Child Support Enforcement: Strong Leadership Required to Maximize
Benefits of Automated Systems (Chapter Report, 06/30/97, GAO/AIMD-97-72).

Pursuant to a congressional request, GAO updated its 1992 report on
child support enforcement, focusing on: (1) the status of state
development efforts, including costs incurred; (2) whether the
Department of Health and Human Services had implemented GAO's 1992
recommendations; and (3) whether the Department was providing effective
federal oversight of state systems development activities.

GAO noted that: (1) it is too early to judge the potential of fully
developed automated systems, yet bringing the benefits of automation to
bear on child support enforcement appears to have played a major role in
locating more noncustodial parents and increasing collections; (2)
according to HHS, in fiscal year (FY) 1995, almost $11 billion was
collected, 80 percent higher than the amount collected in 1990; (3)
while automated state child support systems are being developed, many
may not be certified by the October 1, 1997, deadline; (3) furthermore,
states have underestimated the magnitude, complexity, and costs of their
systems projects; (4) systems development costs for FY 1995 alone were
just under $600 million, and over $2.6 billion has been spent since 1980
for county and statewide systems development; (5) GAO's 1992 report
discussed significant problems in federal oversight and monitoring of
state activity, and made three recommendations; (6) however, only one
has been completely implemented; (7) the Office of Child Support
Enforcement (OCSE) now works with its audit division to identify and
resolve systems problems; (8) GAO's recommendations to suspend federal
funding when major problems exist and to require states to initiate
corrective actions when problems are first identified were only
partially addressed; (9) OCSE's oversight of state child support systems
has been narrowly focused and, as a result, not effective or timely in
assessing the states' systems approaches and progress; (10) OCSE
believes it lacks the technical expertise and resources to be involved
at critical points in the systems development process; (11) OCSE's role
has been primarily limited to document review and after-the-fact
certification when the states request an inspection of completed
systems; (12) therefore, OCSE has allowed some funds to be spent without
ensuring that states were progressing toward effective or efficient
systems; (13) while OCSE has shared some lessons learned, its oversight
has operated on a state-by-state basis; (14) lacking this nationwide
perspective has hindered the agency's ability to provide proactive
leadership to the states; (15) as added systems functional requirements
of the newly enacted welfare reform legislation come into play, it will
be increasingly important that child support enforcement systems work as
envisioned and that OCSE monitor progress on a broader scale; and (16) *

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

 REPORTNUM:  AIMD-97-72
     TITLE:  Child Support Enforcement: Strong Leadership Required to 
             Maximize Benefits of Automated Systems
      DATE:  06/30/97
   SUBJECT:  Child support payments
             Federal/state relations
             Strategic information systems planning
             Requirements definition
             Law enforcement information systems
             Systems design
             Systems evaluation
             State programs
             Computer software
IDENTIFIER:  HHS Temporary Assistance for Needy Families Program
             AFDC
             Aid to Families with Dependent Children Program
             HHS Child Support Enforcement Program
             HHS State System Approval Information System
             
******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO report.  Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved.  Major          **
** divisions and subdivisions of the text, such as Chapters,    **
** Sections, and Appendixes, are identified by double and       **
** single lines.  The numbers on the right end of these lines   **
** indicate the position of each of the subsections in the      **
** document outline.  These numbers do NOT correspond with the  **
** page numbers of the printed product.                         **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************


Cover
================================================================ COVER


Report to Congressional Requesters

June 1997

CHILD SUPPORT ENFORCEMENT - STRONG
LEADERSHIP REQUIRED TO MAXIMIZE
BENEFITS OF AUTOMATED SYSTEMS

GAO/AIMD-97-72

Child Support Enforcement

(511206)


Abbreviations
=============================================================== ABBREV

  ACF - Administration for Children and Families
  AFDC - Aid to Families With Dependent Children
  AIMD - Accounting and Information Management Division
  APD - advance planning document
  APDU - advance planning document update
  CFR - Code of Federal Regulations
  GAO - General Accounting Office
  HHS - Department of Health and Human Services
  IMTEC - Information Management and Technology Division
  OCSE - Office of Child Support Enforcement
  SSAIS - State System Approval Information System

Letter
=============================================================== LETTER


B-275037

June 30, 1997

The Honorable Henry J.  Hyde
House of Representatives

The Honorable Lynn C.  Woolsey
House of Representatives

This report responds to your request that we update our 1992 report
entitled Child Support Enforcement:  Timely Action Needed to Correct
System Development Problems (GAO/IMTEC-92-46, Aug.  13, 1992). 
Specifically, we addressed (1) the status of state development
efforts, including costs incurred, (2) whether the Department of
Health and Human Services had implemented our 1992 recommendations,
and (3) whether the Department was providing effective federal
oversight of state systems development activities.  This report makes
several recommendations designed to increase the likelihood that
states will develop child support systems that perform as required
and to minimize the risk of costly technology decisions and wasted
federal and state expenditures during the development and
implementation of these systems. 

As agreed with your offices, unless you publicly announce the
contents of this report earlier, we will not distribute it until 30
days from the date of this letter.  At that time, we will provide
copies to the Secretary of Health and Human Services; the Director,
Office of Management and Budget; appropriate congressional
committees; and the state child support enforcement offices included
in this review.  We will also make copies available to others upon
request. 

You can reach me at (202) 512-6253 or by e-mail at
[email protected] if you have any questions concerning this
report.  Major contributors to this report are listed in appendix V. 

Joel C.  Willemssen
Director, Information Resources Management


EXECUTIVE SUMMARY
============================================================ Chapter 0


   PURPOSE
---------------------------------------------------------- Chapter 0:1

According to the Department of Health and Human Services (HHS), the
number of child support cases in which collections are being made is
about 20 percent.  As a result, millions of children may not be
adequately provided for or may need to rely on welfare.  Child
support payments will become even more important to recipients who
may cease to be covered by welfare under recently enacted
legislation. 

In an attempt to increase collections of child support, the Congress
in 1980 authorized federal funding to pay up to 90 percent of states'
costs for operating and developing automated child support
enforcement systems.  This has amounted to over $2 billion to date. 
Concerned about how effectively this money has been spent,
Representatives Henry J.  Hyde and Lynn C.  Woolsey asked GAO to
update its 1992 report on this subject,\1

examining (1) the status of state development efforts, including
costs incurred, (2) whether the Department of Health and Human
Services had implemented GAO's 1992 recommendations, and (3) whether
the Department was providing effective federal oversight of state
systems development. 


--------------------
\1 Child Support Enforcement:  Timely Action Needed to Correct System
Development Problems (GAO/IMTEC-92-46, Aug.  13, 1992). 


   BACKGROUND
---------------------------------------------------------- Chapter 0:2

The Office of Child Support Enforcement (OCSE) is part of HHS'
Administration for Children and Families.  The Child Support
Enforcement Program was established in 1975 to help strengthen
families and reduce dependence on welfare by helping to ensure that
the responsibility for supporting children was placed on parents. 
The states operate programs to locate noncustodial parents, establish
paternity, and obtain support orders, along with enforcing actual
collections of those court-ordered support payments.  The federal
government--through OCSE--funds 66 percent of state administrative
and operating costs, including automated systems, as well as 90
percent of expenses associated with planning, designing, developing,
installing, and/or enhancing automated systems. 

The Family Support Act of 1988 required that statewide systems be
developed to track determination of paternity and child support
collections; it set a deadline of October 1, 1995, for implementation
and federal certification of such systems.  However, only a handful
of states met the deadline.  The Congress then passed legislation
extending the deadline by 2 years, to October 1, 1997. 

To meet the criteria for federal funding, state systems were required
to carry out the following specific functions:  case initiation, case
management, financial management, enforcement, security, privacy, and
reporting.  In determining whether a state met these criteria, OCSE
reviewed the advance planning document (APD) that each state had to
develop and submit, describing its proposed system. 

To obtain a broad picture of systems development in the states, GAO
analyzed OCSE and state documents concerning systems development,
visited and conducted structured interviews with officials
responsible for systems development in 6 states and 1 county, held a
focus group discussion with child support enforcement personnel from
14 states and 1 county, and surveyed all 10 HHS regional offices. 
GAO also discussed systems issues with HHS and OCSE officials and
staff, both in Washington, D.C., and at five regional offices. 


   RESULTS IN BRIEF
---------------------------------------------------------- Chapter 0:3

It is too early to judge the potential of fully developed automated
systems, yet bringing the benefits of automation to bear on child
support enforcement appears to have played a major role in locating
more noncustodial parents and increasing collections.  As caseloads
have risen sharply in recent years, the percentage of cases in which
funds are being collected (about 20 percent) has been maintained. 
The increase in total dollars collected has been significant. 
According to HHS, in fiscal year 1995, almost $11 billion was
collected--80 percent higher than the amount collected in 1990. 

While automated state child support systems are being developed, many
may not be certified by the October 1, 1997, deadline.  As of March
31, 1997, only 12 states' systems had been certified.  In fact,
OCSE's director of child support information systems predicted that
as many as 14 of the states--which account for a significant
proportion of the nation's total child support caseload--may miss the
October 1997 deadline.  Furthermore, states have underestimated the
magnitude, complexity, and costs of their systems projects.  Costs
have increased rapidly in the past 5 to 6 years.  Systems development
costs for fiscal year 1995 alone were just under $600 million, and
over $2.6 billion has been spent since 1980 for county and statewide
systems development. 

GAO's 1992 report discussed significant problems in federal oversight
and monitoring of state activity and made three recommendations. 
However, only one has been completely implemented:  OCSE now works
with its audit division to identify and resolve systems problems. 
GAO's recommendations to (1) suspend federal funding when major
problems exist and (2) require states to initiate corrective actions
when problems are first identified were only partially addressed. 

OCSE's oversight of state child support systems has been narrowly
focused and, as a result, not effective or timely in assessing the
states' systems approaches and progress.  The agency does not
evaluate or assess states' systems development projects using a
disciplined, structured approach.  OCSE believes it lacks the
technical expertise and resources to be involved at critical points
in the systems development process.  The agency's role has been
primarily limited to document review and after-the-fact certification
when the states request an inspection of completed systems. 
Therefore, OCSE has allowed some funds to be spent without ensuring
that states were progressing toward effective or efficient systems. 
And, while OCSE has shared some lessons learned, its oversight has
operated on a state-by-state basis.  Lacking this nationwide
perspective has hindered the agency's ability to provide proactive
leadership to the states. 

As added systems functional requirements of the newly enacted welfare
reform legislation come into play, it will be increasingly important
that child support enforcement systems work as envisioned and that
OCSE monitor progress on a broader scale.  Many recipients may find
that they no longer qualify for welfare benefits, with child support
being their only remaining income. 


   PRINCIPAL FINDINGS
---------------------------------------------------------- Chapter 0:4


      SYSTEMS YIELDING BENEFITS,
      DEADLINES MAY NOT BE MET
-------------------------------------------------------- Chapter 0:4.1

Automating child support information systems appears to have improved
caseworker productivity, allowing automatic searches of other
databases--including those containing motor vehicle registrations,
state revenue information, and new employee registries--and
eliminating the need to develop voluminous paper documentation. 
Automated systems also help track court actions relating to paternity
and support orders and amounts of collections and distributions. 

These benefits, however, have been expensive.  Since 1980, states
have spent a combined $2.6 billion on automated systems--with $2
billion of this total being federally funded.  Individual state
estimates of how much will be required to complete their systems are,
in many cases, double initial projections. 

The 12 states that currently have certified systems represent only 14
percent of the national caseload.  Many of the larger states that
OCSE believes may miss this year's deadline have officially reported
to OCSE and to the HHS Office of Inspector General that they will
meet the date.  If they do not, about 44 percent of the national
caseload will not be gaining the benefits of full automation. 


      PROBLEMS IMPEDE PROGRESS;
      EARLIER GAO RECOMMENDATIONS
      NOT FULLY IMPLEMENTED
-------------------------------------------------------- Chapter 0:4.2

Federal and state governments and private industry recognize that
investing time and resources in defining system requirements has a
large program payoff in the development of systems that are completed
on time, are cost-effective, and meet the needs of their users. 
Since major systems decisions hinge on such baseline requirements,
these must be known early in the systems development process.  OCSE
was expected to develop federal requirements for state child support
enforcement systems by 1990, yet final requirements were not issued
until June 1993.  According to OCSE, the delay was caused by its own
failure to use an incremental approach in defining requirements and a
long review process. 

OCSE's delay in providing states with final systems requirements
slowed some states' progress in developing their systems and
contributed to contractor problems.  One state official noted that
the delay contributed to many contract modifications and, eventually,
to the contract's termination.  According to officials of another
state, late functional requirements and unrealistic deadlines
increased costs and delayed the project.  In addition, one of HHS'
regional offices stated that "states began their projects prior to
receiving final requirements; however, [they] were reluctant to
finalize anything until the requirements were issued."

Increased software reuse is an effective means of improving the
productivity of computer software development, improving the
reliability of the software itself, and reducing development time and
cost.  In October 1990, OCSE mandated that states transfer systems
currently in use in other states.  However, it took this step before
assessing the availability of sufficient systems to be used in such
transfers.  In fact, only eight certified systems were then in
operation--and they were based on the 1984 amendments.  No automated
systems had been certified based on the more extensive 1988 act,
making it highly unlikely that the available systems would be
suitable for transfer to other states.\2

As a result, many states attempted to transfer incomplete and/or
incompatible systems, causing added costs and delays. 

Finally, OCSE decided not to fully implement GAO's 1992
recommendation to suspend federal funding when major problems existed
and require states to implement corrective actions as soon as
problems were identified.  OCSE recently stated that it requires
corrective action when problems are noted but explained that it did
not consider withholding funding because it believes the federal
government should work with the states in rectifying deficiencies. 
However, it will now temporarily hold up funding; this has been done
with several states when variations in cost figures were found or
when OCSE had concerns about the system's direction. 


--------------------
\2 The Child Support Enforcement Amendments of 1984 authorized the
federal government to provide states with 90-percent funding for
computer hardware and software to operate certified automated child
support systems.  The Family Support Act of 1988 ended the 90-percent
funding as of September 30, 1995. 


      INADEQUATE FEDERAL OVERSIGHT
      HINDERS STATE SYSTEMS
      DEVELOPMENT
-------------------------------------------------------- Chapter 0:4.3

OCSE does not effectively use the APDs to proactively oversee,
monitor, or control major investments in systems development
projects.  The APD and the states' annual updates--advance planning
document updates (APDUs)--are the basic communications and analysis
tools that OCSE uses in assessing the progress and status of states'
systems, and whether systems meet necessary functional requirements. 
OCSE does not require a disciplined, structured approach for
developing or reviewing systems because, according to agency
officials, it lacks the technical expertise and resources needed to
be involved at critical points in the development process.  Instead,
OCSE primarily focuses on assessing whether states are meeting
systems functional requirements and will meet the October 1, 1997
deadline. 

A disciplined, structured approach to systems development entails
finite phases that must be completed and assessed before moving
forward.  For example, systems planning and analysis must precede
design, design must precede development, and development must precede
implementation.  These are major milestones at which one would expect
and be able to judge progress and determine whether any corrective
actions are necessary.  If work proceeds before an earlier phase is
complete, problems can arise from condensing the work of two or more
phases into one time period, thereby truncating the process.  While
states may provide OCSE--in their APDs or APDUs--with information on
all phases of their systems development, and OCSE and HHS regional
officials may discuss these phases with state officials, the agency
has not used this information effectively at important milestones to
assess system progress and redirect inadequate state development
efforts, losing an opportunity to correct problems early in the
process. 

A critical factor is whether the APD plans are properly carried out
and reflect what the states are actually doing.  State officials
noted that the APDs are not useful for managing systems development. 
According to one, "[APDs] are an administrative exercise to justify
obtaining funding." With OCSE's emphasis on deadlines, states are
often forced to present inaccurate--some feel impossible--schedules
if they are to continue receiving funding. 

Further, while OCSE is required by law to certify state systems,
these certification reviews come too late for timely redirection of
systems development if needed.\3 Since the agency conducts
certification inspections upon state invitation, OCSE is rarely in a
position to promptly intervene and solve problems.  When the agency
does note problems, correction at that point will inevitably be more
time-consuming and expensive than it otherwise may have been. 

In general, OCSE's state-by-state monitoring approach inhibits
effective leadership.  Because of the magnitude of the child support
caseload, the complexity and importance of the automated systems, and
the large amounts of funds invested, a broader, nationwide oversight
that would look for common themes, lessons learned, and systemic
problems, is essential for this program's success.  Without this
perspective, it is difficult to help states manage systems and
control costs.  Post-implementation reviews after state certification
would also give OCSE insight into systems issues and help it to
further assist the states' developmental efforts.  With additional
funding from the welfare reform legislation, OCSE now plans to
conduct post-implementation reviews. 


--------------------
\3 OCSE is required to certify that the states' systems meet the
functional requirements as described in the agency's implementing
regulations. 


      WELFARE REFORM INCREASES
      NEED FOR STRONG FEDERAL
      LEADERSHIP
-------------------------------------------------------- Chapter 0:4.4

New welfare legislation enacted last summer--the Personal
Responsibility and Work Opportunity Reconciliation Act of
1996--dramatically altered the nation's welfare system into one that
requires work in exchange for time-limited assistance.  Since child
support is an integral part of welfare reform, the states are
required to operate a child support enforcement program meeting
federal requirements in order to be eligible for Temporary Assistance
for Needy Families block-grant funding.\4 Guidance the states need to
prioritize systems requirements and changes is not yet available. 
OCSE plans to release the functional requirements on an incremental
basis--issuing requirements on selected systems components as soon as
policy decisions are final.  This will be critical if states are to
meet welfare reform's new requirements for their systems while at the
same time working to complete basic child support enforcement
systems.  Another demand on systems development will be ensuring that
new statewide child support enforcement systems, as well as existing
systems that interface with the new systems, process date-sensitive
information correctly in the year 2000 and beyond.\5


--------------------
\4 This replaced the Aid to Families With Dependent Children program. 

\5 Many older systems that will still be in operation in 2000 were
programmed using 2 digits to represent the year--such as "97" for
1997.  In such a format, 2000 is indistinguishable from 1900. 


   RECOMMENDATIONS
---------------------------------------------------------- Chapter 0:5

GAO is making several recommendations to increase the likelihood of
developing state automated child support systems that will perform as
required.  To maximize the federal government's return on costly
technology investments, GAO recommends that the Secretary of Health
and Human Services direct and ensure that the Assistant Secretary of
the Administration for Children and Families take the following
actions: 

  -- develop and implement a structured approach to reviewing
     automation projects so that significant systems development
     milestones are identified and the costs of project decisions are
     justified during the entire effort;

  -- suspend federal funding for any state that is experiencing
     delays and problems and that is not following generally accepted
     systems development practices until the state redirects its
     approach;

  -- conduct post-implementation reviews to identify any lessons
     learned, to ensure that OCSE incorporates into its oversight
     role a nationwide assessment of child support systems that
     provides a broader perspective on costs, systemic problems,
     potential solutions, and innovative approaches; and

  -- assess the impact of welfare reform on existing child support
     programs and develop timely technical requirements focusing on
     critical systems changes needed by established deadlines. 

GAO is also making other recommendations that are contained in
chapter 6. 


   AGENCY COMMENTS AND GAO'S
   EVALUATION
---------------------------------------------------------- Chapter 0:6

GAO requested written comments on a draft of this report from the
Secretary of Health and Human Services or her designee.  The
Department's comments stressed the difference in perception between
the Department and GAO regarding the appropriate role of HHS'
Administration for Children and Families.  Notwithstanding this
difference, HHS generally agreed with GAO's recommendations that OCSE
evaluate its technical resources, conduct post-implementation
reviews, assess the status of systems nationwide, and define in a
timely manner the requirements arising from recent welfare-reform
legislation. 

In addressing the question of role, HHS commented that it sees OCSE's
role as one of assisting states in meeting congressionally-mandated
deadlines for certification of automated child support enforcement
systems.  The Department disagreed with GAO's view that OCSE should
provide more active, involved monitoring and oversight of state
activities in this area, especially at critical points in the
development cycle.  Along with questioning the appropriateness of
suspending funding for problem projects, officials further stated
that making the recommended changes in the monitoring process would
increase the administrative burden on states. 

GAO believes that HHS has interpreted OCSE's role too narrowly,
failing to take adequate responsibility for helping to ensure the
success of state systems, especially in light of the $2.6 billion
expended on child support enforcement systems--$2 billion of it in
federal funds.  By taking a reactive approach toward
oversight--reviewing state progress annually or only upon request
after major decisions have been made--OCSE does not monitor systems
projects at key points in their development, thereby missing the
opportunity to intervene and help redirect states when problems
arise.  As a result, federal dollars have been invested unwisely on
projects that were allowed to proceed in the wrong direction. 

In terms of administrative burden, GAO believes that the streamlining
it is advocating would not impose an additional administrative
process on states.  In any event, HHS officials did not address the
burden--on states, support recipients, and the federal government--of
failing to effectively complete development of these systems. 

These comments are discussed in chapter 6, and reprinted in appendix
IV. 


INTRODUCTION
============================================================ Chapter 1

The general well-being of children and families is a critical
national policy goal.  Current priorities are aimed at protecting
children and preserving families, including meeting the needs of
millions of parents who annually seek child support for their
eligible children.  However, when noncustodial parents fail to
provide financial support, millions of children must rely on welfare
programs.  In 1995, over 9 million of the 13.6 million people
receiving benefits from the Aid to Families With Dependent Children
(AFDC) program were children.\1


--------------------
\1 As of July 1, 1997, AFDC will be replaced by the Temporary
Assistance for Needy Families block grant. 


   THE CHILD SUPPORT ENFORCEMENT
   PROGRAM
---------------------------------------------------------- Chapter 1:1

The Congress created the national Child Support Enforcement Program
in 1975 as title IV-D of the Social Security Act.  This
intergovernmental program involves federal, state, and local
governments.  The Department of Health and Human Services' (HHS)
regional office staff and the Office of Child Support Enforcement
(OCSE) oversee the state-administered programs.  The purpose of the
program is to increase collections from noncustodial parents and
reduce federal, state, and local welfare expenditures.  As shown in
figure 1.1, reported collections in fiscal year 1995 were 80 percent
higher than they were in 1990. 

   Figure 1.1:  Total Child
   Support Collections, Fiscal
   Years 1990-1995

   (See figure in printed
   edition.)

Source:  Child Support Enforcement--Twentieth Annual Report to
Congress (draft), for the period ending September 30, 1995
(HHS/ACF/OCSE).  We did not independently validate this information. 

The number of reported child support cases has also increased 60
percent--from 13 to 20 million cases over that same time period.  As
a result, according to HHS, the number of cases in which collections
are being made has remained about 18 to 20 percent. 

Families entering the Child Support Enforcement Program require
different combinations of services at different times, and child
support enforcement agencies are directly responsible for providing
these services.  For instance, in some cases the child's paternity
has not been established and the location of the alleged father is
unknown.  In these cases, the custodial parent needs help with every
step:  locating the alleged father, establishing paternity, obtaining
and enforcing a child support order, and collecting the support
payment.  In other cases, the custodial parent may already have a
child support order; in such a case, the child support enforcement
agency must review and possibly modify the order as a result of
changes in the employment status or other circumstances of the
noncustodial parent before tackling enforcement. 

State child support enforcement programs are organized in
significantly different ways.  They report to different state
agencies and follow different policies and procedures.  In addition,
relationships between the state child support enforcement programs
and other state agencies differ.  These characteristics usually vary
by the type of service delivery structure, levels of court
involvement required by state family law, population distribution,
and other variables.  For example, some state child support agencies
manage their programs centrally (operating a number of state
offices), while others allow the counties or other governmental
entities or even private companies to manage the programs locally.\2


--------------------
\2 Child Support Enforcement:  Early Results on Comparability of
Privatized and Public Offices (GAO/HEHS-97-4, Dec.  16, 1996). 


   STATES ARE DEVELOPING FEDERALLY
   FUNDED INFORMATION SYSTEMS TO
   SUPPORT THE PROGRAM
---------------------------------------------------------- Chapter 1:2

Growing caseloads, increased costs, and social demands have given
rise to the need to implement expedited processes for establishing
and enforcing payment of child support.  As such, automation was (and
still is) seen by many, including the federal government, as an
effective tool for addressing this need.  In 1980, the Congress
promoted the development of automated systems that could improve the
performance of the child support program.  Public Law 96-265
authorized the federal government to pay up to 90 percent of the
states' total costs incurred in planning, designing, developing,
installing, or enhancing these systems.\3 The systems are required by
OCSE to be implemented statewide and be capable of carrying out
mandatory functional requirements, including case initiation, case
management, financial management, enforcement, security, privacy, and
reporting.\4 Incorporating these requirements can help locate
noncustodial parents and monitor child support cases.  Since 1981,
the federal government has spent over $2 billion for automated
systems to assist states in collecting child support. 

The Family Support Act of 1988 mandated that by October 1, 1995, each
state have a fully operational automated child support system that
meets federal requirements.  At that time, the 90-percent development
funding was to be discontinued.  In addition, if a state did not have
its system certified as fully operational by this date, the act
declared that the state's child support program may have its program
funding reduced.  However, by October 1, 1995, only five states had
met the deadline.  Therefore, the Congress passed Public Law 104-35,
extending the deadline to October 1, 1997. 

Developing child support enforcement systems is a joint federal and
state responsibility.\5

In providing most of the funding for systems, the federal government,
through OCSE, is responsible for providing leadership, technical
assistance, and standards for effective systems development.  OCSE is
also responsible for assessing states' automated systems and ensuring
that states are effectively using the 90-percent funding.\6

To receive 90-percent federal funding for the development of an
automated child support enforcement system, a state is required to
develop and submit an advance planning document (APD) to OCSE,
describing its proposed system.  The APD is reviewed by OCSE's
Division of Child Support Information Systems and by HHS' regional,
program, and financial management staff to ensure that the proposed
system incorporates the minimum functional requirements and will meet
federal, state, and user needs in a cost-effective manner.  After the
APD is approved, OCSE provides 90-percent funding for the project and
monitors its progress.  Federal regulations and OCSE guidance (1)
require states to update their APDs when projects have significant
changes in budget or scope and (2) give OCSE the authority to suspend
funding if a state's development does not substantially adhere to its
approved plan.  When the state considers its system complete, a state
official requests that the federal government certify that its system
meets requirements.  After certification, a state is authorized to
receive additional funding to maintain its operational system. 


--------------------
\3 For the purposes of this report, the term systems refers to the
hardware and software components of the child support enforcement
systems. 

\4 A functional requirement is a requirement that specifies a
function that a system or system component must be able to perform. 

\5 In HHS' state systems advance planning document guide, the agency
notes that the administration of this program is a cooperative
endeavor, with federal and state governments working together to
implement information systems.  HHS provides leadership and direction
and is responsible for approving, monitoring, and certifying states
programs--and ensuring that federal funds are spent wisely. 

\6 The federal government provides 66 percent of the costs incurred
by states for operating child support programs, which includes
operating and obtaining automated systems.  The "enhanced" funding of
90 percent is paid to develop systems. 


   RECENT WELFARE REFORM
   LEGISLATION
---------------------------------------------------------- Chapter 1:3

While states are still trying to meet the challenges of the 1988 act,
they are also faced with newer challenges.  The Personal
Responsibility and Work Opportunity Reconciliation Act of 1996
requires that states implement specific expedited and administrative
procedures intended to expand the authority of the state child
support agency and improve the efficiency of state child support
programs.  In order to comply with the expedited processes
requirement, states have to meet specific time frames for
establishing paternity and establishing and enforcing support
orders.\7 Under current law their statewide systems must
automatically perform specific locate, establishment, enforcement,
and case management functions and maintain financial management,
reporting, and under the new law states' security and privacy
functions.\8 In addition, under the new law states must enhance their
current statewide systems to electronically interface with other
federal and state agencies.  This is needed to establish, for
example, central case registries and new-hire directories. 
Therefore, to successfully comply with the welfare reform
legislation, it is critical that the states and OCSE have fully
operational child support systems in place. 


--------------------
\7 States must (1) establish support orders within 6 months of
service in 75 percent of title IV-D cases needing orders and within
12 months in 90 percent of cases, (2) take action to enforce
delinquent orders within 30 days (or 60 days if service is needed),
(3) send advance notice of income withholding within 15 days of
location, and (4) meet additional time frames for interstate cases. 

\8 For example, states must establish automated registries of child
support orders and directories of newly hired employees to track and
locate parents owing support. 


   OBJECTIVES, SCOPE, AND
   METHODOLOGY
---------------------------------------------------------- Chapter 1:4

In 1992, in response to a request from the Senate Committee on
Finance, we reviewed HHS' oversight of states' efforts to develop
automated child support enforcement systems.  In August 1992, we
issued a report citing major problems with oversight and monitoring
of these development efforts.\9 We reported that while taking timely
corrective action on known problems is critical to developing well
designed automated systems, OCSE had not required needed changes in
some states facing serious systems problems.  We, therefore, made
recommendations to HHS for improvement.  On June 20, 1996,
Representative Henry J.  Hyde requested that we conduct a follow-up
review.  Later, Representative Lynn C.  Woolsey joined in this
request. 

Our specific objectives were to determine (1) the status of automated
state systems, including costs, (2) whether HHS had implemented our
1992 recommendations, and (3) whether HHS was providing effective
federal oversight of state systems development. 

To accomplish these objectives, we reviewed federal laws and
regulations on OCSE's oversight of state development of automated
systems.  We assessed OCSE systems guidelines, policies and
procedures, and correspondence with the states.  We also interviewed
officials in OCSE's Office of Child Support Information Systems and
Division of Audit to discuss their continued roles and
responsibilities in overseeing the planning, development, and
implementation of state child support enforcement systems. 

To update our knowledge of automated systems issues, we analyzed
state planning documents, OCSE certification reports, and state audit
reports of automated systems.  In addition, we reviewed financial
reports produced by OCSE's statistical and reporting systems;
however, we did not independently verify data contained in these
reports.  We coordinated with the HHS Office of Inspector General and
reviewed, analyzed, and summarized the results of its nationwide
child support systems state survey.  We also interviewed selected
contractors developing and implementing child support enforcement
systems.  Further, we conducted a focus group of 18 state officials,
representing 14 states (California, Connecticut, Delaware, Georgia,
Iowa, Louisiana, Massachusetts, Michigan, Minnesota, Missouri,
Nebraska, New York, Ohio, and Texas) and Los Angeles County to
determine the benefits, barriers, and solutions to developing
automated child support systems. 

We performed our work at OCSE headquarters in Washington, D.C.  We
also surveyed all 10 HHS regional offices and visited 5 (Atlanta,
Dallas, Denver, New York, and Philadelphia) to gain an understanding
of the history of each state's development effort and of OCSE's role
in providing regional oversight and technical assistance.  Further,
we visited six states (Alabama, California, Massachusetts, Ohio,
Texas, and Washington) and Los Angeles County.  We selected these
locations based on the following criteria:  levels of funding
requested, methods used to develop systems (e.g., in-house,
contractor, and combination of in-house and contractor), caseload,
geographic location, phase of development (e.g., pilot,
implementation, enhancement, in operation), and level of
certification.  During our site visits we assessed the systems'
status, best practices, and barriers to implementing systems using
relevant components of our system assessment methodology.\10 We also
reviewed various state and contractor systems-related documents and
correspondence and interviewed state agency officials. 

We conducted this work between August 1996 and March 1997, in
accordance with generally accepted government auditing standards.  We
requested written comments from the Secretary of Health and Human
Services or her designee.  The Inspector General provided us with
written comments, which are discussed in chapter 6 and reprinted in
appendix IV. 


--------------------
\9 Child Support Enforcement:  Timely Action Needed to Correct System
Development Problems (GAO/IMTEC-92-46, Aug.  13, 1992). 

\10 The System Assessment Framework:  A Guide for Reviewing
Information Management and Technology Issues in the Federal
Government (GAO/AIMD-10.1.12, August 1996). 


SYSTEMS YIELD BENEFITS, BUT
DEADLINE MAY NOT BE MET
============================================================ Chapter 2

Many states have made progress in their automation projects, and
state officials report that the systems have already demonstrated
benefits.  However, some states' costs are twice as high as
originally estimated, and the extent of final costs is not yet known. 
Progress in developing systems varies--some states have automated
many features, while others are in the earlier phases of development
and may not be certified or operational by the October 1, 1997,
deadline. 


   AUTOMATED SYSTEMS APPEAR TO BE
   HELPING STATES
---------------------------------------------------------- Chapter 2:1

According to state program and systems managers, child support
enforcement systems have improved program effectiveness and worker
productivity by automating inefficient, labor-intensive processes and
monitoring program activities.  Systems have improved efficiency by
automating the manual tasks of preparing legal documents related to
support orders and calculating collections and distributions,
including interest payments.  Further, automated systems can help
locate absent parents through interfaces with a number of state and
federal databases more efficiently than could the old, manual
process.  These systems have also improved tracking of paternity
establishment and enforcement actions.  The following examples show
specific reported improvements in program performance since states
began developing their automated systems. 

  -- According to one systems official, while it is difficult to
     attribute benefits entirely to the system, "the system has
     changed the way business is done in the child support office."
     The automated system assisted the state's staff in increasing
     the number of parents located from almost 239,000 in fiscal year
     1995 to over 581,000 in fiscal year 1996--a reported increase of
     143 percent.  Additionally, while using the system from July
     1994 through December 1996, the staff increased the number of
     support orders established by over 78 percent, the number of
     paternities established by almost 89 percent, and child support
     collections by almost 13 percent. 

  -- Officials from another state noted that staff performance
     increased with the new system because it gives staff a new tool
     to use to improve their productivity.  Collections per employee
     have more than doubled--from about $162,000 to $343,000
     annually.  The system has also helped the state reduce the time
     required to process payments:  The turnaround time for checks
     for nonwelfare custodial parents dropped from 29 days to
     processing a payment in only seconds and issuing checks within
     24 hours. 

  -- Officials from the same state also reported that the automated
     system allowed cases to be viewed on-line by several individuals
     simultaneously, eliminating the bottlenecks created by manually
     searching, retrieving, and delivering hard copy case files. 
     Before the state implemented its new system, 11 percent of the
     child support staff was dedicated to the manual process of
     retrieving files. 


   BILLIONS SPENT FOR AUTOMATED
   SYSTEMS NOT YET IN COMPLIANCE
   WITH FEDERAL MANDATES
---------------------------------------------------------- Chapter 2:2

States have spent billions of dollars on automated child support
enforcement systems.  Costs for developing and operating these
systems continue to mount, while progress in developing systems
varies.  Despite the escalating costs, only 12 systems have been
certified, and as many as 14 states may not meet the October 1997
deadline. 


      SYSTEMS COSTS CONTINUE TO
      INCREASE; VARY WIDELY FROM
      STATE TO STATE
-------------------------------------------------------- Chapter 2:2.1

According to OCSE records, states have spent over $2.6 billion since
the early 1980s to develop, operate, maintain, and modify county and
statewide automated child support systems.  Of these costs, the
federal government has paid 66 to 90 percent of states' systems
costs, amounting to more than $2 billion.  Since 1980, federal
expenditures for child support enforcement systems have risen
dramatically.  Figure 2.1 shows the history of federal funding for
these systems from fiscal year 1981 through fiscal year 1996.  As the
chart reveals, federal spending escalated as states began working to
comply with the 1988 act.  Appendix I provides the total reported
costs for each state's child support system and the federal and state
shares of those expenditures, and appendix II provides the enhanced
and regular federal expenditures.\1

   Figure 2.1:  Cumulative Funds
   Spent on Child Support
   Enforcement Systems, Fiscal
   Years 1981-1996

   (See figure in printed
   edition.)

Note:  We did not independently validate funds spent on child support
enforcement systems. 

Source:  HHS. 

Although the 90-percent enhanced funding ended on October 1, 1995,
the Congress later retroactively extended it to October 1, 1997, for
those states having approved enhanced funding in their APDs as of
September 30, 1995. 

States generally underestimated the costs of developing and operating
child support enforcement systems.  During our seven site visits, we
compared the projected costs in the original APDs with the most
recent estimates.  While two sites' original estimates were fairly
accurate, the remaining five were significantly understated.  For
example, in total, current cost estimates for the states we visited
are about twice as high as originally planned.  In addition, at least
10 states are now discovering that their systems will cost more to
operate once they are completed.  While these states expected cost
increases as a result of added system functionality, increased
information storage, and the use of sophisticated databases,
estimated operating costs for some new systems may be even higher
than anticipated.  For example, one state's initial estimate showed
the new system's data processing costs would be three to five times
higher than that of the old system.  However, according to a state
official, those costs will likely be six to seven times higher than
the current system's operating costs.  Operating that state's new
system may cost nearly $7 million more annually than the old system. 

Further, costs for developing and operating child support systems
have varied greatly among the states--from a low of $1.5 million to a
high of $344 million.  The difference can be attributed to a variety
of factors, including caseload size, whether the states or the
counties administer the child support program, the number of attempts
states made to develop child support enforcement systems, the way the
systems were developed--by modifying an existing system or developing
a new one, and the kind of system being developed. 


--------------------
\1 We did not independently validate OCSE's state and federal child
support systems costs. 


      FEW STATES CERTIFIED AS
      DEADLINE APPROACHES
-------------------------------------------------------- Chapter 2:2.2

Only five systems were certified and seven conditionally certified as
of March 31, 1997.  OCSE grants full certification when a system
meets all functional requirements and conditional certification when
the system needs only minor corrections that do not affect statewide
operation.  Figure 2.2 indicates which states have certified and
conditionally certified systems as of March 31, 1997. 

   Figure 2.2:  Certified and
   Conditionally Certified Child
   Support Enforcement Systems as
   of March 31, 1997

   (See figure in printed
   edition.)

Source:  HHS. 

The certified and conditionally certified states represent only 14
percent of the nation's reported child support caseload.  Further,
according to OCSE's director of state child support information
systems, as many as 14 states--6 with caseloads over 500,000--may not
have statewide systems that fully meet certification requirements in
place by this October 1.  These states represent 44 percent of the
nation's child support caseload.  OCSE does not yet know how long it
will take or how much it will cost to bring these states into
compliance with federal requirements.  In addition, 2 of these states
that chose to update their existing systems rather than develop new
child support enforcement systems may need to redesign their systems. 

Responding to an HHS Office of Inspector General survey, 36 of the 42
states\3 that are not certified reported that they will meet the 1997
deadline.  However, this task may well present a challenge for many
of them.  While almost two-thirds of the states reported that they
were either enhancing operational systems to meet certification
requirements or in the conversion or implementation phases, the
remaining one-third of the states responding to the 1996 survey
stated that parts of their systems are only in the design,
programming, or testing phases of systems development--with major
phases to be completed, including conversion and statewide
implementation.\4 State systems officials in our focus group
considered conversion to be one of the most difficult problems and a
barrier to successful implementation. 


--------------------
\3 We are using the term state to refer to the 50 states plus the
District of Columbia, Guam, Puerto Rico, and the Virgin Islands. 

\4 The Office of Inspector General's survey was sent to the states in
April 1996 and completed and returned by August 13, 1996. 


PROBLEMS IMPEDED EARLIER PROGRESS
AS RECOMMENDATIONS NOT FULLY
IMPLEMENTED
============================================================ Chapter 3

All too often an organization's inability to take basic but necessary
steps to decrease systems development risks leads to failure. 
Problems consistently identified in reviews by GAO and others include
information systems that do not meet users' needs, exceed cost
estimates, or take significantly longer than expected to complete. 
In its efforts to assist in the development of automated, statewide
child support enforcement systems, OCSE is no exception.  The agency
did not define requirements promptly, adequately assess systems
issues prior to mandating a transfer policy, or seek to identify and
aggressively correct problems early in the development process.  The
lack of sound, timely federal guidance, coupled with some states' own
inadequate systems approaches, caused systems development activities
to proceed with increased risks. 


   DELAYED FUNCTIONAL REQUIREMENTS
   SLOWED PROGRESS
---------------------------------------------------------- Chapter 3:1

The federal and state governments and private industry recognize that
an investment of time and resources in requirements definition has
the biggest program payoff in the development of systems that are on
time, cost-effective, and meet the needs of its users.  Major systems
decisions hinge on baseline requirements; these requirements,
therefore, must be defined early.  Without them, reasonable estimates
of the scope, complexity, cost, and length of a project cannot be
adequately developed.\1 In addition, failure to clearly and
accurately define requirements may preclude alternatives, restrict
competition, and further increase the risk of cost and schedule
overruns.\2

According to OCSE's director of state child support information
systems, the agency was expected to develop federal requirements for
the statewide systems by October 1990.  However, OCSE did not publish
federal regulations--which described in general the program and
automated systems--until October 1992.  The agency did provide draft
systems development guidance--functional requirements--to the states,
but it was not disseminated until July 1992.  OCSE did not provide
the states with final systems functional requirements until June
1993.\3

OCSE acknowledges that the federal requirements were late; it
attributes this primarily to its not using an incremental approach in
releasing the requirements to the states.  Rather than issuing
certain requirements as they were defined, it waited until all
requirements could be issued together.  This was done because OCSE
believed that certain underlying policy issues needed to be resolved
before requirements could be made final and released.  In addition,
because OCSE's minimum requirements were extensive and difficult
policy issues needed to be resolved, the review and approval process
also contributed to the delayed issuance of requirements.  According
to OCSE, it took time to assess policy issues, determine the most
effective way to automate related changes, and accurately define
related requirements.  For example, according to OCSE, before it
could even begin to define requirements related to the replacement of
monthly mail-in notices (e.g., of a client's child support status)
with telephone recordings, complex policy decisions had to be
considered. 

Another example of a policy issue needing resolution, according to
OCSE, was related to guidance on financial distributions.  For
instance, OCSE had to assess how systems would handle the required
"$50 pass-through" policy of the program.\4 This policy states that
the first $50 of current child support payments collected for a child
also covered under Aid to Families With Dependent Children (AFDC)
must be delivered to the mother of that child rather than to the
state AFDC office.  While this policy sounds simple, it presented a
certain degree of administrative complexity, especially for cases in
which support payments were not made on time.  In such a case,
regardless of how many months a payment has been in arrears, only $50
(for the current month) goes directly to the mother. 

We have indicated in the past that an agency in the process of
defining and analyzing requirements should assess the impact of
changes on other organizational elements;\5

therefore, we agree that policy issues such as these must be
addressed prior to developing detailed requirements.  We also agree
with OCSE that where possible, it should have made final and issued
certain requirements sooner, using an incremental approach. 

Delays in issuing final systems functional requirements meant more
than just a late start; they compounded other problems. 
Uncertainties about final requirements slowed some states'
development activities and contributed to contractor problems.  Seven
of the 10 regions we surveyed indicated that the delay in
requirements had an adverse effect on their states' development.  The
following excerpts illustrate the impact of the delay on certain
state projects. 

  -- An official of one region (representing four states) indicated
     that the requirements were issued much later than needed and
     ranked the "lack of timely final functional requirements" the
     number one impediment to states' systems development.  This
     official added that all four of its states "wanted better
     clarification in black and white as to .  .  .  what the system
     should look like, how it should operate .  .  .  etc." The
     regional official added that its states were "left to figure
     this out for themselves, then they [had] to pass a certification
     review that is totally subjective in the areas of functionality
     and level of automation."

  -- Another regional official said that all five of its states began
     their projects prior to receiving final requirements; however,
     they were reluctant to finalize anything until final
     requirements were issued.  Another HHS regional official said,
     "the delay in getting official regulations published impacted
     contracts with the vendors and was an embarrassment to ACF
     [Administration for Children and Families] [yet, because of the
     deadline imposed], development efforts went forward."

  -- Still another region surveyed indicated that requirements were
     somewhat late and, for four of its six states, this was an
     impediment.  However, a regional official alluded to strong
     contractor relations as one of the primary reasons that the
     delay was not a problem (but otherwise could have been) for two
     of the region's states.  The official stated, "the effect of
     timing on [these two states was] unique.  Each was building one
     system that supported a few offices throughout each
     jurisdiction.  Each also had a fairly good working relationship
     with its vendor.  Because of the nature of the projects, the
     type of environment and working relationship with the vendor,
     [both states were not as affected by the delay]."

Likewise, during our visits to individual states, four of them
attributed their systems development problems to late functional
requirements.  One state official noted that the late functional
requirements and short time frames contributed to development delays
and increased costs.  Another state noted that the delay in
functional requirements contributed to many changes in the contract
and, eventually, to contract termination.  Finally, in one state, the
child support systems' development contract had to be amended to
address modifications to the functional requirements.  In fact, three
work segments in the contract had to be added as a result of these
modifications, increasing project costs by at least $210,000 due to
reprogramming. 

The HHS Office of Inspector General has similarly reported that
several states experienced systems problems as a result of late
functional requirements.\6 In response to the Office of Inspector
General survey, one state official noted that the state's system was
designed to meet the requirements set out in the draft guidelines;
once the final requirements were issued, the state had to shift to a
new initiative and virtually start over.  Another state official said
".  .  .  the late issuance of the certification guide was a major
factor in the decision to delay statewide implementation of the
system.  The delay in receiving the guide caused [the state] to
compress the development cycle of [its] subsystem, putting a higher
risk on the success of the overall system."


--------------------
\1 Mission Critical Computer Resources Management Guide, Defense
Systems Management College (September 1988). 

\2 Information Technology:  An Audit Guide for Assessing Acquisition
Risks (GAO/IMTEC-8.1.4, December 1992). 

\3 Automated Systems for Child Support Enforcement:  A Guide for
States, U.S.  Department of Health and Human Services, Administration
for Children and Families (Revised June 1993).  (This guide replaced
draft guidance distributed in July 1992.)

\4 The "$50 pass-through" (also referred to as the "$50 disregard")
is the amount collected and distributed as payment to families and
disregarded in consideration of welfare program eligibility.  The $50
pass-through was eliminated effective October 1, 1996, by the
Personal Responsibility and Work Opportunity Reconciliation Act of
1996. 

\5 GAO/IMTEC-8.1.4, December 1992. 

\6 Implementation of State Child Support Certified Data Systems,
Department of Health and Human Services, Office of Inspector General
(OEI-04-96-00010, April 1996). 


   PREMATURE POLICY MANDATE ALSO
   DELAYED DEVELOPMENT, INCREASED
   COSTS
---------------------------------------------------------- Chapter 3:2

In addition to being hindered by the delay in functional
requirements, states encountered delays in developing systems and
incurred more costs as a result of OCSE's policy requiring states to
transfer systems.  Two years after the passage of the 1988 act, OCSE
required states to "transfer" existing child support systems from
other states or counties rather than building entirely new systems. 
While certainly a reasonable approach for saving money, at the time
of this policy mandate, only a few available systems had been
certified as meeting OCSE's old requirements and no systems were
certified based on the more extensive 1988 act.  As a result, states
had difficulty transferring these systems and adapting them for their
own programs.  OCSE had intended for the transfer policy to be an
efficient method of building systems; however, in many cases, the
transfer policy actually slowed systems development and led to
increased systems costs when states attempted to transfer
incompatible, incomplete, or inadequately tested systems. 


      POLICY MANDATE POORLY
      PLANNED AND IMPLEMENTED
-------------------------------------------------------- Chapter 3:2.1

Before issuing its transfer policy, OCSE did not perform sufficient
analyses to support the requirement that states transfer systems.  By
not thoroughly evaluating the available alternatives, OCSE had no
assurance that states would be able to transfer systems in an
effective and efficient manner.  As a result, states were faced with
choosing from a limited number of systems, some of which were
incomplete or unsuitable for their systems environments. 

OCSE established the transfer policy on October 9, 1990; it said that
all states must transfer a child support system from another state or
county.  OCSE noted that states needed to review other states'
systems and determine how these systems varied from their own systems
requirements.  According to OCSE, this sharing of technology among
states would decrease the installation time for automated systems and
reduce the risk of systems failures due to poor system design or
inadequate planning. 

The transfer policy required states to reuse software.  Software
reuse can be an appropriate part of systems development projects. 
According to the National Bureau of Standards, one of the most
effective means of improving the productivity of software development
is to increase the proportion of software that is reused.  Reusable
software not only increases productivity, but also improves
reliability and reduces development time and cost.  However, many
technical, organizational, and cultural issues usually need to be
resolved before widespread reuse of software should be mandated.\7 In
this case, this was not done. 

If properly implemented, OCSE's transfer policy could have saved
states time and money in developing child support systems.  As early
as 1987, we reported that sharing state systems could save OCSE time
and money.\8 However, we also noted at that time that OCSE had not
adopted standards or provided adequate oversight of states' efforts
to develop compatible and transferable automated systems.  Careful
and detailed alternatives analyses are required prior to selecting
software to be transferred.  Analyses should consider functional
requirements; standardization of data elements; compatibility of
software and hardware platforms; and other factors, such as caseload
processing, organizational structure, state and contractor expertise
and skills, and any unique state requirements.  For alternatives,
state agencies should consider only completed systems that have been
tested, validated, and successfully used in operation to ensure that
benefits will be achieved. 

Two factors intensified the need for adequate planning for software
reuse:  states' different organizational structures and methods of
administering the program and the magnitude of changes required by
OCSE's implementation of the 1988 act.  Since states administer the
federal child support program, each state determines how its program
will be organized and operated.  These differences in state programs
affect systems development.  For example, county-administered states
faced an additional challenge in complying with the 1988 requirement
to have one statewide child support system.  Those states had to
build systems that considered the needs of users in all of their
county offices, complicating system design.  Careful planning for
software reuse was especially important.  Under the 1988 act and
implementing regulations, states were required to obtain and track
more detailed information on each absent parent, child, and custodial
parent.  OCSE mandated that states transfer systems before making the
requirements final, so they were unable to first evaluate the ability
of the transfer systems to meet those requirements. 

Despite its lack of requirements, oversight, and alternatives
analyses, OCSE mandated the transfer policy without performing
sufficient analyses or feasibility studies of existing certified
systems as potential transfer candidates.  Only eight certified
systems were available when the mandate was issued, and these were
certified based on the 1984 requirements.  At the same time, no
automated systems were certified based on the more extensive 1988
act, making it highly unlikely that the available systems would be
suitable for transfer to other states.  While, in July 1994, OCSE
changed its transfer policy making it optional, this was after most
states had attempted to transfer systems and when systems efforts had
progressed further. 

Only one state we visited noted that it had successfully transferred
another system.  It was among the last to transfer a system,
initiating the transfer in 1994.  Moreover, the state began its
project after the final federal requirements were issued, conducted
thorough analyses of three potential systems, and transferred a
system that had already been certified as meeting the 1988
requirements.  In addition, the system selected was the result of a
successful, earlier transfer from another location. 

However, some states we visited did not take as thorough an approach
and faced difficulties in attempting to transfer existing child
support systems. 

  -- One project team we visited spent almost $400,000 attempting to
     transfer a system from another state, only to discover that the
     transfer was not possible because the system was not compatible
     with its existing operation. 

  -- Another state's official explained that, to meet the mandate and
     the 1995 deadline, his state attempted to transfer a system that
     was immature, incomplete, and inadequately tested.  While the
     state started to implement parts of the transfer system, the
     entire system was not delivered until a year later, increasing
     project costs. 

  -- Yet another state attempted to have a contractor modify a
     transfer system that was also incomplete.  Later, the entire
     effort was abandoned, wasting over a million dollars and
     contributing to a delay of several years. 

The HHS Office of Inspector General recently reported that 71 percent
of states said that their attempts to transfer a system delayed,
rather than enhanced, development of an automated system.  In
response to that survey, one state official noted, ".  .  .  when we
began our project there were no systems certified to the 1988 level. 
We chose one state's system as our transfer model and wasted about a
year documenting all of its deficiencies in order to justify not
transferring it." In addition, we were told by officials in several
other states that they transferred only concepts from other
systems--that the amount of computer code actually transferred was
negligible. 


--------------------
\7 William Wong, Management Guide to Software Reuse, Department of
Commerce, National Bureau of Standards' Special Publication 500-155
(April 1988). 

\8 Letter to the Administrator, Family Support Administration, HHS,
Feb.  20, 1987 (B-221220). 


   GAO RECOMMENDATIONS ONLY
   PARTIALLY ADDRESSED DESPITE
   DELAYS AND ESCALATING COSTS
---------------------------------------------------------- Chapter 3:3

In our 1992 report, we stated that efforts to develop child support
enforcement systems were plagued by problems, particularly in the
area of federal oversight provided to states.\9

According to laws and regulations, OCSE is responsible for
continually reviewing and assessing the planning, design,
development, and installation of automated systems to determine
whether such systems will meet federal requirements.  OCSE is
required to monitor 90-percent federally funded child support systems
to ensure that they are successfully developed and are
cost-effective.  If this is not the case and if a state is not
substantially adhering to its approved plan, OCSE is authorized to
suspend federal funding. 

Past compliance reviews conducted by OCSE's systems division
identified many deficiencies with states' development of automated
child support systems and escalating costs.  For example, development
of three severely flawed systems continued at a total cost of over
$32 million before they were stopped and redirected by OCSE.  Rather
than directing needed remedial actions when these problems were
identified, OCSE informed the states of the deficiencies yet
continued to fund the systems based on states' assurances that the
problems would be addressed.  Further, OCSE's systems division did
not routinely use audit division reports to help monitor development
because it was not required to do so. 

In 1992, we recommended that OCSE (1) work with the audit division to
identify and resolve systems-related problems, (2) use its authority
to suspend federal funding when major problems existed, and (3)
require states to implement needed corrective actions when first
identified.  Despite the seriousness of the problems we identified in
1992 and recurring problems since then, the only recommendation fully
implemented by OCSE was the first one, regarding working with the
audit division to identify and resolve systems-related problems. 
OCSE reviews audit reports prior to certification visits and the
auditors are now members of the certification review teams.  In
addition, officials in both the systems and audit areas stated that
communication and coordination between the two have improved
substantially since our 1992 report. 

Even though systems costs were escalating, OCSE did not fully
implement the other two recommendations.  It continues to assert that
the federal government should work with the states to correct
deficiencies rather than take enforcement actions.  However, law and
regulations require that OCSE monitor the 90-percent federally funded
child support systems to ensure that they are successfully developed. 
OCSE is authorized to suspend federal funding if a state is not
substantially adhering to its approved plan.  While HHS regional
staff noted that OCSE either held up, reduced, or stopped funding to
18 states since the 1988 act, almost 60 percent of these reported
disruptions were due to insufficient information on the required APD
or the state's exceeding its authorized funding level.  OCSE believes
that suspending funds is counterproductive to helping states meet the
deadline. 

Even when funding was held up for major systems-related problems,
efforts to correct these problems did not appear to be timely.  For
example, one HHS regional official suggested to OCSE that it hold up
funding for projects in his region, yet, according to this official,
the agency did not stop any funding until one of those projects'
"initial efforts crashed." In another instance, OCSE had serious
concerns about the status and future direction of one state's
project, staffing levels, and methods of incorporating changes in
code.  As a result, it held up funding for several months.  However,
some problems were not identified until 2 years into the project's
life cycle; as of the end of our review, OCSE was still working with
the state to resolve them.  In another case, OCSE identified problems
with poorly documented code, inadequate planning and guidance, and
contract management but did not stop federal funding. 

According to OCSE's systems division director and analysts, states
have primary responsibility for developing their systems and
therefore the federal government should not assume a primary role in
directing how states should develop systems and remedy problems.  We
disagree with OCSE's approach of continuing to fund systems with
serious problems that endanger the projects' success.  Such an
approach involves the risk of needing to fix serious problems later
in the development process, when it is much more costly and
time-consuming to do so. 


--------------------
\9 GAO/IMTEC-92-46, Aug.  13, 1992. 


INADEQUATE FEDERAL OVERSIGHT AND
LEADERSHIP CONTINUE TO HINDER
STATES' DEVELOPMENT EFFORTS
============================================================ Chapter 4

Because of the complexity, costs, and large caseloads associated with
the child support program, effective development of automated systems
requires continuing oversight and strong leadership.  Yet despite the
pressure that federal and state agencies are under to improve their
child support enforcement services, the development of statewide
automated systems is hampered by ineffective federal leadership and
some inadequate state development approaches.  Major mechanisms OCSE
uses to oversee states' systems development projects include reviews
of the states' advance planning documents (APDs); advance planning
document updates (APDUs); and certification reviews, which are
assessments to determine if projects meet federal requirements. 
While these reviews provide OCSE information on states' plans for
designing automated systems, the agency does not effectively use the
APDs to oversee, monitor, or control the systems development
projects, and the certifications are performed too late in the
process to detect and correct problems. 

In short, critical systems development decision points are not
monitored by OCSE and reviews of states' systems are primarily
focused on determining whether all federal requirements have been
met.  Further, OCSE has not completed nationwide analyses or
post-implementation reviews to effectively assess lessons learned,
hindering its ability to provide more thorough, helpful leadership. 
OCSE acknowledges these facts, yet cites a management approach that
holds states responsible for developing their systems; OCSE believes
it lacks the technical expertise and resources needed to be involved
at critical points in the development process. 


   APD USE INADEQUATE TO REDIRECT
   STATES
---------------------------------------------------------- Chapter 4:1

While the review of APDs is one of OCSE's principal vehicles for
monitoring states' systems activities, the agency's review is
inadequate for systems monitoring because OCSE does not require a
disciplined, structured approach for developing or reviewing systems. 
As such, the APDs are not used to measure systems development
progress at key decision points; rather, according to state
officials, APDs are used primarily as a funding approval mechanism
only, requiring the states to report information and plans on the
basis of a deadline that may or may not be realistic.  Consequently,
systems problems may go undetected until much later in the process
when they are considerably more difficult and expensive to correct. 

APDs are written plans of action submitted by states to request
federal funds for designing, developing, and implementing their
systems.  According to an OCSE guide, the three primary purposes of
the APD process are to (1) describe in broad terms a state's plan for
managing the design, development, implementation, and operation of a
system that meets federal, state, and user needs in an efficient,
comprehensive, and cost-effective manner, (2) establish system and
program performance goals in terms of projected costs and benefits,
and (3) secure federal financial participation for the state.\1 The
document contains the state's statement of needs and objectives,
requirements analysis, and alternatives analysis.  The APD also sets
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 APDU to report the system's status and to request
additional funding--annually or, if needed, more frequently. 

Public Law 100-485 requires that states submit APDs to OCSE and that
based on the APDs, OCSE review, approve, and fund information
systems.  This review focuses on ensuring that each state
incorporates the minimum functional requirements by the legislatively
mandated date of October 1, 1997.  OCSE is also required to review
the security requirements, intrastate and interstate interfaces,
staff resources, hardware requirements, and the feasibility of the
proposed plan.  HHS regional staff support OCSE by monitoring states'
development efforts and, at times, assist the states in preparing
their APDs and APDUs.  In addition, regional staff provide states
technical assistance and suggestions to help the states comply with
systems requirements. 

A well-defined, disciplined structure for systems development,
covering the status of systems at critical design points, is key to
preventing software development problems and encouraging strong,
effective management oversight.  For example, phases include systems
planning and analysis, design, development, and implementation. 
These are major milestones in any system development project and are
to be used to identify risks, assess progress, and identify
corrective actions needed before proceeding to the next phase. 
Appendix III provides an overview of typical systems development
phases.  The key is to identify system risks and problems early in
the design, in order to avoid major failures and abandoned projects,
and to help ensure that major segments do not have to be extensively
recoded or redesigned.\2

While states recognize the benefits of using a structured systems
development approach and may provide OCSE with information on each of
these phases in their APDUs, OCSE does not effectively use this
information for determining the adequacy or progress of the systems
development projects.  APDUs are revised annually, rather than
corresponding to major system phases.  According to OCSE's director
of state child support information systems, the agency has not
monitored the projects using a structured approach because it lacks
the technical expertise and resources needed to be involved at
critical points in the development process.  Recent funding from the
welfare reform legislation has allowed OCSE to conduct more frequent
reviews; it is critical to assess systems development activities at
each phase and to identify problems and any potential risks early,
before moving forward to the next phase. 

Further, OCSE has not provided specific guidance describing the
information needed in the APDUs to assess different phases of
development.  For example, while OCSE requires that the states
provide schedules of systems development activities, these
schedules--formats, descriptions, and structure--vary from state to
state and, even within a state, may vary from year to year, making it
difficult to effectively assess state progress and monitor systems
development.  Specifically, some APDUs do not provide information on
how much data have been converted, code written, modules produced,
nor portions and results of the modules tested, again hindering
OCSE's ability to effectively measure progress. 

While state officials indicated that the APDUs were useful for
budgetary purposes, officials at six out of seven locations we
visited said that APDUs are not useful in helping them manage their
systems developments.  "They are an administrative exercise to
justify obtaining funding," said one.  Officials also noted that the
deadline seemed to be OCSE's primary concern; even when the deadline
seemed impossible to meet, the states were forced to present
inaccurate schedules. 


--------------------
\1 State Systems APD Guide, Department of Health and Human Services,
Administration for Children and Families and Health Care Financing
Administration (September 1996). 

\2 GAO/AIMD-10.1.12, August 1996. 


      PROBLEMS IN STATE
      DEVELOPMENT APPROACHES NOT
      CORRECTED OR REDIRECTED
-------------------------------------------------------- Chapter 4:1.1

Because of the significant financial investment in information
systems and their crucial role in helping locate noncustodial
parents, a structured systems development approach is essential to
reducing major systems risks.  In several instances, even when
states' APDUs contained adequate information to identify significant
problems in approach, OCSE has not required the states to correct the
deficiencies.  As a result, more money was spent and systems
underwent further development without disciplined, structured systems
development approaches, increasing the likelihood of system failures. 

In the case of one state we visited, as shown in figure 4.1, despite
three revisions to time estimates for systems testing, OCSE approved
each updated estimate.  Only after the state officials told OCSE that
delays and problems were occurring did the agency, in November 1996
and in February 1997, review systems progress.  State officials told
us that they wanted the OCSE review to identify that the time for
testing was insufficient to support the state in vendor discussions,
yet the agency had not, as of March 31, 1997, reported on this
matter.  According to OCSE's director of child support information
systems, the review focused only on one specific functional
requirement and was not a comprehensive systems review, and would
not, therefore, address the software testing issue. 

   Figure 4.1:  State Plans for
   Software Programming and Unit
   Testing

   (See figure in printed
   edition.)

Source:  State's child support enforcement systems advanced planning
document updates. 

In systems development, building and testing sections of computer
code and then using test results to refine the software are
critical.\3 Testing results need to be considered in correcting
errors and improving identified inefficiencies.  Despite this
essential need for systems testing, in the example in figure 4.1,
OCSE continued to approve the APDUs and fund the project, without
assessing the increasing risks.  The APDUs also showed that the state
was not following a structured systems development approach and was
developing software while at the same time attempting to integrate
software in preparation for systems testing.  According to this
state's systems project director, the project's costs increased from
$30 million to over $50 million, in part because of the poor quality
of software developed.  If work on a later phase proceeds before an
earlier phase is completed, one risks problems from condensing work
and truncating the process.  Completion of the automated system is
now uncertain. 

In another state we visited, OCSE had ample indication that the
state's system was experiencing difficulty, yet failed to act until
just recently.  In August 1993, the state initially scheduled
implementation to last until April 1996--a period of 39 months.  The
estimated completion date was subsequently extended twice--first in
September 1995 by an additional 10 months (to February 1997), and
then again in October 1996 by an additional 5 months (to July 1997). 
These extensions, taken together, prolonged the implementation phase
by almost 40 percent, from an initially envisioned 39 months to 54
months.  The 1995 updated estimate should have signaled to OCSE that
the state's system needed help.  However, not until the state
requested a substantial increase--$133 million in project costs--did
OCSE, in January 1997, question the state's systems progress. 
According to OCSE's director of child support information systems,
the agency should have visited this state sooner to provide technical
assistance.  Given that the October 1996 update judged implementation
to be less than half finished, it is questionable whether the July
1997 target is even realistic. 

OCSE has not effectively provided program oversight by detecting and
redirecting states' approaches that are inadequate and threaten
successful systems development.  When planning and developing a
system, a state must ensure that it meets users' needs, provides the
intended benefits to users and their constituencies, and is developed
on time; otherwise, it will not be effective.\4 Gaining commitment
and support from key decisionmakers is particularly important for
states developing child support systems because they serve a variety
of users--district attorneys, county and state officials, and court
officers.  Key users' needs should be considered, and general
agreement is critical in developing systems' requirements and design. 
However, not all states clearly defined their functional requirements
and user needs. 

In one state, the system design was insufficient because it did not
clearly define functional requirements and did not gain the support
and involvement of key county officials.  While OCSE reviewed
requirements and reported minor deficiencies, it did not mention any
serious problems in the state's systems development approach.  Yet
just 1 year later, after the state had an independent contractor
conduct a risk assessment of the project, state officials
acknowledged that the approach needed to be completely revamped. 
According to OCSE's director of child support information systems,
the agency's review for this state focused only on whether functional
requirements were being met rather than the systems approach being
followed.  She further noted that the state is responsible for
developing the system and obtaining system buy-in from the users. 
Because of OCSE's large financial investment in systems--the approval
of $154 million for this state alone--and the importance of gaining
user acceptance for successful systems development, we believe the
state and federal agencies are both responsible. 

Two states we visited also encountered similar problems by not
involving all key players--program, systems, and management
officials--in the decision-making process.  In one state, the
information resources management official was not involved in the
systems planning and implementation.  Later a disagreement occurred
in policy decisions related to operating the system, resulting in an
abrupt change in management of the system.  Consequently, the
project's direction is uncertain and the system conversion progress
has been delayed.  Another state allowed its child support program
office to develop its system without ensuring that key officials with
technical program and managerial expertise were adequately involved. 
Systems problems were identified late in development, and project
management was changed to put more emphasis on technical project
experience.  If this technical expertise had been involved initially,
the system might have been more successfully planned and implemented. 


--------------------
\3 GAO/AIMD-10.1.12, August 1996. 

\4 GAO/AIMD-10.1.12, August 1996. 


   CERTIFICATION REVIEWS NARROW IN
   FOCUS AND PERFORMED TOO LATE
   FOR EFFECTIVE OVERSIGHT
---------------------------------------------------------- Chapter 4:2

While OCSE does review and certify states' systems to determine if
functional requirements are being met, as described in the 1988 act
and the agency's implementing regulations, these reviews are
conducted toward the end of a systems development project, at the
states' request, and are not performed at critical decision points
(such as analysis, design, coding, and testing).  Thus, these reviews
are too late to identify problems and redirect approaches.  Despite
being the predominant stakeholder for over $2.6 billion worth of
child support systems, OCSE maintains that its philosophy is to work
with the states as opposed to suspending funding until problems are
corrected, allowing federally financed projects to proceed when such
projects do, in fact, require redirection. 

Two types of reviews are conducted by OCSE:  functional and
certification reviews.  During a functional review, OCSE helps the
state work on specific system requirements.  For example, it may
telephone or visit the state to discuss how to meet automation
requirements for the noncustodial parent locating function.  The
certification review comprises a two-level process, level 1 and level
2.  A level 1 review is performed when an automated system is
installed and operational in one or more pilot locations.  (OCSE
created this level of review in 1990 due to state requests for agency
guidance prior to statewide implementation.) A level 2 review is
conducted when OCSE visits the state to determine if the system meets
certification requirements.  The certification review normally takes
a week to perform; the OCSE review team is made up of headquarters
systems officials and HHS regional personnel representing the
systems, fiscal, program, and audit functions. 

Prior to a certification review, OCSE provides a questionnaire
including questions on how the system meets specific federally
required functions (such as case initiation and the detail supporting
this function) and a test deck of financial transactions (mainly test
cases of different types of child support payment distributions) to
be run on the state's system.  OCSE also supports, through federal
funds, child support user group meetings so state officials can meet
and share related systems experiences. 

OCSE's reviews--functional and certification--of state systems focus
on whether functional requirements are met, while they lack but need
comprehensive assessments of the systems' development approaches and
schedules.  Since 1991, OCSE has visited, assessed, and reported on
31 states' system development projects.\5 We reviewed all of these
and found that 28 of the 31 functional and certification reviews
focused primarily on systems' functional requirements.  Specifically,
OCSE determines whether the system can initiate a new case, locate an
absent parent, establish a support order, manage cases, enforce
cases, perform financial distributions, perform management reporting,
and maintain security/privacy.  Suggestions for improving the
efficiency and effectiveness of the systems' operations are discussed
in reports following these reviews.  For example, recommendations
include having the states consider automating more functions, such as
developing an automatic tickler file to remind caseworkers of
required actions or an approaching time frame.  While suggestions for
improved efficiency in automation are valuable, OCSE's reviews lack a
comprehensive assessment of the states' systems development approach,
including the overall design, project management, user involvement,
and delays in major milestones and critical tasks.  While 3 of the 31
reviews did address systems development approaches and overall
project management, OCSE officials noted that they only do this type
of review for states that are experiencing delays and significant
problems. 

Even though OCSE required states to submit APDs by October 1, 1991,
all but 1 of the 31 functional and certification reviews conducted by
OCSE were performed in October 1995 or later.  Since OCSE only does
certification reviews upon state request and toward the end of
systems development, much of the federal funding had already been
spent. 

Even when OCSE identifies state systems problems and notifies the
states, corrective actions do not always follow.  In one state, OCSE
performed two reviews and noted that the state had serious managerial
problems.  There was no project manager for extended periods of time,
and users did not support the project, despite the fact that a key
early step in designing a proposed system is identifying and
satisfying users' needs.\6 Even after the second review was performed
and the problems noted in the first review had not been corrected,
OCSE took no action to stop or delay the project or to suspend
funding.  Again, according to OCSE systems analysts, the agency's
main focus is to help states fix problems as they arise; agency
officials believe that withholding funds is counterproductive to
meeting deadlines.  As a result, the federal government has approved
over $50 million for this one system; it has been in partial
operation--primarily in the smaller counties--since mid-1993, and its
expected completion is still not known. 

In another example, OCSE approved a state's proposal to meet
functional requirements by developing a distributed system estimated
to cost nearly $178 million to complete and maintain through 2000. 
This system involves developmental costs for at least 23 county
databases, plus additional costs to maintain separate systems.  While
OCSE questioned the state about the additional costs of personnel,
the agency's certification process does not require that a state
receive an approval before moving from one phase of development to
the next.  By not visiting the project to review critical design
documents, the agency was not able to effectively assess the
distributed processing strategy and associated costs or to suggest an
alternative approach.  This state now estimates that the system will
cost about $311 million to complete and maintain through 2000. 
Further, according to state officials, the system will not meet the
October 1997 deadline. 

Systems managers from three states we visited indicated their desire
that OCSE play a more active role.  Systems that are being developed
are very sophisticated, and the officials said that it was important
for OCSE to assess the management and direction of the project early
to avoid or minimize problems later.  One state official said that
OCSE should see itself as a stakeholder in the development of these
systems and not just a reviewer to see if certain requirements were
followed.  Another official noted that OCSE is scrupulously hands-off
with all, especially with the private sector.  Another state official
told us that because the contractor's work was so poor and late in
delivery, the state had to support the project with more of its own
staff, contributing to an increased cost of $20 million.  When the
state asked for assistance from OCSE on how to handle the contractor,
they were told that "it was a state contract and it had to be
resolved at that level."

Further, while a recent HHS Office of Inspector General survey noted
that 70 percent of the states felt that OCSE's guidance was good to
excellent--including the certification guide, clarification of
requirements, and questions and answers on functional
requirements--43 percent said they needed additional technical
assistance.  One state official said "there has been very little
monitoring to date.  The delays will come when we request
certification and OCSE doesn't like what it sees." Another state
official noted, "It would have been helpful to have had more
compliance reviews and technical reviews of designs, but it was
probably not practical given OCSE's [limited] staffing resources."
Finally, another state official noted "I've only known OCSE to be the
gatekeeper with regard to funding and ultimate certification.  I
don't have any experience to see them in another role."


--------------------
\5 Since 1991, OCSE has visited and reviewed 34 state systems
projects; however, the agency had only issued reports on 31 of these
reviews as of March 31, 1997. 

\6 GAO/AIMD-10.1.12, August 1996. 


   LACK OF NATIONWIDE PERSPECTIVE
   INHIBITS EFFECTIVE LEADERSHIP
---------------------------------------------------------- Chapter 4:3

The narrowness of OCSE's reviews limits its ability to gain a
nationwide perspective on the status of states' systems development;
this, in turn, hinders effective leadership in earlier stages.  OCSE
has completed neither a comprehensive, nationwide analysis nor
post-implementation reviews to determine whether state systems are
sound financial investments.  Such analyses are essential to
assessing the ongoing progress of child support systems and
evaluating the impact of automated systems on program goals and
objectives--including any lessons learned.\7 While the Administration
for Children and Families (ACF) developed an automated system to help
the agency perform individual state and nationwide analyses, we found
the system was not being fully used, was not user friendly, and
contained errors in about half of the states' data. 

Because OCSE has not conducted a nationwide assessment, it has not
analyzed the hardware, software, database structures, and networks
supporting state child support systems; as a result, state officials
have had to discuss these issues through informal means.  If OCSE had
performed a nationwide analysis, it would have a sound basis for
encouraging states to share innovative database designs, software,
and other technologies for greater efficiencies and cost savings. 
Further, analyses of systems costs, benefits, and schedules from a
nationwide perspective would help identify where improvements are
needed in a timely manner.  Aggregate data on projects help identify
recurring problems, successes, and other trends for decision-making
purposes.\8

To collect states' data more effectively, assess systems costs, and
monitor projects nationwide, ACF created the State System Approval
Information System (SSAIS).  This was designed to establish a more
accurate way of tracking state systems projects.  SSAIS tracks the
historical data on automated systems projects--including the child
support program--on a state-by-state basis.  Users may access the
following data on any state project:  (1) funding requests, reviews,
and approvals, (2) names of contractors, (3) completed systems
reviews, and (4) notes from systems reviews. 

While we commend ACF's actions to develop this system, we found that
SSAIS recorded incorrect funding levels for nearly half of the
projects.  During our comparison of OCSE's approval letters and the
data in SSAIS, we found errors, such as duplicate entries and entries
for systems projects no longer underway, and inconsistencies, such as
entries for some states that included planning costs while entries
for other states did not.  Specifically, SSAIS showed that the states
were authorized to spend nearly $1.9 billion on automated child
support systems projects, while hard copy approval letters maintained
by OCSE indicated that the states were authorized to spend about $100
million less. 

OCSE's director of state child support information systems
acknowledged that SSAIS is not yet comprehensive, complete, or user
friendly.  She said that for these reasons, OCSE was not fully using
or relying on the SSAIS entries at the time of our review.  However,
she acknowledged that our review led the division to assign a higher
priority to correcting discrepancies in SSAIS and to establishing a
consistent policy regarding entries for planning costs. 

In addition to tracking systems development nationwide,
post-implementation reviews are the basic means of ensuring that
systems meet program objectives and identify techniques for improving
work processes, data integrity, and project management--thereby
avoiding costly systems mistakes.\9 This information is also helpful
in identifying the benefits of information technology projects and
prioritizing technology investments that best meet mission needs.  To
date, OCSE has not completed post-implementation reviews on any of
the 12 certified or conditionally certified child support systems. 

OCSE's information systems director called post-implementation
reviews important.  However, she said that until recently, resource
constraints limited OCSE's ability to complete such reviews. 
Instead, OCSE focused its resources on assessing the states' progress
in meeting system certification requirements.  Under welfare reform,
1 percent of the federal share of child support payments collected
annually will be provided for state child support programs and
systems oversight.\10 According to OCSE's information systems
director, this will allow OCSE to conduct post-implementation
reviews. 


--------------------
\7 Assessing Risks and Returns:  A Guide for Evaluating Federal
Agencies' IT Investments Decision-making (GAO/AIMD-10.1.13, February
1997). 

\8 GAO/AIMD-10.1.13, February 1997. 

\9 GAO/AIMD-10.1.13, February 1997. 

\10 For 1996, the federal share of child support payments collected
amounted to $1.2 billion.  One percent of this--$12 million--is
provided for program and systems oversight. 


WELFARE REFORM PLACES NEW DEMANDS
ON CHILD SUPPORT SYSTEMS
============================================================ Chapter 5

Recently enacted welfare reform legislation substantially increases
the importance of collecting child support payments from noncustodial
parents, since for welfare recipients who lose eligibility, child
support may be their only remaining source of income.  As such, it
will be even more important that state automated systems operate
correctly and efficiently in helping eligible child support
recipients collect funds due them.  The law also places specific new
requirements on states relating to the functioning of their systems. 
Recognizing the importance of these developments, OCSE plans to
change its approach and issue functional requirements incrementally,
and has taken steps to work with the states; however, the impact of
welfare reform and associated costs is not yet known.  Another demand
on systems development will be monitoring and ensuring that new
statewide child support enforcement systems, as well as existing
systems that interface with the new systems, will process
date-sensitive information correctly in the year 2000 and beyond. 


   WELFARE REFORM FEATURES
   AUTOMATED SYSTEMS
---------------------------------------------------------- Chapter 5:1

In August 1996, the Congress enacted the Personal Responsibility and
Work Opportunity Reconciliation Act, fundamentally changing the
nation's welfare system into one that requires work in exchange for a
5-year program of assistance; implementing many of its most critical
features involves automated systems.  The law contains work
requirements, a performance bonus that rewards states for moving
welfare recipients into jobs, and comprehensive child support
enforcement measures.  It also provides support for families moving
from welfare to work--including increased funding for child care and
guaranteed medical coverage.  Provisions are also included to improve
automation in order to increase paternity establishment, obtain more
information on work and residence locations of noncustodial parents,
and process child support orders and collections. 

Both the states and ACF will be required to address these provisions
by developing new databases or enhancing existing automated child
support systems before October 1, 2000.  A $400 million cap has been
placed on enhanced federal matching funds\1 through 2001 for
development costs of automated systems, and funds are to be allocated
to the states on the basis of existing workloads and level of needed
automation. 

Welfare reform further underscores the need for streamlined business
processes and automated systems for the child support program.  Since
welfare reform establishes time limits and eligibility restrictions
on individuals in the Temporary Assistance for Needy Families block
grant, states are being faced with the need to increase child support
collections.  According to experts, this likely will force some
states to manage child support cases differently and require
modifications to existing laws governing child support operations.\2
States will be required to automate many child support operations to
more efficiently disclose, exchange, and compare information on
noncustodial parents owing delinquent support payments.  Some of the
more significant state welfare reform systems requirements include

  -- establishing, by October 1, 1997, or following the close of the
     next regular legislative session, a statewide system for
     tracking paternity orders and acknowledgements of paternity;

  -- developing, by October 1, 1997, a new-hire registry on which
     employers will report information on employees recently hired,
     with the capability of reporting the information to a national
     database and issuing wage withholding notices to employers
     within 2 business days, and making data comparisons with the
     case registry database by May 1, 1998;

  -- developing, by October 1, 1998, a central case registry for all
     child support cases and support orders established or modified
     in the state after that date and, as of that date, capable of
     making data comparisons; and

  -- establishing, by October 1, 1998 (or by October 1, 1999, if
     court administered), a centralized unit to collect and disburse
     child support payments, and by October 1, 2000, a statewide
     child support system that meets all requirements.\3

To comply with these mandates, many states will not only have to
reassess the way child support cases are managed administratively but
electronically as well.  This may include developing new databases
and electronic links to other public and private organizations,
including financial institutions; credit bureaus; the Internal
Revenue Service; and state agencies--including judicial, corrections,
licensing, business ownership, motor vehicle, labor, vital
statistics, and Medicaid.  For example, the requirement to develop a
registry of paternities will likely result in the development of a
new database that interfaces with departments or bureaus responsible
for tracking statewide births.  However, tracking these data
electronically may be a challenge since some states have not
automated the departments or bureaus responsible for birth
certificates.\4

The states will also have to develop automated interfaces with the
national Federal Parent Locator Service and the yet-to-be-developed
federal case registry and new-hire registry databases.  The Federal
Parent Locator Service is an electronic system that cross-matches
data to help locate noncustodial parents across state lines through
links with other state systems and existing national databases, such
as those of the Internal Revenue Service, Social Security
Administration, and Department of Labor. 

Both the new-hire registry and the federal case registry will also
need to be designed to receive and compare state data on child
support cases and noncustodial parents.  The states, in turn, will be
required to develop similar databases that electronically interface
with these national systems.  The federal new-hire and case registry
systems must be completed by October 1, 1997, and October 1, 1998,
respectively.\5 OCSE has contracted with several vendors to develop
the software for these databases and to provide states with technical
assistance. 

Many states will have to adapt their existing systems and change laws
governing child support operations to implement many of these systems
requirements.  For example, three states we visited indicated that
existing laws governing the child support program and new-hire
reporting requirements for employers would have to be amended or
rescinded before mandated systems requirements could be implemented. 
Other states will have to make welfare reform system changes while
finishing work on child support systems mandated by the 1988 act. 


--------------------
\1 The legislation lowered the federal reimbursement rate for
enhanced funding from 90 percent to 80 percent.  The cap applies only
to this enhanced funding, not to regular funds, which continue to be
reimbursed at a rate of 66 percent. 

\2 Vickie Turetsky, Child Support Administrative Processes:  A
Summary of Requirements in the Personal Responsibility and Work
Opportunity Reconciliation Act of 1996, Center for Law and Social
Policy (January 1997). 

\3 Vickie Turetsky, Child Support Administrative Processes. 

\4 Vickie Turetsky, Child Support Administrative Processes. 

\5 Welfare Reform Act--Telephone Seminar:  Personal Responsibility
and Work Opportunity Reconciliation Act of 1996, American Bar
Association family law section and KRM Information Services, Inc. 
(Oct.  2, 1996). 


   OCSE PLANS TO ADDRESS WELFARE
   REFORM
---------------------------------------------------------- Chapter 5:2

While OCSE has initiated some steps to identify welfare reform
systems issues and plans to change its approach in issuing functional
requirements, the agency does not yet know the impact the legislation
will have on state systems.  As of March 31, 1997, OCSE had not
developed functional requirements for implementing welfare reform or
fully analyzed the impact of these provisions on existing state child
support systems.  Guidance for developing the new-hire registries had
not been completed, even though these registries are due to be in
operation by October 1, 1997.  However, OCSE does plan to apply
lessons learned from technology projects mandated as part of the 1988
Family Support Act and release technical guidance for systems changes
required by welfare reform incrementally.  The early release of
technical guidance should help states decide on systems requirements
as soon as possible, minimizing project delays. 

With the increased funding being made available through welfare
reform, OCSE plans to conduct more on-site reviews of child support
systems projects to help identify and prevent costly systems
development problems during earlier stages of the projects.  The
agency has also supported ACF user groups and established electronic
information bulletin boards to identify and share information on
systems issues.  Further, OCSE is participating in welfare reform
work groups with the states to discuss policy and systems-related
issues.  In January 1997, the agency also created a federal, state,
and local government initiative to work with the eight largest
states--representing almost 50 percent of the child support cases. 
This initiative focuses on improving program performance, which may
include automated systems issues.  In addition, the agency recently
queried the states to identify technical support needs and planned to
issue, in May 1997, a national plan to better address OCSE technical
assistance to the states' systems activities.  Despite these early
attempts to work with the states, according to OCSE's information
systems director, the agency does not know the impact the welfare
reform will have on the states' child support systems.  She noted
that until the requirements are defined, the extent of systems
changes and their costs are not known. 


   MILLENNIUM CHANGES ARE ALSO
   IMPORTANT TO AUTOMATION
---------------------------------------------------------- Chapter 5:3

The change in century could have a significant impact on state
systems that process date-dependent information related to child
support.  Ensuring that all state child support enforcement systems
adequately address the processing of information that is
date-dependent is critical.  Correcting noncompliant year-2000
software may be expensive.  Among these systems are those that must
interface directly and provide information to the newly-developed
child support enforcement systems.  Many older state systems that
will still be in operation in 2000 were programmed using 2 digits to
represent the year--such as "97" for 1997.  However, in such a format
2000 is indistinguishable from 1900. 

OCSE has stated that it has informed the states that both the new
child support systems and their applications software under
development, as well as the existing systems that must still
interface with the new statewide systems and their applications
software, must be year-2000 compliant. 


CONCLUSIONS, RECOMMENDATIONS, AND
AGENCY COMMENTS
============================================================ Chapter 6

Thanks to technology, many states are better able to locate
noncustodial parents who owe child support payments, seize government
tax refunds or benefits, and issue child support payments to families
more efficiently.  But this progress has been expensive.  The cost of
developing state child support enforcement systems has risen over the
past 15 years to the point that it now exceeds $2.6 billion, of which
$2 billion is federal funds.  The amount remaining to be spent to
bring all states into full legal compliance is unknown.  At the time
of our review, most states did not yet have fully functional child
support enforcement systems.  Aside from new requirements resulting
from welfare legislation, only 12 states had federally certified
child support enforcement systems; as many as 14 states--responsible
for about 44 percent of the national caseload--may well miss the
October 1, 1997, deadline for completing their automated systems. 

The causes are widespread.  States have underestimated the magnitude,
complexity, and costs of their projects and operations, and they
could have received better guidance and assistance from the federal
government, specifically OCSE.  The lack of progress in the
development of state child support systems also can be partly
attributed to the agency's limited leadership and oversight and some
states' inadequate systems approaches.  OCSE's release of final
functional requirements for the state systems was late, which
encouraged some states to automate many tasks without adequate
requirements management or control.  Though ready, some states
hesitated to make their systems' requirements final; it must be
remembered that deadlines loomed, with or without final requirements. 
Another factor was OCSE's mandated transfer policy, which was
premature and poorly implemented.  This alone caused long-term
problems, increased costs, and delays. 

Against this backdrop, which included the failure to fully implement
recommendations we made some 5 years ago, OCSE allowed state systems
with serious problems to proceed, thus escalating spending with no
assurance that effective, efficient systems would result--and many
indicators to the contrary.  Specifically, OCSE did not establish
levels of oversight and technical review commensurate with the size
and complexity of this nationwide undertaking.  It did not require
states to follow a structured systems development approach; nor did
OCSE assess progress at critical decision points, thereby missing
opportunities to intervene and successfully redirect systems
development. 

OCSE relied on required annual planning documents, which were
optimistic projections and for many states did not relate to the
critical phases of system development.  Certifications were narrowly
focused and conducted only at a state's request, when the state was
ready.  While OCSE has supported state-to-state interaction with
users' groups, the agency itself cannot develop a truly nationwide
perspective without an understanding of the trends that typify
development of individual state systems.  Lacking this knowledge,
OCSE cannot disseminate valuable information to states in earlier
stages of development. 

During the last 5 years, when much of the money has been spent and
when it was most critical for OCSE to take a leadership role and
evaluate states' efforts, agency and HHS regional officials noted
that their oversight was hindered by limited technical expertise and
resources.  Critical areas of systems expertise--systems development,
systems engineering, and program management--are essential to assess
how effectively systems are being implemented. 

Because of the magnitude of the caseload, the funds being provided,
and the importance of the program's mission, it is essential that
both federal and state officials take responsibility for developing
effective and efficient automated child support systems.  While
evaluating states' efforts is one major component of OCSE's role, it
is important that the agency considers itself a stakeholder in these
efforts.  The problem appears to stem from OCSE's view of its
role--one of merely monitoring requirements and approving funds
rather than being held accountable for effective systems development
approaches. 

With the enactment of welfare reform, OCSE's role becomes much more
important:  for those who may no longer be eligible for welfare
benefits and rely solely on child support, the effectiveness of their
state's system will be critical.  Effective, strong federal
leadership will be necessary if we are determined to support those
who rely on these automated systems. 


   RECOMMENDATIONS
---------------------------------------------------------- Chapter 6:1

We are making several recommendations to increase the likelihood of
developing state automated child support systems that will perform as
required.  To maximize the federal government's return on costly
technology investments, we recommend that the Secretary of Health and
Human Services direct and ensure that the Assistant Secretary of the
Administration for Children and Families take the following actions. 

  -- Develop and implement a structured approach to reviewing
     automation projects to ensure that significant systems
     development milestones are identified and that the costs of
     project decisions are justified during the entire effort.  We
     recommend each major systems phase be reviewed and, at critical
     points--analysis, design, coding, testing, conversion, and
     acceptance--that OCSE, according to preestablished criteria,
     formally report to the state whether it considers the state
     ready to proceed to the next milestone or phase. 

  -- Develop a mechanism for verifying that states follow generally
     accepted systems development practices to minimize project risks
     and costly errors.  OCSE should revise the guidance for the APDs
     and APDUs to ensure that these documents provide the information
     needed to assess different phases of development and are
     consistent from year to year.  This information should include
     clearly defined requirements; schedules reflecting the amount of
     data converted, code written, modules produced, and the results
     of testing; and other measures to quantify progress. 

  -- Use an evaluative approach for planned and ongoing state
     information technology projects that focuses on expected and
     actual cost, benefits, and risks.  OCSE should require states to
     implement needed corrective actions for federally funded systems
     when problems and major discrepancies in cost and benefits are
     first identified.  If a state experiences delays and problems
     and is not following generally accepted systems development
     practices, OCSE should suspend funding until the state redirects
     its approach. 

  -- Evaluate current staff systems knowledge, skills, and abilities
     and identify what additional technical expertise is needed. 
     Develop the technical skills needed to allow OCSE to become more
     actively involved with the states at critical points in their
     development processes, and enhance the skills of existing
     systems reviewers through additional training.  This expertise
     should include program management, software development, and
     systems engineering. 

  -- Conduct timely post-implementation reviews on certified child
     support systems to determine whether they are providing expected
     benefits, identify any lessons learned, and assess innovative
     technical solutions. 

  -- At least annually assess the progress of child support systems
     projects nationwide to gain and share with the states a broader
     perspective on costs, systemic problems, potential solutions,
     and innovative approaches.  Information should be shared with
     other states to help reduce costs and improve effectiveness of
     the child support program nationally--especially any practices
     or systems that could benefit states attempting to develop or
     implement welfare reform systems requirements. 

  -- Assess the impact of welfare reform on existing child support
     programs--including automated systems and business
     operations--and determine whether states will be able to
     implement systems requirements within established time frames
     and without exceeding the $400 million cap.  This assessment
     should also include an estimate of additional regular rate
     funding for automated systems that states may need to comply
     with the requirements of welfare reform. 

  -- Provide the states with technical requirements for implementing
     welfare reform systems, including the new-hire, central case,
     centralized collection, and disbursement registries in enough
     time to allow the states to meet the legislatively mandated
     deadlines of October 1997, 1998, and ultimately 2000. 


   AGENCY COMMENTS AND OUR
   EVALUATION
---------------------------------------------------------- Chapter 6:2

HHS disagreed with our recommendations on agency monitoring and
oversight and suspending federal funding for flawed state systems,
while generally agreeing with our other recommendations.  HHS
officials' primary concern with the report was the degree of federal
stewardship appropriate in the effective development of automated
state child support enforcement systems.  The Department notes that
we have a different perception of the appropriate federal role in
state automated systems development than is authorized.  The
Department indicated that reviewing state systems at critical phases
would increase the administrative burden on the states, and result in
OCSE's "micromanagement" of state projects.  Further, officials
reiterated their belief that withholding funds was counterproductive
to developing automated systems.  Finally, HHS expressed concern
about our presentation of the level of state automation and systems
costs.  The Department did, however, generally agree with our
recommendations regarding assessing OCSE's technical resources,
conducting post-implementation and nationwide systems reviews, and
defining--in a timely manner--welfare reform requirements. 

We have reviewed HHS' comments, and found no reason to change our
conclusions and recommendations.  HHS views OCSE's role narrowly, as
one of monitoring requirements and approving funds.  The agency
believes its role is to assist states in meeting mandated deadlines,
rather than more actively monitoring and overseeing state systems
development activities with an eye towards helpful intervention.  We
disagree with HHS' approach.  Instead of placing the responsibility
solely with the states, OCSE is also accountable for effective state
systems development.  According to statutory requirements, the agency
should review, assess, and inspect systems throughout development. 
Given the significance of state child support enforcement systems to
the operation of the program and the magnitude of expenditures, it is
critical that these systems be developed correctly and efficiently
from the beginning.  A philosophy of providing funds, despite serious
systems problems and costly mistakes, misses the opportunity to
reduce the risk of systems failure and save taxpayer dollars.  OCSE
does not evaluate or assess states' systems development projects
using a disciplined, structured approach.  The agency's reviews are
narrowly focused and, as a result, not effective or timely in
assessing the states' systems approaches and progress.  A summary of
the Department's comments and our evaluation is provided below.  HHS'
comments are reprinted in appendix IV of this report. 


      HHS' ROLE NARROWLY FOCUSED
-------------------------------------------------------- Chapter 6:2.1

We do not agree with HHS' position that OCSE's role be primarily
focused on providing technical assistance and guidance.  HHS'
statutory responsibilities, as set forth in the Social Security Act,
delineate a leadership role in developing child support enforcement
systems.  Section 452(a) of this act provides that a "designee of the
Secretary" (Office of Child Support Enforcement) shall, .  .  . 
"review and approve state plans" for child support enforcement
programs, "establish standards" for state programs, "review and
approve state plans," and "evaluate the implementation of state
programs." With regard to child support management information
systems, Section 452(d) provides that OCSE shall, "on a continuing
basis, review, assess, and inspect the planning, design, and
operation of management information systems .  .  .  with a view to
determining whether, and to what extent, such systems meet and
continue to meet requirements imposed [under the act]."

The agency's advance planning document (APD) guide,\1 also refers to
its leadership role in approving, monitoring, and certifying state
systems programs to ensure that federal expenditures are made wisely. 
In HHS' response to our report, it noted elsewhere that the agency
"has the authority, which it frequently exercises, to require states
to send an as-needed APD at critical milestones in its life cycle
methodology." And in response to comments from states concerning the
extent of OCSE's reviews, HHS said that it intends to continue
monitoring state systems projects, noting that it has
"responsibilities for assuring that the expenditure of federal funds
on state systems is necessary for the effective and efficient
operation of the programs."\2 This is consistent with recent
legislation that requires more effective oversight of systems
development activities.  The Clinger-Cohen Act of 1996\3 and recent
Office of Management and Budget guidance for managing information
technology investments\4 specify the need for greater accountability
for systems during the critical development and implementation
phases. 

The federal government is a major stakeholder in these systems,
paying about $2 billion dollars over the last 15 years.  As such, it
is critical that OCSE's current approach to monitoring and overseeing
state systems be improved to ensure that the federal government's
investment in systems is spent wisely.  Our recommendations reflect
systems development practices that are widely used in both the
private and public sectors.  Monitoring at critical points in the
development process allows earlier intervention and greater
opportunity to correct problems before they become more costly.  We
disagree that this approach constitutes micromanagement; we believe
that to do less constitutes lax management.  These activities also
need not create an administrative burden.  OCSE does not have to
impose additional reporting requirements on the states; it must
simply streamline its existing reporting process to ensure the
inclusion of key pieces of information at critical phases.  It is
imperative that HHS take advantage of its legislatively authorized
oversight and monitoring role.  In the absence of such action, HHS is
likely to continue to provide little added value to states; instead,
it will remain merely a bureaucratic hurdle for states to climb to
fund critically important systems. 

HHS asserts that OCSE should provide technical assistance to the
states, rather than suspend federal funding--especially given the
statutory deadline.  We disagree.  Federal regulations provide for
the suspension of federal funding when states' systems under
development cease to substantially comply with requirements and other
provisions of the APD.\5 Further, irrespective of the deadline,
allowing systems to be developed ineffectively and inefficiently at
the expense of the taxpayers is not supporting the goals and
underlying intent of the legislation.  Allowing a state to go forward
before correcting inadequacies in approach contributes to rising
systems' costs. 

In this report and in our 1992 report, we pointed out that OCSE
continues to fund systems with serious problems--problems that
threaten their very success.\6 As discussed in this report, such an
approach invites the need to correct serious problems later in the
development process, when it is more costly and time-consuming to do
so. 

OCSE has periodically suspended state funding for automation
projects.  As we reported, however, almost 60 percent of these
disruptions were due to insufficient information on the required APD,
or for states' exceeding their authorized funding levels--not for
more substantive issues on the soundness of the development approach
itself.  Even when funding was held up for major systems-related
problems, efforts to correct these problems did not appear to be made
in a timely fashion.  Further, in cases in which an HHS regional
official suggested that OCSE hold up funding for a project, the
agency did not stop funding until the project "crashed."


--------------------
\1 State Systems APD Guide, Administration for Children and Families
and Health Care Finance Administration, Department of Health and
Human Services (September 1996). 

\2 The Federal Register, vol.  61, no.  148 (July 31, 1996). 

\3 This act requires, among other things, that agencies manage risks
associated with information technology projects from initiation to
completion. 

\4 Evaluating Information Technology Investments, OMB Office of
Information and Regulatory Affairs, Information Policy and Technology
Branch, November 1995. 

\5 45 CFR 307.40. 

\6 GAO/IMTEC-92-46, Aug.  13, 1992. 


      WEAKNESSES IN OCSE'S CURRENT
      PROCESS
-------------------------------------------------------- Chapter 6:2.2

HHS agreed that the states are encountering automation problems and
that states need a greater degree of oversight.  The Department
stated that OCSE follows a structured approach in reviewing states
APD submissions.  We believe this approach is not sufficient
oversight.  While we described the agency's review process--APD and
certification reviews, we believe that the APD process should ensure
that systems development activities at critical decision points are
evaluated.  OCSE does not consistently monitor state systems
development at critical milestone points, such as the completion of
design or requirements development.  We noted that OCSE's
certification reviews are usually conducted toward the end of the
development process, and as such are often too late to help identify
problems and redirect the approach. 


      STRUCTURED METHODOLOGY
      SUPPORTS SYSTEMS VARIABILITY
-------------------------------------------------------- Chapter 6:2.3

Another issue raised by HHS was that with a variety of concurrent
systems development activities, including the staggering welfare
reform deadlines, it may be difficult to use our suggested "one
structured methodology fits all." We are certainly not advocating
that OCSE require or impose a "one structured methodology fits all"
approach.  A structured approach to reviewing systems
development--irrespective of the particular methodology used--would
also allow for systems variability, including the differences in
project size, scope, and complexity.  The key to a structured
approach is the identification of critical milestones that are the
basis for systems reviews.  States value this process; one state
official noted that its project would be unmanageable without it. 
Other state officials told us that OCSE should play a more active
role, and that they considered it important that the agency assess
the management and direction of the project early to avoid or
minimize later problems. 


      POST-IMPLEMENTATION REVIEWS
      PLANNED
-------------------------------------------------------- Chapter 6:2.4

HHS concurred with our recommendation to conduct post-implementation
reviews on certified child support systems.  The agency has plans in
fiscal year 1998 to conduct both technical assistance visits and
post-implementation reviews to further identify lessons learned and
assess innovative technical solutions. 


      NATIONWIDE ASSESSMENT
      INITIATED
-------------------------------------------------------- Chapter 6:2.5

HHS agreed on the importance of a nationwide assessment of child
support systems projects and has taken steps in that direction.  The
Department noted that it has developed the State System Approval
Information System (SSAIS) and uses electronic means for information
sharing.  However, it is critical that OCSE continue to maintain and
use accurate information from the SSAIS and develop a sound
nationwide basis for encouraging states to share innovative database
designs, software, and other technologies for greater efficiencies
and costs savings, and for identifying recurring problems.  With a
systematic and comparison-based assessment, OCSE could recognize
trends and identify best practices that could be shared.  Identifying
and taking action on such issues would be a significant benefit of
increased oversight. 


      MAGNITUDE OF WELFARE REFORM
-------------------------------------------------------- Chapter 6:2.6

HHS agreed on the importance of providing states with technical
requirements associated with recent welfare reform legislation.  As
HHS noted, states' systems funding was not limited to the $400
million enhanced funding.\7 We recognize that the welfare reform
requirements are substantial, and that the legislation will allow
states to be reimbursed at the regular and enhanced rate.  The
monetary magnitude of accommodating welfare reform systems
requirements further underscores the importance of effective federal
oversight, including comprehensive assessments of systems
implications and timely issuance of systems requirements. 


--------------------
\7 For the enhanced funding, welfare legislation provides 80-percent
reimbursement from the federal government. 


      NONCERTIFIED SYSTEMS SHOW
      PARTIAL BENEFITS; YET
      SYSTEMS COSTS ARE
      SIGNIFICANT
-------------------------------------------------------- Chapter 6:2.7

HHS believes our report incorrectly noted that noncertified states
have "no automation" to enforce child support collections.  We
acknowledge that state systems that are not yet certified may have
some automation.  In fact, we noted that state officials indicated
that partially automated systems have improved their capability to
locate noncustodial parents, increased paternity establishment and
collections, and provided greater staff efficiency.  We also noted,
however, that OCSE officials said that as many as 14 states may not
meet the deadline for certification, leaving about 44 percent of the
national caseload without the full benefits of automation. 

As envisioned by the Congress in implementing this legislation, some
states have attained benefits; however, child support enforcement
systems costs continue to increase, and the extent of final costs is
not yet known.  We acknowledge that systems' costs include developing
new systems and maintaining and enhancing systems that were certified
prior to 1988.  OCSE, however, does not track these costs separately. 
Some states did not design new systems; rather, they built upon
existing ones.  In these cases, the costs attributable to the 1988
act would be less than those of states that developed entirely new
systems.  However, states that updated their existing systems may
now, as the October 1, 1997, deadline draws near, need to
significantly redesign their systems to fully meet child support
certification requirements and support welfare reform legislation. 

HHS indicated that the costs of automation should be carefully placed
in perspective and also compared automation costs to the agency's
administrative costs.  We recognize that systems costs may be a small
percentage of the total administrative costs; however, we do not
believe that $2.6 billion is inconsequential.  We recognize and cite
examples where OCSE's weak oversight contributed to the rising costs
that could have been avoided if more of a proactive leadership role
was demonstrated by the agency.  These systems will play a critical
role in effectively administering the child support and welfare
programs in the future.  As such, it is incumbent upon the Department
and OCSE to ensure that the dollars invested in these systems are
spent wisely, and provide an effective return on investment. 


TOTAL COST FOR EACH STATE'S CHILD
SUPPORT ENFORCEMENT SYSTEM AND THE
FEDERAL AND STATE SHARES OF THESE
EXPENDITURES FOR FISCAL YEARS
1981-1996
=========================================================== Appendix I

                                                                 Total
State                      Federal share    State share   expenditures
-------------------------  -------------  -------------  =============
Alabama                      $49,222,159    $13,253,068    $62,475,227
Alaska                        11,071,718      4,653,294     15,725,012
Arizona                       47,063,019      9,509,389     56,572,408
Arkansas                      21,886,630      5,673,408     27,560,038
California                   248,299,485     95,658,921    343,958,406
Colorado                      26,907,957      5,005,726     31,913,683
Connecticut                   13,981,700      4,623,786     18,605,486
Delaware                       8,010,878      3,238,003     11,248,881
District of Columbia           8,896,407      2,059,990     10,956,397
Florida                       60,114,807     22,418,473     82,533,280
Georgia                       41,611,875     12,697,507     54,309,382
Guam                           1,357,280        179,404      1,536,684
Hawaii                        15,070,420      3,968,630     19,039,050
Idaho                         26,870,077      8,052,556     34,922,633
Illinois                      64,753,779     24,385,762     89,139,541
Indiana                       27,479,878      6,914,774     34,394,652
Iowa                          31,310,719      7,975,500     39,286,219
Kansas                        24,449,198      7,407,991     31,857,189
Kentucky                      30,360,881      7,996,230     38,357,111
Louisiana                     14,004,464      2,542,250     16,546,714
Maine                         13,446,950      3,246,462     16,693,412
Maryland                      31,432,325      7,420,924     38,853,249
Massachusetts                 33,751,067      7,409,793     41,160,860
Michigan                      66,577,358     10,713,170     77,290,528
Minnesota                     56,661,743     18,197,130     74,858,873
Mississippi                   16,031,897      3,943,213     19,975,110
Missouri                      44,337,248     13,204,378     57,541,626
Montana                        9,324,373      2,743,493     12,067,866
Nebraska                      33,942,309     10,731,967     44,674,276
Nevada                        17,107,631      4,001,180     21,108,811
New Hampshire                 18,451,398      4,930,405     23,381,803
New Jersey                    64,364,051     23,196,922     87,560,973
New Mexico                    23,169,261      6,372,763     29,542,024
New York                      96,145,382     34,290,071    130,435,453
North Carolina                43,056,136      9,723,106     52,779,242
North Dakota                   3,753,318      1,059,255      4,812,573
Ohio                          55,426,872     18,155,838     73,582,710
Oklahoma                      28,279,064      9,265,488     37,544,552
Oregon                        15,168,684      4,015,436     19,184,120
Pennsylvania                  79,416,821     31,326,468    110,743,289
Puerto Rico                   22,882,169      4,506,891     27,389,060
Rhode Island                  13,727,135      2,863,771     16,590,906
South Carolina                41,414,812     15,273,044     56,687,856
South Dakota                   3,427,507        938,347      4,365,854
Tennessee                     33,337,753     11,458,018     44,795,771
Texas                        123,790,741     51,140,939    174,931,680
Utah                          32,295,631     11,058,131     43,353,762
Vermont                        3,291,186        949,286      4,240,472
Virginia                      66,684,309     25,108,768     91,793,077
Virgin Islands                 5,132,533        818,705      5,951,238
Washington                    58,487,304     24,186,599     82,673,903
West Virginia                 17,011,816      4,625,153     21,636,969
Wisconsin                     63,489,909     18,934,783     82,424,692
Wyoming                        8,824,173      1,552,959     10,377,132
======================================================================
Total                      $2,016,364,19   $645,577,518  $2,661,941,71
                                       7                             5
----------------------------------------------------------------------
Source:  HHS.  GAO did not independently verify this information. 


TOTAL FEDERAL COST FOR EACH
STATE'S CHILD SUPPORT ENFORCEMENT
SYSTEM INCLUDING ENHANCED AND
REGULAR FUNDING FOR FISCAL YEARS
1981-1996
========================================================== Appendix II

                                                         Total federal
                                Enhanced        Regular  expenditures\
State                      federal share  federal share              a
-------------------------  -------------  -------------  =============
Alabama                      $33,760,795    $15,461,364    $49,222,159
Alaska                         2,118,686      8,953,032     11,071,718
Arizona                       36,906,092     10,156,926     47,063,019
Arkansas                      13,561,791      8,324,839     21,886,630
California                    87,823,438    160,476,046    248,299,485
Colorado                      23,468,483      3,439,474     26,907,957
Connecticut                    7,945,490      6,036,209     13,981,700
Delaware                       3,576,863      4,434,015      8,010,878
District of Columbia           6,246,565      2,649,843      8,896,407
Florida                       20,889,857     39,224,951     60,114,807
Georgia                       21,425,098     20,186,777     41,611,875
Guam                           1,286,507         70,773      1,357,280
Hawaii                        10,262,676      4,807,744     15,070,420
Idaho                         20,073,452      6,796,625     26,870,077
Illinois                      20,890,710     43,863,070     64,753,779
Indiana                       21,568,032      5,911,845     27,479,878
Iowa                          21,720,806      9,589,913     31,310,719
Kansas                        12,465,611     11,983,587     24,449,198
Kentucky                      20,287,100     10,073,780     30,360,881
Louisiana                     11,584,227      2,420,237     14,004,464
Maine                          8,997,291      4,449,659     13,446,950
Maryland                      22,376,517      9,055,808     31,432,325
Massachusetts                 24,232,685      9,518,382     33,751,067
Michigan                      58,297,798      8,279,560     66,577,358
Minnesota                     30,840,617     25,821,126     56,661,743
Mississippi                   11,533,446      4,498,451     16,031,897
Missouri                      30,714,110     13,623,138     44,337,248
Montana                        5,348,540      3,975,833      9,324,373
Nebraska                      21,982,444     11,959,866     33,942,309
Nevada                        13,836,986      3,270,645     17,107,631
New Hampshire                 11,231,502      7,219,896     18,451,398
New Jersey                    23,567,719     40,796,332     64,364,051
New Mexico                    18,715,453      4,453,809     23,169,261
New York                      36,806,611     59,338,771     96,145,382
North Carolina                34,289,025      8,767,111     43,056,136
North Dakota                   2,370,153      1,383,166      3,753,318
Ohio                          25,313,569     30,113,303     55,426,872
Oklahoma                      15,366,150     12,912,914     28,279,064
Oregon                         9,295,461      5,873,224     15,168,684
Pennsylvania                  25,242,187     54,174,634     79,416,821
Puerto Rico                   19,873,039      3,009,130     22,882,169
Rhode Island                  10,483,904      3,243,231     13,727,135
South Carolina                20,183,220     21,231,592     41,414,812
South Dakota                   2,313,346      1,114,161      3,427,507
Tennessee                     19,932,879     13,404,874     33,337,753
Texas                         31,166,306     92,624,435    123,790,741
Utah                          14,285,463     18,010,168     32,295,631
Vermont                        1,892,442      1,398,744      3,291,186
Virginia                      21,992,655     44,691,655     66,684,309
Virgin Islands                 4,514,866        617,667      5,132,533
Washington                    12,529,291     45,958,013     58,487,304
West Virginia                 10,565,333      6,446,482     17,011,816
Wisconsin                     37,614,490     25,875,419     63,489,909
Wyoming                        8,349,121        475,052      8,824,173
======================================================================
Total                      $1,043,916,89   $972,447,302  $2,016,364,19
                                       4                             7
----------------------------------------------------------------------
\a Some totals may not add due to rounding of component figures. 

Source:  HHS.  GAO did not independently verify this information. 


OVERVIEW OF SYSTEMS DEVELOPMENT
PHASES
========================================================= Appendix III



   (See figure in printed
   edition.)

Source:  Adapted from Roger F.  Pressman, Ph.D., Software
Engineering:  A Practitioner's Approach (New York:  McGraw-Hill,
1992). 




(See figure in printed edition.)Appendix IV
COMMENTS FROM THE DEPARTMENT OF
HEALTH AND HUMAN SERVICES
========================================================= Appendix III



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)



(See figure in printed edition.)


MAJOR CONTRIBUTORS TO THIS REPORT
=========================================================== Appendix V

ACCOUNTING AND INFORMATION
MANAGEMENT DIVISION, WASHINGTON,
D.C. 

Christie M.  Motley, Assistant Director
Leonard J.  Latham, Technical Assistant Director
Robert F.  Gerkin, Business Process Analyst
Norman F.  Heyl, Business Process Analyst
Gwendolyn M.  Adelekun, Business Process Analyst
Michael P.  Fruitman, Communications Analyst
Sharon O.  Byrd, Senior Auditor

ATLANTA REGIONAL OFFICE

Carl L.  Higginbotham, Senior Computer Specialist
Amanda C.  Gill, Staff Evaluator

DENVER REGIONAL OFFICE

Yvonne J.  Vigil, Information Systems Analyst

*** End of document. ***