GAO/HRD-92-108, Sept. 24, 1992
).

18. The need to address data security and confidentiality issues can
complicate interagency efforts to share data, especially in areas such as
child welfare.

19. Before welfare reform, families were automatically enrolled in Medicaid
if they were eligible for cash assistance under AFDC. PRWORA eliminated the
link between eligibility standards for cash assistance and Medicaid,
allowing states to set their own eligibility standards for Medicaid within
certain parameters. To ensure continued Medicaid coverage for low-income
families, PRWORA generally required state Medicaid eligibility standards to
be no more restrictive than the AFDC levels in effect on July 16, 1996.

20. The transitional Medicaid program entitles certain families leaving
welfare as a result of employment or increased earnings to remain eligible
for an additional year of Medicaid.

21. For more information on the effects of welfare reform on Medicaid, see
Medicaid: Early Implications of Welfare Reform for Beneficiaries and States
(GAO/HEHS-98-62, Feb. 24, 1998 ); Medicaid Enrollment: Amid Declines, State
Efforts to Ensure Coverage After Welfare Reform Vary (GAO/HEHS-99-163, Sept.
10, 1999 ); and Marilyn Ellwood, The Medicaid Eligibility Maze: Coverage
Expands, Enrollment Problems Persist: Findings From a Five-State Study
(Washington, D.C.: The Urban Institute, Dec. 1999).

22. See Julie Darnell and others, Medicaid and Welfare Reform: States' Use
of the $500 Million Federal Fund (Washington, D.C.: The Kaiser Commission on
Medicaid and the Uninsured, Oct. 1999).

23. For example, while those who have cycled on and off welfare may be able
to obtain a job fairly easily, they may need assistance to help facilitate
job retention and advancement.

24. For information on the capabilities of state automated systems to
provide information on other characteristics of TANF caseloads, see Richard
P. Nathan and Thomas L. Gais, Implementing the Personal Responsibility Act
of 1996: A First Look (Albany, N.Y.: Nelson A. Rockefeller Institute of
Government, 1999).

25. The average education level of adults receiving welfare in the locality
is about the seventh or eighth grade, according to local officials.

26. Officials explained that the services appropriate for assisting a client
working full-time at a low wage level would likely be different from those
for assisting a client earning a substantially higher wage but working only
part-time.

27. See GAO/HEHS-99-163, Sept. 10, 1999, and Food Stamp Program: Various
Factors Have Led to Declining Participation (GAO/RCED-99-185, July 2, 1999
).

28. The sixth state, Georgia, was still in the process of modifying its
automated systems to track the time limit.

29. See HHS, ACF, Report to the Congress on Data Processing (Washington,
D.C.: HHS, Dec. 1997).

30. One option is for the Congress to review the information in HHS' report
and select a specific system architecture. Another option is to specify that
an evolutionary approach be used to develop a system, whereby progressively
more sophisticated systems could be built in phases using the alternative
system architectures identified by HHS. A third option is to authorize and
appropriate funds for a steering committee that would be responsible for
determining the approach to system development and implementation.

31. This report focused on food stamp participation in California, Florida,
New York, and Texas. See Food Stamp Overpayments: Households in Different
States Collect Benefits for the Same Individuals (GAO/RCED-98-228, Aug. 6,
1998 ).

32. See Mathematica Policy Research, Inc., Options for a National Database
to Track Participation in Federal Means-Tested Public Assistance Programs:
Report to Congress (Washington, D.C.: USDA, Nov. 1999).

33. National Association of State Information Resource Executives, Welfare
Reform and State Human Service Information Systems (Lexington, Ky.: NAIRE,
May 1998).

34. For example, programs sometimes use different conventions for
identifying clients, such as Social Security numbers or system-generated
identifiers.

35. For more information on the process of developing functional and data
requirements, see Strategic Information Planning: Framework for Designing
and Developing System Architectures (GAO/IMTEC-92-51, June 1992 ).

36. States' welfare and workforce development systems have traditionally
been located in separate agencies. For further information on how welfare
reform has affected these systems, see Welfare Reform: States' Experiences
in Providing Employment Assistance to TANF Clients (GAO/HEHS-99-22, Feb. 26,
1999 ).

37. Ohio has a county-administered, state-supervised TANF program.

38. See Welfare to Work: JOBS Automated Systems Do Not Focus on Program's
Employment Objective ( AIMD-94-44, June 8, 1994).

39. See Welfare Programs: Opportunities to Consolidate and Increase Program
Efficiencies (GAO/HEHS-95-139, May 31, 1995 ).

40. Texas officials also pointed out that the separation of Medicaid from
TANF has increased the difficulty of designing systems to determine Medicaid
eligibility.

41. The report noted that generally accepted information technology
standards assume that the average life of a large-scale computer system
ranges from 5 to 7 years, and after that time, new technology advances make
it advantageous to implement a replacement system. See HHS, ACF, Report to
Congress on Data Processing (Washington, D.C.: HHS, Dec. 1997).

42. See Social Service Privatization: Ethics and Accountability Challenges
in State Contracting (GAO/HEHS-99-41, Apr. 5, 1999 ).

43. APDs must contain a state's statement of needs and objectives,
requirements analysis, and alternatives analysis, and they must also set
forth the project management plan with a cost-benefit analysis, proposed
budget, and prospective cost allocations. To obtain continued federal
funding throughout the system's life, a state submits an APD update to
report the system's status and to request additional funding annually or, if
needed, more frequently.

44. APHSA is the organization that represents managers of state and local
human services agencies. See Tools for Information Systems Reform
(Washington, D.C.: APHSA, distributed Dec. 1998), which includes APHSA's
recommendations for improving the federal procurement process.

45. See Welfare Programs: Ineffective Federal Oversight Permits Costly
Automated System Problems (GAO/IMTEC-92-29, May 27, 1992 ) for a more
detailed discussion of the recommendation and related findings.

46. The American Public Welfare Association subsequently changed its name to
the American Public Human Services Association.

47. For example, the action item to propose performance and accountability
standards for systems projects to replace the existing federal up-front
review, approval, and monitoring processes was not implemented.

48. We have issued several guides to general best practices in information
technology. See Executive Guide: Leading Practices in Capital
Decision-Making (GAO/AIMD-99-32, Dec. 1, 1998 ), Executive Guide: Measuring
Performance and Demonstrating Results of Information Technology Investments
(GAO/AIMD-98-89, Mar. 1, 1998 ), Business Process Reengineering Assessment
Guide (GAO/AIMD-10.1.15, Apr. 1997 ), and Executive Guide: Improving Mission
Performance Through Strategic Information Management and Technology
(GAO/AIMD-94-115, May 1, 1994 ).

49. PRWORA does authorize states to use their maintenance-of-effort funds to
provide aid to families who have reached their time limits.

50. The types of automated systems links that states identified in their
responses to this question included daily, nightly, weekly, and monthly
batched data exchanges and the capacity of one system to query data from
another system.

51. As discussed earlier, states were eligible for 90 percent in federal
matching funds before April 1994 if they combined separate systems for AFDC,
food stamps, and Medicaid into a single, integrated Family Assistance
Management Information System. In addition, the Family Support Act of 1988
required that statewide systems be developed to track determination of
paternity and child support collections, and that these systems be linked
with state systems for AFDC.

52. We do not assume that all of the programs listed in fig. 6 should be
linked to the program for TANF eligibility or that these are the only
programs that might be linked. For example, some officials said that it
would be useful to have the system for TANF work activities linked to the
system for job listings to facilitate tracking TANF recipients' referrals to
job listings. There is a wide range of other system links that are
potentially useful for welfare reform, including links to automated systems
for public education, mental health services, and criminal justice. For a
survey of projects under way in 26 states to link data from social service
programs, see UC Data, An Inventory of Research Uses of Administrative Data
in Social Services Programs in the United States−1998 (Berkeley,
Calif.: University of California, Feb. 1, 1999).
*** End of document. ***