Child Support Enforcement: Early Results on Comparability of Privatized
and Public Offices (Letter Report, 12/16/96, GAO/HEHS-97-4).

Growing caseloads and resource limitations have led some states to
privatize their child support enforcement offices. Fifteen states have
turned to privatization as a way to boost performance and handle
caseloads that, in some instances, are approaching 1,000 per worker. For
some offices, privatization has also been a response to state
restrictions on hiring additional public employees. In a series of
comparisons, GAO found that fully privatized enforcement offices
performed as well or better than public child support programs in
locating noncustodial parents and collecting payments. However, these
results are limited to the cases GAO reviewed and do not reflect the
performance of public or private offices overall within the states
selected. Moreover, because the full-service privatization of child
support enforcement is relatively new, the extent to which it offers
comparable performance and cost-effectiveness remains open for
evaluation over the long term.

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

 REPORTNUM:  HEHS-97-4
     TITLE:  Child Support Enforcement: Early Results on Comparability 
             of Privatized and Public Offices
      DATE:  12/16/96
   SUBJECT:  Child support payments
             Privatization
             Cost effectiveness analysis
             Collection procedures
             State-administered programs
             Tax information confidentiality
             Tax refunds
             Federal/state relations
             Administrative costs
             Law enforcement
IDENTIFIER:  Arizona
             Tennessee
             Virginia
             Aid to Families with Dependent Children Program
             AFDC
             HHS Temporary Assistance for Needy Families Program
             Hampton (VA)
             Portsmouth (VA)
             Yavapai County (AZ)
             Mohave County (AZ)
             Knoxville (TN)
             Nashville (TN)
             IRS Project 1099
             IRS Refund Offset Program
             HHS Child Support Enforcement Program
             
******************************************************************
** 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 the Chairman, Committee on the Budget, House of
Representatives

December 1996

CHILD SUPPORT ENFORCEMENT - EARLY
RESULTS ON COMPARABILITY OF
PRIVATIZED AND PUBLIC OFFICES

GAO/HEHS-97-4

Full-Service Child Support Privatization

(106606)


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

  AFDC - Aid to Families With Dependent Children
  FPLS - Federal Parent Locator Service
  HHS - Department of Health and Human Services
  IRS - Internal Revenue Service
  OCSE - Office of Child Support Enforcement
  TANF - Temporary Assistance for Needy Families

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


B-270419

December 16, 1996

The Honorable John R.  Kasich
Chairman, Committee on the Budget
House of Representatives

Dear Mr.  Chairman: 

Since 1984, state child support enforcement programs have been
challenged by a 150-percent increase in the demand for services.  In
fiscal year 1995, the number of cases rose to about 20 million and
collections of child support reached a record high of nearly $11
billion.  While faced with this unprecedented workload, child support
programs continue to confront state budgetary constraints and
increasing federal requirements for mandated services under a series
of child support reforms.  For example, the recent welfare reform
legislation--the Personal Responsibility and Work Opportunity
Reconciliation Act of 1996 (P.L.  104-193, Aug.  22, 1996)--requires
states to establish automated registries of child support orders and
directories of newly hired employees to track and locate parents
owing support.  To meet these growing demands, many states are moving
to privatize child support enforcement services that have
traditionally been delivered by the public sector.  In some places,
states have contracted with private sector firms to assume all local
child support services, which we refer to as full-service
privatization. 

This report responds to your request for information on states'
efforts to fully privatize local offices.  Specifically, it addresses
(1) states' rationale for full-service privatization; (2) how the
performance and cost- effectiveness of full-service privatization
efforts compare with publicly managed child support enforcement; and
(3) what, if any, issues could affect future full-service
privatization contracts. 

To determine states' rationale for privatization, we interviewed
child support officials in the 12 states with 21 local full-service
privatization contracts identified in our November 1995 report.\1 To
analyze the performance and cost-effectiveness of full-service
privatization, we reviewed four fully privatized local offices in
three states--Arizona, Tennessee, and Virginia.  Specifically, we
examined the cost-effectiveness of each of the offices and for a set
of new cases we reviewed the degree of success each had in locating
noncustodial parents, establishing paternity and support orders, and
collecting support.  In each state, we compared a privatized office
with a similar public office; additionally, we compared pre- and
postprivatization outcomes at one other office.  We analyzed the
cost-effectiveness in all four office comparisons, but because of
data limitations, we analyzed performance in only three of the office
comparisons.  Appendix I provides a more detailed explanation of our
methodology, analysis methods, and results.  Finally, to identify
what, if any, issues affect full-service privatization, we
interviewed federal Office of Child Support Enforcement (OCSE) and
Internal Revenue Service (IRS) officials, as well as state and
contractor officials in the three states reviewed. 

Our results represent the performance of these offices on a small set
of cases for an abbreviated time period and do not necessarily
represent long-term public and private performance overall within
these states. 


--------------------
\1 Child Support Enforcement:  States and Localities Move to
Privatized Services (GAO/HEHS-96-43FS, Nov.  20, 1995). 


   RESULTS IN BRIEF
------------------------------------------------------------ Letter :1

Fifteen states have turned to full-service privatization of selected
local child support enforcement offices as a way to improve
performance and handle growing caseloads that are reaching or
exceeding 1,000 cases per worker in some instances.  For some
offices, privatization has also been a response to state restrictions
on hiring additional public employees. 

In the three comparisons of performance we conducted, fully
privatized offices performed at least as well as or, in some
instances, better than public child support programs in locating
noncustodial parents, establishing paternity and support orders, and
collecting support owed.  For example, a privatized office in
Tennessee established paternity for 4 percent of the cases needing
this service in the first year of privatization compared with about 3
percent in the last year before it was privatized.  In Virginia, the
privatized office collected support payments from 41 percent of the
cases we reviewed, a rate almost twice that of the public office we
compared it with. 

The relative cost-effectiveness\2 of the privatized versus public
offices, however, differed among the comparisons we made. 
Specifically, Virginia's and Arizona's privatized offices were more
cost-effective--60 percent and 18 percent, respectively--than their
public counterparts.  However, in Tennessee, one public office was 52
percent more cost-effective than the privatized office we reviewed,
while the remaining privatized office in Tennessee was about as
cost-effective as its public counterpart. 

According to state and contractor officials, differences in
performance and cost-effectiveness among private and public offices
may have resulted from the increased flexibility contractors have in
acquiring resources and managing staff, contractors' greater access
to technology, differences in the complexity of the caseloads, and
varying payment rates to contractors for child support enforcement
services. 

An issue of contractor access to IRS data that could have impeded
future full-service privatization has been partially addressed by
recent welfare reform legislation.  The Personal Responsibility and
Work Opportunity Reconciliation Act of 1996 authorizes state child
support agencies to disclose to contractors certain, but not all,
restricted tax data that are useful in locating parents and enforcing
payment. 


--------------------
\2 We defined cost-effectiveness as the ratio of each office's
administrative costs to collections, expressed as the cost to collect
$1. 


   BACKGROUND
------------------------------------------------------------ Letter :2

In 1975, the Congress created the federal child support enforcement
program as title IV-D of the Social Security Act.  The program's
purpose is to strengthen state and local child support enforcement
efforts for obtaining child support for families who receive Aid to
Families With Dependent Children (AFDC)\3 benefits and for any
non-AFDC individuals who apply for services.  Services provided to
these clients include locating noncustodial parents, establishing
paternity and child support orders, and collecting support owed.\4
Appendix II contains a glossary of child support enforcement
services. 

Child support enforcement is a joint federal and state
responsibility.  Within the federal government, OCSE, in the
Department of Health and Human Services (HHS), is responsible for
providing leadership, technical assistance, and standards for
effective state programs.  States or local offices under state
supervision deliver child support services to families.  The federal
government and the states share program costs, including contract
costs, at the rate of 66 percent and 34 percent, respectively.  In
1995, administrative costs for the program were $3.1 billion and
collections totaled $10.8 billion.  About 19 percent of the cases in
the child support program nationwide received a payment in 1995. 

To help meet the demands of growing caseloads in an environment of
resource constraints and increasing federal requirements, some states
have turned to privatization.  In our November 1995 report to you, we
provided information on states' privatization of child support
enforcement services and identified 37 states that had privatized or
planned to privatize portions of their child support caseloads.  Some
of these states chose to augment their public programs by privatizing
particular services, such as locating noncustodial parents or
collecting support owed,\5 while 12 others looked to private sector
contractors to assume all local child support services.  Under
full-service privatization, states contract out all or most of the
services traditionally performed by a local public entity, usually a
county or judicial district.  Additionally, since our November 1995
report, Ohio, South Carolina, and West Virginia have initiated
full-service privatization efforts.  Figure 1 shows the states in
which we have identified fully privatized local offices as of
September 1996.  Appendix III provides further information on each of
these fully privatized offices, including the local jurisdictions
they serve, the contractor, and the contract length and its payment
terms. 

   Figure 1:  States With Planned
   and Ongoing Fully Privatized
   Local Offices as of September
   1996

   (See figure in printed
   edition.)

Most contracts for full-service privatization provide for the
contractor to be paid an amount equal to a percentage of the
collections generated.  Therefore, as a contractor's collections
increase or decrease, the contractor's payment--or the administrative
costs paid to operate the office--rises or falls proportionately.  In
our November 1995 report, we noted payment rates in effect ranging
from about 10 percent to 32 percent of collections.  Some contracts
provide for variable rates that decline over the life of the
contract.  In these cases, a higher initial payment rate is intended
to help the contractor defray start-up costs.  Most contracts provide
for 2 to 5 years of service. 


--------------------
\3 As of July 1, 1997, AFDC will be replaced by block grants under
the Temporary Assistance for Needy Families (TANF) program. 

\4 Other services include client intake, customer service,
enforcement, and updating and adjusting support orders. 

\5 For further information on states' experiences with private
agencies' collection of child support payments, see Child Support
Enforcement:  States' Experience with Private Agencies' Collection of
Support Payments (GAO/HEHS-97-11, Oct.  23, 1996). 


   STATES PRIVATIZE TO IMPROVE
   CHILD SUPPORT SERVICES, HANDLE
   GROWING CASELOADS, AND OBTAIN
   ADDITIONAL RESOURCES
------------------------------------------------------------ Letter :3

Overall, officials in the states with full-service privatization
efforts most frequently cited a desire to improve the child support
services offered, the need to serve their soaring caseloads, and the
ability to deploy additional child support staff as reasons why they
fully privatized local offices.  At two offices, state officials
estimated that caseloads had reached about 1,000 per worker, and they
could not get the authority to hire additional public staff.  Other
major reasons cited by officials include the local child support
offices' difficulties in meeting increasing federal program
requirements; legislative or executive directives mandating
privatization as a way to improve state performance; and local public
entities, such as the district attorney, ceasing to provide child
support services when faced with rising caseloads, staffing
constraints, and additional federal requirements. 

State officials in the three states we reviewed cited similar
rationales for privatizing the four offices we reviewed within their
states.  In Virginia and Arizona, officials cited the need to serve
growing caseloads.  Arizona officials also wanted to provide better
service to areas previously considered underserved.  At both the
offices we examined in Tennessee, state officials eventually turned
to privatization after the local district attorneys no longer wanted
to operate the local child support offices.  State officials had been
encouraging the district attorney at one office to improve office
performance.  Both district attorneys believed that, without
additional staff, their offices could no longer meet the federal
performance standards and, consequently, both terminated their child
support services.  Faced with the loss of the governmental service
providers, the state contracted out the operation of both child
support offices. 

Some states have implemented full-service privatization in a way that
minimizes displacement of public employees.  In Virginia and Arizona,
existing districts were subdivided to create new offices in February
and March 1994, respectively.  Creating new privatized offices in an
existing district is considered easier than replacing an established
office because no public workers are displaced, according to state
officials.  At both Tennessee offices, which were privatized in July
1992 and July 1993, the public workers lost their jobs, but according
to state officials, most were offered positions with the private
contractor. 


   OUTCOMES ARE COMPARABLE, BUT
   COST-EFFECTIVENESS VARIED
------------------------------------------------------------ Letter :4

Our analysis suggests that fully privatized offices can produce
performance outcomes comparable to those of public child support
programs.  In outcome measures such as locating noncustodial parents,
establishing paternity and support orders, and obtaining collections
for the cases we reviewed,\6 the privatized offices in Arizona and
Tennessee performed about as well as their public counterparts and
the one in Virginia did significantly better.  While performance
outcomes show that the privatized offices did at least as well or
better than their public counterparts, the cost-effectiveness results
were more mixed for the periods reviewed. 


--------------------
\6 In Tennessee, these results reflect the performance of a selected
office before and after privatization.  We did not compare the
performance outcomes of our paired offices in Tennessee because of
the relatively small number of cases that met our criteria for
inclusion in our analysis (see app.  I for a description of the kinds
of cases excluded from analysis).  We did, however, develop
cost-effectiveness data for these offices. 


      VIRGINIA'S PRIVATIZED OFFICE
      MATCHED OR EXCEEDED PUBLIC
      OFFICE OUTCOMES AND WAS MORE
      COST-EFFECTIVE
---------------------------------------------------------- Letter :4.1

In Virginia, the privately run office that we studied in Hampton
performed as well as or better than the public office in Portsmouth
that we compared it with.  These offices serve similar client
populations in suburban and rural areas in the same part of the
Tidewater area of Virginia.  As table 1 illustrates, the Hampton
office established paternity and support orders and made collections
for higher percentages of the selected cases than did the Portsmouth
office during our 18-month review period.  The difference for
collections was statistically significant after controlling for
differences in caseload characteristics between the offices, such as
the percentage of AFDC cases and the services that cases needed,
while the differences for establishing paternity and support orders
approached statistical significance.\7

The difference in location rates was not statistically significant. 



                          Table 1
          
          Virginia: Performance Outcomes Compared
            for 18 Months on New Cases Opened in
                         July 1994

                               Of cases needing service,
                                percent with successful
                                       outcome\a
                            -------------------------------
                            Private office    Public office
Service needed                   (Hampton)     (Portsmouth)
--------------------------  --------------  ---------------
Location\b                            73.9             58.5
Paternity establishment               40.0             19.2
Support order                         35.7             16.8
 establishment
Collection                            40.8             22.3
-----------------------------------------------------------
\a Except for location rates, the differences between offices were
statistically significant or approached significance after
controlling for the percentage of cases receiving AFDC and needing
specific services (see app.  I). 

\b Actual location rates for both the public and private offices may
be understated because the only information available for analysis
was whether the office had a valid mailing address for an individual
at the end of the 18-month study period.  Therefore, because some
individuals may have been successfully located at some point during
the 18-month period but may not have had a valid address at the end
of the period, our data may underestimate actual location rates. 

With regard to the cost-effectiveness of the two offices, Hampton's
cost to collect $1 of support was 60 percent lower than that for
Portsmouth during the period reviewed.\8 \9 \10 As table 2 shows, the
contractor was paid 11.5ï¿½ for each $1 collected, while the public
office spent 18.4ï¿½.  Also, according to the contract, the
contractor's payment will decline to 9.95ï¿½ per $1 collected in the
final year of the 5-year contract. 



                          Table 2
          
          Virginia: Administrative Costs Compared
             to Collections, July 1994 through
                       December 1995

                            Private office    Public office
                                 (Hampton)     (Portsmouth)
--------------------------  --------------  ---------------
Administrative costs            $1,791,733       $2,238,482
Collections                    $15,553,480      $12,197,214
Cost to collect $1                   11.5ï¿½            18.4ï¿½
-----------------------------------------------------------
Additionally, to determine whether privatization resulted in
disproportionate increases in state costs, we compared administrative
costs before and after privatization.  In Virginia, the privatized
office's administrative costs did not increase dramatically after
privatization.  In state fiscal year 1994, the first year of
privatization, the privatized office's administrative costs grew by
less than 1 percent over the previous year's imputed costs.\11

In the second year after privatization, the privatized office's
administrative costs increased by about 22 percent, a growth rate
similar to the public office's 21-percent increase. 


--------------------
\7 Statistical significance indicates that if the cases we analyzed
represent a statistical sample of all new cases handled by the
offices, there is a less than 5-percent chance of being wrong when
concluding that there is a difference between the offices.  We
considered the paternity and support order results to be approaching
statistical significance because they were associated with a slightly
greater than 5-percent risk (5.3 and 5.6 percent, respectively) of
this type of error. 

\8 Program cost-effectiveness can either be stated as the cost to
collect $1 of child support or as the amount of child support
collected for each $1 spent on the program.  Because contractors are
generally paid a percentage of collections, we chose to use the
cost-to-collect-$1 method. 

\9 While we controlled for the percentage of the cases that received
AFDC in our comparison of the performance of pairs of offices, we did
not do so for the analysis of overall cost-effectiveness. 
Nevertheless, before analysis, the office pairs were matched on the
basis of several criteria, including AFDC caseload. 

\10 For public offices, we asked state officials for administrative
costs incurred solely at the local level.  For privatized offices,
administrative costs to the state were defined as the payment to the
contractor.  This excludes some applicable administrative costs such
as those incurred to contract for the services and to subsequently
monitor them. 

\11 We imputed administrative costs for Hampton because its caseload
was drawn from the caseload of its parent office, Newport News. 


      ARIZONA'S PRIVATIZED OFFICE
      PERFORMED AS WELL AS PUBLIC
      OFFICE AND WAS MORE
      COST-EFFECTIVE
---------------------------------------------------------- Letter :4.2

In Arizona, performance outcomes on our review cases revealed that
the privately run office in Yavapai County did about as well as the
public office in Mohave County that we compared it with; however, the
privatized office was more cost-effective during our 18-month review
period.  Although some differences in outcomes between the public and
private offices appear to exist (see table 3) none of the differences
was large enough to be statistically significant either before or
after controlling for AFDC status and kinds of services needed. 



                          Table 3
          
           Arizona: Performance Outcomes Compared
            for 18 Months on New Cases Opened in
                         July 1994

                                Of cases needing service,
                                 percent with successful
                                        outcome\a
                              -----------------------------
                              Private office  Public office
Service needed                     (Yavapai)       (Mohave)
----------------------------  --------------  -------------
Location\b                              37.5           22.9
Paternity establishment                  8.7            9.7
Support order establishment             14.3            5.8
Collection                              15.1            5.3
-----------------------------------------------------------
\a None of the differences between the public and private offices'
performance was found to be statistically significant either before
or after controlling for AFDC status and the kinds of services
needed. 

\b Location rates represent cases needing location services at case
opening and being located at some point during the review period. 

Regarding overall cost-effectiveness, the privatized office had a
lower cost per $1 collected during the period reviewed.  Mohave
County spent 34.8ï¿½ per $1 collected, while the contractor received
29.5ï¿½ per $1 collected for operating the Yavapai County program (see
table 4).  In addition, under the terms of the contract, the
contractor's payment rate will fall to 24.0ï¿½ for every $1 collected
in the final year of the 4-year contract. 



                          Table 4
          
           Arizona: Administrative Costs Compared
             to Collections, July 1994 through
                       December 1995

                              Private office  Public office
                                   (Yavapai)       (Mohave)
----------------------------  --------------  -------------
Administrative costs              $987,700\a     $1,296,092
Collections                       $3,387,792     $3,720,575
Cost to collect $1                   29.5ï¿½\a          34.8ï¿½
-----------------------------------------------------------
\a These cost figures reflect the payment to the contractor for
distributed collections of $3,348,948, not the collections figure
shown. 

We were unable to examine the rate of change in the privatized
office's administrative costs before and after privatization and
compare it with the public office's rate of change.  Arizona
officials could not break out preprivatization administrative costs
for Yavapai County only, which had been served at one time by a
multicounty unit.  Additionally, the officials believed that such a
comparison would be inappropriate, because Yavapai's cases had been
underserved before privatization. 


      TENNESSEE'S PRIVATIZED
      OFFICES SHOW MIXED RESULTS
---------------------------------------------------------- Letter :4.3

Performance and cost-effectiveness results were mixed in Tennessee,
where we reviewed two privatized offices.  In the first privatized
office, we did not have a sufficient number of cases to compare the
public and privatized offices' performance in locating noncustodial
parents, establishing paternity and support orders, and collecting
support owed.  However, we were able to compare the offices'
collections and administrative costs for the review period.  At this
office, our comparison of collections and administrative costs showed
that the public Fifth Judicial District office provided services at a
lower cost per $1 collected during the 18-month review period (see
table 5).  The public office cost was 9.9ï¿½ per $1 collected, while
the privatized office was paid 15.0ï¿½.  Both offices serve rural
populations within the Knoxville area. 



                          Table 5
          
          Tennessee: Administrative Costs Compared
            to Collections, January 1994 through
                         June 1995

                              Private office  Public office
                                    (Seventh         (Fifth
                                    Judicial       Judicial
                                   District)      District)
----------------------------  --------------  -------------
Administrative costs                $594,930       $354,025
Collections                       $3,966,143     $3,578,920
Cost to collect $1                     15.0ï¿½           9.9ï¿½
-----------------------------------------------------------
Finally, at the second privatized office we reviewed in Tennessee,
where we analyzed performance and cost-effectiveness before and after
privatization, the privatized office generally maintained comparable
performance and cost-effectiveness while serving a more challenging
caseload.\12 \13 This office, Tennessee's Twentieth Judicial
District, which includes Nashville, was privatized in state fiscal
year 1994.  Table 6 shows the percentage and table 7 shows the number
of cases receiving needed services for a 5-year period beginning with
state fiscal year 1991.  Five years of data are presented to show
data fluctuations from year to year.\14

Comparing the data over the 5-year period shows that while the
percentage of cases receiving needed services from the privatized
office remained about the same or declined in certain years, the
actual numbers of cases receiving most services rose dramatically. 
Even for the service that showed a decline--the number of cases
receiving a collection--the dollar amount of collections in the
Twentieth Judicial District increased over 40 percent in the first 2
years of privatization, as compared with 29 percent in the rest of
the state. 



                                Table 6
                
                Tennessee: Performance of the Twentieth
                 Judicial District, State Fiscal Years
                                1991-95

                         (Figures are percents)

                                        Public             Private
                                ----------------------  --------------
Cases with successful outcome     1991    1992    1993    1994    1995
------------------------------  ------  ------  ------  ------  ------
Location\a                        38.9    37.2    14.3    18.1    34.4
Paternity established              3.6     4.4     3.2     4.0     5.9
Support order established          7.4     4.5     3.2     7.7     4.5
Collection                        12.5    24.1    12.5     9.0    10.6
----------------------------------------------------------------------
\a Location rates may include multiple locations for the same
individual. 



                                Table 7
                
                 Tennessee: Cases Receiving Service in
                 the Twentieth Judicial District, State
                          Fiscal Years 1991-95

                                        Public             Private
                                ----------------------  --------------
Cases with successful outcome     1991    1992    1993    1994    1995
------------------------------  ------  ------  ------  ------  ------
Location\a                       3,657   3,601   1,414   3,608   7,998
Paternity established              424     472     354     927   1,346
Support order established          932     590     406   2,107     915
Collection                       7,404  16,200   5,572   3,554   4,349
----------------------------------------------------------------------
\a Location data may include multiple locations for the same
individual. 

State officials attribute the increase in the number of cases
receiving services to the contractor's automated case management
system and the employment of additional child support caseworkers and
attorneys.  This increase in the number of services provided,
however, is not reflected in the percentage of cases receiving
services because of large increases in identified case needs.  State
officials attributed this increase to the contractor's systematic
review of case files before entering the cases into its automated
system. 

With regard to cost-effectiveness, the privatized program in the
Twentieth Judicial District was about as cost-effective as the public
program had been.  As shown in table 8, in state fiscal year 1993,
the year before privatization, the cost for each $1 collected was
10.4ï¿½ for the district attorney's office, rising to 12.1ï¿½ in the
first year of privatization and declining to 10.7ï¿½ the following
year. 

While the cost-effectiveness of the program was about the same,
administrative costs did increase after privatization.  In state
fiscal year 1994, this office's administrative costs increased by
28.5 percent over the previous year, in contrast to an increase of
6.1 percent in the rest of the state.  The following year,
administrative costs grew by 32.7 percent in the Twentieth Judicial
District, virtually the same rate as in the rest of the state, 32.6
percent. 



                                 Table 8
                 
                 Tennessee: Administrative Costs Compared
                 to Collections in the Twentieth Judicial
                   District, State Fiscal Years 1991-95

                            Public                       Private
              ----------------------------------  ----------------------
                    1991        1992        1993        1994        1995
------------  ----------  ----------  ----------  ----------  ----------
Collections   $7,748,590  $12,078,72  $12,667,58  $14,409,94  $17,798,66
                                   8           9           8           2
Administrati    $986,341  $1,023,744  $1,315,643  $1,745,667  $1,900,231
 ve
 costs
Cost to            12.7ï¿½        8.5ï¿½       10.4ï¿½       12.1ï¿½       10.7ï¿½
 collect $1
------------------------------------------------------------------------
Note:  Collections figures before privatization are the subject of a
dispute between the state and the contractor.  The contractor claims
that the figures may be overstated, which may understate our
calculation of the cost to collect $1. 


--------------------
\12 During the period reviewed, the percentage of AFDC cases in this
office's caseload grew from about 29 percent to about 54 percent. 
AFDC cases are generally perceived by child support officials as
being more difficult to work.  While we were able to control for the
percentage of AFDC cases at our other comparison offices, AFDC
caseload growth could have affected this comparison. 

\13 To examine the possible effect of major changes in statewide
policy on the Twentieth Judicial District's performance outcomes, we
examined the performance of offices in the rest of the state before
and after the Twentieth Judicial District was privatized.  In
general, we found nothing to suggest that major policy changes were
affecting statewide performance outcomes. 

\14 State officials cited various reasons for this fluctuation,
including the implementation of a statewide data system beginning in
state fiscal year 1994, heightened efforts to achieve large numbers
of case closures in certain years, varying degrees of cooperation
with the judicial system in the area of paternity establishment, and
other factors related to the transition to privatization. 


      FACTORS BELIEVED TO AFFECT
      PERFORMANCE AND
      COST-EFFECTIVENESS
---------------------------------------------------------- Letter :4.4

State and contractor officials believed that several factors affect
an office's performance and cost-effectiveness.  Factors generally
believed to benefit contractors include the increased flexibility
contractors have in acquiring resources, managing staff, and having
greater access to technology.  For example, in Virginia, the
contractor was able to obtain computer and other equipment for the
new office in Hampton within 90 days of signing the contract.  In
contrast, state officials described a situation in an unrelated
office where delivery of needed equipment was expected to take 7
months under an expedited state acquisition process.  Likewise, in
Arizona, the contractor had the flexibility to open two privatized
offices within 2 months of signing the contract and to replace two
managers within the first year of operation.  Arizona state officials
told us that the removal of state managers could not be accomplished
as quickly.  Officials in all three states also cited contractors'
advantages in technology, such as automated systems for case
management, as a possible factor affecting performance differences. 
Furthermore, in Virginia, the contractor provided additional
technologies including extensive databases for locating noncustodial
parents, bar-coded files for data management, and the use of videos
during intake for consistent and complete orientation of both
custodial and noncustodial parents. 

In addition to the state officials' beliefs, contractor officials
also suggested factors that they believed favorably affected
performance outcomes, such as their emphasis on timely and efficient
processing of new cases, the co-location of child support workers in
AFDC offices for case intake, and expanded evening and weekend office
hours.  Contractor officials in Arizona, however, believe that their
performance was negatively affected by unexpectedly high staff
turnover during our review period. 

In addition to factors affecting performance outcomes, the major
factor affecting the cost-effectiveness of full-service privatization
is how much a state has to spend to acquire these services.  We
identified contract payment rates ranging from about 10 percent to 32
percent of collections.  These contract rates are affected by the
level of contractor competition and the volume, composition, and
collection potential of the caseload.  For example, because AFDC
cases are considered more labor-intensive to work, state officials
said that an office with a high percentage of AFDC cases may require
a higher payment rate than one with fewer AFDC cases.  Finally, all
payment rates are subject to the vagaries of the marketplace. 
Payment rates that are initially beneficial to a state could change
when the service is rebid or reawarded. 


   WELFARE REFORM PARTIALLY
   RESOLVES THE ISSUE OF ACCESS TO
   IRS DATA
------------------------------------------------------------ Letter :5

One major issue that could have impeded future full-service
privatization--contractors' access to IRS tax information--has been
partially resolved by the recent enactment of welfare reform
legislation.  The issue focuses on whether full-service child support
enforcement contractors have the same authority to access IRS data
for locating noncustodial parents and enforcing child support orders
as public offices have under the law.  The Personal Responsibility
and Work Opportunity Reconciliation Act of 1996 partially resolves
this issue by authorizing contractors access to certain tax
information. 

Currently, IRS, through OCSE, provides states and local child support
offices a number of valuable location and collection services.  For
example, OCSE's Federal Parent Locator Service (FPLS) can request
from IRS the social security number and address from a noncustodial
parent's most recent tax return.  In addition, IRS' Project 1099
program offers child support programs information on sources of
earned and unearned income, such as banks and mutual funds, that are
reported on IRS Form 1099 and can be used to locate noncustodial
parents and their assets.  Also, IRS' income tax refund offset
program has been the second largest source of collections for the
child support program.  In 1995, it offset over $828 million of
delinquent child support payments from about 1.2 million delinquent
noncustodial parents' tax refunds, in addition to providing offices
with filers' addresses.  OCSE and the states have granted
full-service contractors access to these IRS services and information
in the belief that these contractors act in the same capacity as
public child support offices; that is, as a designated local child
support agency of the state. 

Before enactment of the welfare reform legislation, however, IRS
officials took the position that section 6103(l) of the Internal
Revenue Code did not authorize child support contractors access to
this information.  In reports it issued in January 1995 and March
1996, IRS found Nebraska and Tennessee out of compliance with the
Internal Revenue Code during tax data safeguard reviews\15

because the states had granted contractors access to IRS data.  IRS
officials are concerned about their ability to safeguard tax
information and oppose further disclosure of IRS information for
nontax administrative purposes.  Federal, state, and contractor
officials told us that prohibiting access to IRS data would affect
the expansion and continuing operation of privatized full-service
child support offices. 

At the same time that OCSE and IRS have been working on this issue,
the Personal Responsibility and Work Opportunity Reconciliation Act
of 1996 has partially resolved this access issue.  Section 316(g)(4)
of the act amends section 6103(l) of the Internal Revenue Code to
permit state child support agencies to disclose to contractors the
addresses and social security numbers of noncustodial parents and the
amount of tax refunds withheld for past-due child support.  OCSE
officials said that this permits contractors access to FPLS and tax
refund offset data, but denies contractors access to unearned income
information currently received under the Project 1099 program. 

OCSE officials believe that the act partially addresses the issue of
access to data, but does not address contractors' access to the Form
1099 information that states and contractors believe serves as a
valuable enforcement tool.  One possible solution to this continuing
problem, according to IRS, is to have a state child support entity
verify this information with its original source.  Once the data are
verified, the information is not subject to nondisclosure
requirements.  Therefore, in order for the contractor to gain access
to Project 1099 program information, a state entity would have to
verify it with its source (for example, a financial institution),
thus entailing a separate process.  However, according to state
officials, requiring this additional level of verification by public
employees may negate some of the perceived benefits of full-service
privatization.  Alternatively, under the new law, contractors may
have another avenue to obtain unearned income data.  Section 372 of
the act requires states to match data quarterly with in-state
financial institutions.  However, HHS noted that it will be some time
before data from these matches are available in all states. 

While welfare reform legislation provides contractors limited access
to restricted tax data, it is too early to assess what effect these
changes will have on contractors' use of the full range of
enforcement tools currently available to public offices. 


--------------------
\15 A safeguard review is an on-site evaluation of the measures used
by agencies to protect federal tax returns and tax return information
received from IRS. 


   CONCLUSIONS
------------------------------------------------------------ Letter :6

As child support programs continue to face resource constraints,
full-service privatization appears to offer states the opportunity to
supplement their child support enforcement services.  The results of
our examination of full-service privatization show that the offices
we reviewed performed at least as well as public offices.  However,
these results are limited to the cases we reviewed and do not reflect
the performance of public or private offices overall within the
states selected.  Furthermore, because full-service privatization of
child support enforcement is relatively new, the extent to which it
offers comparable performance and cost-effectiveness remains an issue
for additional evaluation over the long term. 


   AGENCY COMMENTS
------------------------------------------------------------ Letter :7

In commenting on a draft of this report (see app.  V), HHS said that
the report can be a valuable resource to states as they consider
full-service privatization of child support enforcement functions. 
With respect to our observation that the Personal Responsibility and
Work Opportunity Reconciliation Act of 1996 has partially resolved
the issue of access to IRS data, HHS noted that it will be some time
before financial institution data are available in all states.  We
revised the final report to reflect the availability of such data. 

We also received technical comments from HHS, IRS, and the three
states and three contractors that we reviewed and incorporated them
in the final report as appropriate. 


---------------------------------------------------------- Letter :7.1

We are sending copies of this report to the Chairmen and Ranking
Minority Members of the Senate Committee on Finance and its
Subcommittee on Social Security and Family Policy and the House
Subcommittee on Human Resources, Committee on Ways and Means; the
Secretary of HHS; and HHS' Assistant Secretary for Children and
Families.  We also will make copies available to others on request. 
If you or your staff have any questions about this report, please
contact David P.  Bixler, Assistant Director, at (202) 512-7201 or
Kevin M.  Kumanga, Senior Evaluator, at (202) 512-4962.  Other major
contributors to this report are Christopher Morehouse, Steven
Machlin, Catherine Pardee, and Suzanne S.  Sterling. 

Sincerely yours,

Mark V.  Nadel
Associate Director, Income Security Issues


SCOPE AND METHODOLOGY
=========================================================== Appendix I

We compared three of the four fully privatized offices we selected
for our review with similar public offices, as illustrated in table
I.1.  The privatized offices selected for these comparisons were
Yavapai County, Arizona; the Seventh Judicial District of Tennessee
(Anderson County); and the Hampton District Office in Hampton,
Virginia.  We also compared another privatized office, the Twentieth
Judicial District of Tennessee (Davidson County), with the public
office that preceded it because it could not be matched with a
comparable public office in the state.  These private programs were
selected because they are among the most mature examples of
full-service privatization, having been privatized between 1992 and
1994 (see app.  IV for additional information); represent a diversity
of geographical and urban/suburban/rural program settings; and
include all three of the major full-service contractors.  At the time
of our review, Policy Studies, Inc., was the contractor for Yavapai
County, Arizona; Maximus ran the programs in Tennessee's Seventh and
Twentieth Judicial Districts; and Lockheed Martin IMS managed the
office in Hampton, Virginia. 



                               Table I.1
                
                  Offices Selected for Performance and
                      Cost-Effectiveness Analysis

                   Private                  Public
            ----------------------  ----------------------
            Local                   Local       Public      Methodolog
State       office      Contractor  office      entity      y used
----------  ----------  ----------  ----------  ----------  ----------
Virginia    Hampton     Lockheed    Portsmouth  State       Paired
            District    Martin IMS  District    Department  comparison
                                                of Social
                                                Services

Arizona     Yavapai     Policy      Mohave      State       Paired
            County      Studies,    County      Department  comparison
                        Inc.                    of
                                                Economic
                                                Security

Tennessee   Seventh     Maximus     Fifth       District    Paired
            Judicial                Judicial    Attorney    comparison
            District                District
            (Anderson               (Blount
            County)                 County)


Tennessee   Twentieth   Maximus     Twentieth   District    Before and
            Judicial                Judicial    Attorney    after
            District                District                comparison
            (Davidson               (Davidson
            County)                 County)
----------------------------------------------------------------------
To identify public offices for comparison and analysis, we asked
state officials to nominate offices for review that had performance
records similar to the privatized office before privatization.  We
then compared these offices with the privatized office, using the
selection criteria identified in table I.2, and selected the office
that most closely resembled the privatized program before
privatization. 



                               Table I.2
                
                 Selection Criteria for Public Offices

                                    Data used for comparison with
Selection criteria                  privatized program
----------------------------------  ----------------------------------
Performance                         Location of noncustodial parents
(for 3 years before privatization)  Paternity established
                                    Support order established
                                    Collection

Staffing                            Number of caseworkers

Caseload                            Size
                                    Composition: AFDC and non-AFDC
                                    Numbers of new cases, closed
                                    cases, and interstate cases

Cost                                Administrative costs

Demographics\a                      Population
                                    Minority population
                                    Median household income
                                    Percent of families below poverty
                                    Percent of households headed by a
                                    single parent
                                    Urbanization
----------------------------------------------------------------------
\a Determined from data for the most recent year for which data were
available. 

For each pair of offices, we defined our sets of cases to review as
all those opened in a given month (July 1994 for Arizona\16 and
Virginia, January 1994 and randomly selected additional cases from
February 1994 for Tennessee).  We chose to review new cases to
provide both public and private offices the greatest opportunities
for successful outcomes and to ensure that neither office had the
advantage of any experience with any of the cases.\17

We tracked performance outcomes for an 18-month period.  To further
assure comparability of cases between the two offices, we excluded
cases that (1) closed and did not reopen during the 18-month
period;\18 (2) were forwarded to other states to be worked because
the noncustodial parent resided out of state; (3) needed specialized
services, usually medical support only; or (4) were mistaken
referrals, duplicate cases, or otherwise inappropriate for analysis. 
Table I.3 shows the number of cases in our original cohorts for each
pair of offices, the number of cases excluded, and the number of
cases remaining for analysis of outcomes.  In each state, we gathered
case data for analysis by reviewing case files or automated case
management data.  Additionally, each state provided collections and
administrative cost data\19 for the cost-effectiveness analysis, and
Tennessee provided the data for the analysis of the Twentieth
Judicial District's performance before and after privatization.  We
did not verify the state-provided data. 



                               Table I.3
                
                Cases Identified, Excluded, and Examined
                          for Cohort Analysis

                           Virginia        Arizona        Tennessee
                        --------------  --------------  --------------
                                Privat          Privat          Privat
                        Public       e  Public       e  Public       e
----------------------  ------  ------  ------  ------  ------  ------
Cases in original          214     211     145      89      89     102
 cohort
Cases excluded             102     108      87      36      69      50
Cases analyzed in          112     103      58      53      20      52
 final cohort
----------------------------------------------------------------------
The results of our review represent the performance of these offices
on a small set of new cases for an abbreviated time period and do not
necessarily represent long-term public and private performance
overall within these states.  However, because both new and existing
cases must be equally served, the results can be viewed as a
reasonable indication of the overall performance of these offices for
that time period. 

We also did not compare the performance and cost outcomes of the
selected contracts across states because significant differences
exist in state demographics, caseloads, child support enforcement
tools available under state law, judicial and administrative systems,
and levels of state automation, among other factors. 


--------------------
\16 Cases were randomly selected for the public office in Arizona,
due to the volume of cases opened that month. 

\17 Child support officials told us that, generally, new cases result
in quicker, more positive outcomes than older cases because the
information is more current and the cases are easier to work. 
However, by federal regulations, both new and existing cases must be
equally served. 

\18 These cases closed for a variety of reasons.  For example, they
closed when the noncustodial parent died; the custodial parent did
not cooperate with the child support agency, moved out of the county
or state, or was no longer eligible to receive child support; the
noncustodial and custodial parents reunited; blood tests excluded the
individual identified as the putative father; or the noncustodial
parent assumed custody of the child upon the custodial parent's
incarceration. 

\19 Administrative costs did not include indirect costs that states
may have incurred in the process of preparing requests for proposals,
evaluating bids, and monitoring contracts. 


   OUTCOME ANALYSIS
--------------------------------------------------------- Appendix I:1

To compare the outcomes of the private and public offices for cases
that were in the system for 18 months, we analyzed differences
between the paired offices for Virginia and Arizona in outcomes in
each of the following four areas:  location, paternity establishment,
support order establishment, and collection.  All the cases were
included in the analysis for collections because all needed
collections.  However, for each of the other services, cases that did
not need the particular service at case opening were excluded from
the analysis.  Tables I.4 and I.5 contain the number of cases and
successful outcomes observed for the paired offices in Virginia and
Arizona, respectively. 



                               Table I.4
                
                 Numbers of Cases Needing Services and
                 With Successful Outcomes Among Cohort
                       Cases in Virginia Offices

                                        Private office  Public office
                                        --------------  --------------
                                                 Cases           Cases
                                         Cases    with   Cases    with
                                        needin  succes  needin  succes
                                             g    sful       g    sful
                                        servic  outcom  servic  outcom
Service needed                               e       e       e       e
--------------------------------------  ------  ------  ------  ------
Location                                    69      51      53      31
Paternity establishment                     50      20      73      14
Support order establishment                 84      30      95      16
Collection                                 103      42     112      25
----------------------------------------------------------------------


                               Table I.5
                
                 Numbers of Cases Needing Services and
                 With Successful Outcomes Among Cohort
                        Cases in Arizona Offices

                                        Private office  Public office
                                        --------------  --------------
                                                 Cases           Cases
                                         Cases    with   Cases    with
                                        needin  succes  needin  succes
                                             g    sful       g    sful
                                        servic  outcom  servic  outcom
Service needed                               e       e       e       e
--------------------------------------  ------  ------  ------  ------
Location                                    40      15      48      11
Paternity establishment                     23       2      31       3
Support order establishment                 35       5      52       3
Collection                                  53       8    57\a       3
----------------------------------------------------------------------
\a We could not determine whether or not one of the original 58 cases
received any collection. 

We did not conduct a similar analysis for the paired offices in
Tennessee because of the small number of cases at the public office
that met our criteria for inclusion.  Table I.6 shows summary data
for the cases from the private office in Tennessee. 



                               Table I.6
                
                  Summary Data for Cases From Private
                          Office in Tennessee

                                                                 Cases
                                                                  with
                                                                succes
                                                                  sful
                                                                outcom
Service needed                      Cases needing service            e
------------------------------  ------------------------------  ------
Location                                      24                    16
Paternity establisment                        18                     5
Support order establishment                   40                    12
Collection                                    52                    13
----------------------------------------------------------------------

   LOGISTIC REGRESSION MODELS
--------------------------------------------------------- Appendix I:2

In addition to calculating simple observed differences for the four
activities, we used a multivariate statistical technique--logistic
regression--to assess whether differences between offices were
statistically significant after adjusting for differences in caseload
characteristics.  In each model, the dependent or outcome variable
reflected whether an attempted action was successful or not (1 if
successful, 0 otherwise), while the primary independent variable of
interest was office type (1 if public, 0 if private).  We included
other independent variables in the models to adjust for caseload
characteristics that could influence an office's performance.  These
variables included AFDC status at case opening (1 if AFDC, 0
otherwise) and, if relevant to the outcome analyzed, whether the case
needed a particular action at case opening (1 if needed, 0
otherwise).\20 Table I.7 shows the distribution of cases on these
characteristics for each office in our analysis. 



                                    Table I.7
                     
                     Distribution of Cohort Cases on Selected
                                 Characteristics

                             Virginia                        Arizona
                  ------------------------------  ------------------------------
                      Public         Private          Public         Private
                  --------------  --------------  --------------  --------------
                          Percen          Percen          Percen          Percen
                  Number       t  Number       t  Number       t  Number       t
----------------  ------  ------  ------  ------  ------  ------  ------  ------
Cases                112     100     103     100      58     100      53     100
AFDC cases            81      72      54      52      53      91      36      68

Cases needing
--------------------------------------------------------------------------------
Location              53      47      69      67      48      83      40      75
Paternity             73      65      51      50      31      53      23      43
 establishment
Support order         95      85      84      82      52      90      35      66
 establishment
Location and          39      35      32      31      26      45      22      42
 paternity and
 support order
 establishment
Collection only        9       8       3       3       0       0       8      15
--------------------------------------------------------------------------------
Table I.8 contains the coefficients obtained from the logistic
regression models and their respective levels of statistical
significance (p-value).  The model coefficients reflect the magnitude
of the relationship between each independent and outcome variable. 
In general, the more the coefficient differs from 0, the more closely
the variables are related.  The p-value for each coefficient
represents the probability that our analysis would have produced a
coefficient of that magnitude or greater (that is, more different
from 0) if the two variables were not related.  We considered
coefficients with p-values lower than .05 to be statistically
significant.  We considered coefficients with p-values between .05
and .06 to be approaching statistical significance.  Our analysis
assumes that the cases we analyzed represent random samples of all
new cases handled by the offices.  While we analyzed all cases
meeting certain criteria that were opened in a particular month, for
the purpose of assessing statistical significance we considered these
cases as point-in-time samples that are representative of each
office's general performance on new cases. 



                               Table I.8
                
                  Logistic Regression Coefficients and
                  Corresponding P-Values Obtained From
                                 Models

                                           Virginia        Arizona
                                        --------------  --------------
                                                    P-
                                        Coeffi  value\  Coeffi      P-
Model                                    cient       a   cient   value
--------------------------------------  ------  ------  ------  ------
Location
----------------------------------------------------------------------
Office type                                  -     .21       -     .27
                                        0.5162          0.5360
AFDC status                                  -     .01       -     .04
                                        1.3183          1.3104

Paternity establishment
----------------------------------------------------------------------
Office type                                  -     .05       -   .76\b
                                        0.8411          0.3268
                                                            \b
AFDC status                                  -     .01      \b      \b
                                        1.1565
Need for location                            -     .67       -   .06\b
                                        0.1838          2.1196
                                                            \b

Support order establishment
----------------------------------------------------------------------
Office type                                  -     .06       -     .26
                                        0.7309          1.0242
AFDC status                                  -     .00       -     .55
                                        1.4297          0.7319
Need for location                       0.3849     .33       -     .08
                                                        1.5279
Need for paternity                           -     .34       -     .12
                                        0.3783          1.4314

Collection
----------------------------------------------------------------------
Office type                                  -     .04       -     .39
                                        0.8638          0.6956
AFDC status                                  -     .00       -     .16
                                        1.4324          1.2626
Need for location                            -     .87       -     .99
                                        0.0655          0.0048
Need for paternity                           -     .34       -     .06
                                        0.3965          2.1475
Need for support order                       -     .00  0.5387     .53
                                        3.3023
----------------------------------------------------------------------
Note:  To highlight the results for the main variable of interest
(office type) the corresponding coefficients and p-values are bolded. 

\a The p-value represents the probability of obtaining a coefficient
of this magnitude or further from 0 if the independent variable was
not related to the dependent variable.  P-values lower than .05 are
commonly considered to indicate a statistically significant effect. 

\b AFDC status was excluded from model because only two cases that
needed paternity establishment did not receive AFDC. 

We conducted our study between August 1995 and October 1996 in
accordance with generally accepted government auditing standards. 


--------------------
\20 As table I.7 shows, in our analysis of office pairs, the public
offices had a higher percentage of AFDC cases.  State officials said
that AFDC cases are generally harder to work for various reasons,
including limited case information and difficulties in securing the
cooperation of custodial parents, compared to non-AFDC cases.  The
actions we controlled for are location and paternity and support
order establishment. 


BASIC CHILD SUPPORT ENFORCEMENT
SERVICES
========================================================== Appendix II

Location includes efforts at local, state, and federal levels to
identify a noncustodial parent's address, social security number,
place of employment, and other characteristics.  This might include
efforts to directly contact individuals; contacts with public and
private institutions, such as credit bureaus, state and federal
income tax agencies; and the use of computer tape matches. 

Paternity establishment is the identification of the legal father of
a child, usually through the courts or expedited through hearings in
a quasi-judicial or administrative body.  Paternities are established
in either of two ways:  (1) through voluntary acknowledgment by the
father or (2) if contested, through a determination made on the basis
of scientific and testimonial evidence. 

Support order establishment involves the development of a support
order that legally obliges the noncustodial parent to pay child
support and provide medical insurance coverage when available at
reasonable cost.  The child support enforcement agency must assist
custodial parents in initiating an action in court or through an
administrative or expedited legal process that will produce such an
order.  The child support enforcement agency helps in the
determination of a child's financial needs and the extent to which
the noncustodial parent can provide financial support and medical
insurance coverage.  Support orders are subject to periodic review
and adjustment at least every 3 years in AFDC cases and upon parental
request in non-AFDC cases. 

Collections and enforcement involves enforcing, monitoring, and
processing payments.  To enforce payment on delinquent cases or to
ensure regularity and completeness of current accounts, child support
enforcement agencies have a wide array of techniques at their
disposal.  These techniques include bonds and security deposits,
federal and state tax intercepts, garnishments, liens, and wage
withholding, among others.  Noncustodial parents' payments must also
be monitored, recorded, and distributed. 


CONTRACT INFORMATION FOR
FULL-SERVICE CHILD SUPPORT
PRIVATIZATION INITIATIVES IN 15
STATES AS OF SEPTEMBER 1996
========================================================= Appendix III

                Fully                           Start date
                privatized                      (contract       Payment
State           local offices   Contractor      length\a)       terms\b,c
--------------  --------------  --------------  --------------  ----------------
Arizona
--------------------------------------------------------------------------------
                Santa Cruz and  Policy          3/94 (4 years)  32-24 percent\d
                Yavapai         Studies, Inc.
                Counties        


Arkansas
--------------------------------------------------------------------------------
                Jefferson       Hunt Law Firm   9/95 (1.8       Fixed fee of
                County                          years)          $810,000,
                                                                reimbursable
                                                                expenses up to
                                                                $490,000 plus
                                                                incentives

                Greene County   Greene County   9/95 (1.8       Fixed fee of
                                Child Support   years)          $175,000,
                                Enforcement,                    reimbursable
                                Inc.                            expenses up to
                                                                $290,000 plus
                                                                incentives

                Garland County  Owen Support    7/95 (2 years)  Fixed fee of
                                Services, Inc.                  $390,000,
                                                                reimbursable
                                                                expenses up to
                                                                $460,000 plus
                                                                incentives

                Craighead       Brent Davis,    7/95 (2 years)  Negotiated
                County          Esq.                            contract\e

                3rd Judicial    Randolph        7/95 (2 years)  Negotiated
                District        County Judge/                   contract\e
                (Clay,          Multi
                Jackson,        Services, Inc.
                Lawrence,
                Randolph, and
                Sharp
                Counties)

                14th Judicial   Multi           7/95 (2 years)  Fixed fee of
                District        Services, Inc.                  $300,000,
                (Baxter,                                        reimbursable
                Boone, Marion,                                  expenses up to
                and Newton                                      $330,000
                Counties)


Colorado
--------------------------------------------------------------------------------
                El Paso County  Maximus         1/96 (5 years)  19-10 percent\d


Georgia\f
--------------------------------------------------------------------------------
                Cobb, De Kalb,  Policy          7/94 (4 years)  11.5-10
                and Fulton      Studies, Inc.                   percent\d
                Counties


Iowa\g
--------------------------------------------------------------------------------
                Fremont,        Policy          1/96 (5 years)  Flat fee\h
                Harrison,       Studies, Inc.
                Mills,
                Montgomery,
                Page,
                Pottawattamie,
                and Taylor
                Counties;
                Muscatine and
                Scott
                Counties; and
                Boone, Dallas,
                Jasper,
                Madison,
                Marion, Polk,
                Storey, and
                Warren
                Counties


Maryland
--------------------------------------------------------------------------------
                Baltimore City  Lockheed        11/96 (4        22.95-20.55
                                Martin IMS      years)          percent\d

                Queen Anne's    Lockheed        11/96 (4        9.67 percent
                County          Martin IMS      years)


Mississippi
--------------------------------------------------------------------------------
                Hinds and       Maximus         7/94 (5 years)  40.1-27.6
                Warren                                          percent\d,i
                Counties


Nebraska
--------------------------------------------------------------------------------
                Douglas County  Policy          2/93 (5 years)  15-13 percent\d
                                Studies, Inc.


Ohio\j
--------------------------------------------------------------------------------
                Hamilton        Maximus         6/96 (3 years)  9.9-6.6
                County                                          percent\d,k


Oklahoma
--------------------------------------------------------------------------------
                Pittsburg and   Kibois          1993 (4 years)  Cost-
                Haskell         Community                       reimbursement
                Counties        Action
                                Foundation
                                (nonprofit)

                Comanche and    Great Plains    1993 (4 years)  Cost-
                Cotton          Improvement                     reimbursement
                Counties        Foundation
                                (nonprofit)


South Carolina
--------------------------------------------------------------------------------
                Georgetown,     To be awarded   1/97 (5 years)
                Horry, Marion,
                and
                Williamsburg
                Counties


Tennessee
--------------------------------------------------------------------------------
                7th Judicial    Maximus         7/92 (5 years)  16 percent
                District
                (Anderson
                County)

                20th Judicial   Maximus         7/93 (5 years)  12-10.5
                District                                        percent\d
                (Davidson
                County)

                10th Judicial   Policy          7/96 (5 years)  15.75-13.5
                District        Studies, Inc.                   percent\d
                (Bradley,
                McMinn,
                Monroe, and
                Polk Counties)

                29th Judicial   Policy          2/92 (5 years)  19-15 percent\d
                District (Dyer  Studies, Inc.
                and Lake
                Counties)

                27th Judicial   Policy          1/95 (5 years)  17-14 percent\d
                District        Studies, Inc.
                (Obion and
                Weakley
                Counties)

                21st Judicial   To be awarded   1/97 (5 years)
                District
                (Hickman,
                Louis, Perry,
                and Williamson
                Counties)


Virginia\l
--------------------------------------------------------------------------------
                Hampton\m and   Lockheed        2/94 (5 years)  11.45-9.95
                Chesapeake      Martin IMS                      percent\d
                District
                Offices

                Alexandria and  To be awarded   1997 (5 years)
                Arlington\n
                District
                Offices

West Virginia

                Kanawha County  Policy          9/96 (3 years)  18 percent\o
                                Studies, Inc.


Wyoming
--------------------------------------------------------------------------------
                Districts 1,    Policy          6/95 (4 years)  17.5-16
                2, and 3        Studies, Inc.                   percent\d
                (Albany,
                Carbon,
                Laramie,
                Lincoln,
                Sweetwater,
                and Uinta
                Counties)

                Districts 8     Gray &          5/95 (4 years)  $724,000 + 8
                and 9           Associates                      percent of
                (Converse,                                      collections over
                Fremont,                                        $2.5 million
                Goshen,
                Niobrara,
                Platte,
                Sublette, and
                Teton
                Counties)
--------------------------------------------------------------------------------
\a Length of contract can include possible annual renewals. 

\b Unless otherwise noted, payment terms are expressed as a
percentage of contractor-generated collections. 

\c Payment terms vary, depending on factors such as caseload volume
and composition and use of multiple or single contractors. 

\d First figure is the payment rate for the contract's first year;
second figure is the rate reached by the end of the contract period. 

\e Payment terms of negotiated contracts are not disclosed because
they are considered private information. 

\f The contract provides for service to non-AFDC clients only; the
state continues to provide services to AFDC clients.  Additionally,
in Cobb and De Kalb Counties, the contractor serves in-state cases
only. 

\g This contract provides for the privatization of location,
paternity establishment, administrative support order establishment,
and certain customer relations services in the areas shown.  The
state continues to provide other services. 

\h This performance-based contract is subject to monetary penalties
and liquidated damages. 

\i Terms are for a statewide operation of a full-service program that
has not yet been implemented and is pending legislative approval.  In
addition, the contractor received $2.98 million for start-up costs
between April and June 1994 and $14,180,262 as a flat fee for the
first 15 months of the program.  As amended for the period from July
1, 1996, through June 30, 1997, the contract provides for
compensation of $4.39 million, including incentives, and additional
bonuses paid on the basis of collections. 

\j Under this performance-based contract, the caseload--both AFDC and
non-AFDC--is shared by the contractor and the county child support
agency.  Currently, the contractor's share of the county caseload is
estimated by the state child support director to be about 40 percent. 

\k This payment rate applies to non-AFDC cases only.  Other rates
apply to other types of cases, for example, from 26.8 to 17.8 percent
of collections for AFDC cases, and 10.5 to 6.6 percent of collections
for outgoing interstate cases.  Compensation is not to exceed
$6,308,554. 

\l The state legislature has authorized the creation of two new
district offices in addition to those listed.  The new district
offices that would serve areas yet to be determined are to be
established in state fiscal year 1997 and state fiscal year 1998. 

\m The Hampton District Office serves the city of Hampton and
Gloucester, Mathews, Middlesex, Poquoson, and York Counties. 

\n The Arlington District Office will serve the cities of Arlington
and Falls Church. 

\o Payment not to exceed $2.2 million in the contract's first year. 


SELECTED DEMOGRAPHIC
CHARACTERISTICS OF THE LOCALITIES
REVIEWED
========================================================== Appendix IV

Unless otherwise noted, the source for all demographic data in this
appendix is U.S.  Bureau of the Census, County and City Data Book
1994 (Washington, D.C.:  U.S.  Bureau of the Census, 1994). 


   VIRGINIA
-------------------------------------------------------- Appendix IV:1

Located in the Norfolk-Virginia Beach-Newport News metropolitan area,
the Hampton District Office serves both suburban and rural areas.  In
1992, the population in Hampton was about 137,000.  Located in the
same metropolitan area, the Portsmouth District Office serves
Portsmouth City, with a 1992 population of about 105,000.  In 1989,
the median income in Hampton was $30,144 and in Portsmouth, $24,601. 
In 1989, 8.8 percent of the families in Hampton had incomes below the
poverty level, compared with 14.9 percent in Portsmouth. 

Like Portsmouth, in 1992, Hampton was among the top 25
counties/jurisdictions in the country in federal civilian employment. 
Hampton's June 1994 caseload of 15,000 was about 42 percent AFDC and
58 percent non-AFDC.  At the same time, Portsmouth served a caseload
of about 14,000, about 51 percent of which were AFDC cases and 49
percent non-AFDC. 

Statewide in 1992, there were 36.9 births to unmarried teenagers
between 15 and 19 years old per 1,000 females.\21 According to 1988
Bureau of the Census data, 11.7 percent of all births in Hampton were
to mothers under 20 years old, compared with 18.6 percent in
Portsmouth and 11.2 percent statewide. 


--------------------
\21 Annie E.  Casey Foundation, Kids Count Data Book 1995 (Baltimore,
Md.:  Annie E.  Casey Foundation, 1995). 


   ARIZONA
-------------------------------------------------------- Appendix IV:2

Yavapai County is north of Phoenix; its 1992 population was about
116,000, compared with about 106,000 in Mohave County.  While Yavapai
County is outside the state's metropolitan areas, it has one major
town within its borders--Prescott.  Mohave County is in a
metropolitan area (Las Vegas, Nevada); its major urban areas include
Kingman, Bullhead City, and Lake Havasu City.  In 1989, the median
income in Yavapai was $22,060, and in Mohave, $24,002.  In 1989, 9.8
percent of Yavapai's families had incomes below the poverty level,
compared with 8.7 percent in Mohave.  In state fiscal year 1994, the
Yavapai office had about 10,000 child support cases; about 24 percent
were AFDC and 76 percent non-AFDC.  At that time, Mohave's caseload
was about 15,000, 36 percent of which were AFDC cases. 

Mohave County stands out among Arizona counties in two respects. 
Mohave's population is highly mobile:  from 1985 to 1990, the county
led the state in the percentage of movers, 61.7 percent.  Mohave
County also had the highest population growth rate of any county in
the state between 1980 and 1992:  89.8 percent. 

Statewide, the unmarried teenage birth rate in 1992 was 62.7 per
1,000.\22 Census data show that 15.3 percent of all births for
Yavapai County in 1988 were to teenage mothers, 15.1 percent for
Mohave, and 13.8 percent statewide. 


--------------------
\22 Kids Count Data Book 1995. 


   TENNESSEE
-------------------------------------------------------- Appendix IV:3

Both the Seventh and Fifth Judicial Districts--Anderson County and
Blount County, respectively--are in the Knoxville metropolitan area,
with Anderson County lying north of the city and Blount County to its
south.  Anderson County's 1992 population was about 70,500, compared
with about 90,400 in Blount County.  Major towns include Clinton and
Oak Ridge in Anderson County and Maryville in Blount County.  In
1989, Anderson County's median income was $26,496, compared with
$25,575 in Blount County.  The percentage of families with incomes
below poverty in 1989 was 11.5 percent in Anderson County and 10
percent in Blount County.  From June 1994 through December 1995, the
Seventh Judicial District served a caseload of about 4,600, almost
evenly divided between AFDC and non-AFDC cases.  During the same
period, the Fifth Judicial District's caseload was about 6,300, of
which 73 percent were AFDC cases and 27 percent non-AFDC cases. 

The Twentieth Judicial District serves Davidson County, which
includes Nashville, which had a 1992 population of about 495,000.  In
1989, the county's median income was $27,821 and 13.4 percent of the
county's families had incomes below the poverty level.  The child
support caseload in state fiscal year 1994 was about 39,400, and 57
percent were AFDC cases. 

Statewide, in 1992, there were 46.3 births to unmarried teenagers 15
to 19 years old per 1,000 females, close to the national average of
42.5.\23 According to 1988 Census data, teenage mothers accounted for
16.0 percent of all births in Anderson County, 14.2 percent in Blount
County, 15.2 percent in Davidson County, and 17.2 percent statewide. 



(See figure in printed edition.)Appendix V

--------------------
\23 Kids Count Data Book 1995. 


COMMENTS FROM THE DEPARTMENT OF
HEALTH AND HUMAN SERVICES
========================================================== Appendix IV



(See figure in printed edition.)


RELATED GAO PRODUCTS
=========================================================== Appendix 0

Child Support Enforcement:  Reorienting Management Toward Achieving
Better Program Results (GAO/HEHS/GGD-97-14, Oct.  25, 1996). 

Child Support Enforcement:  States' Experience with Private Agencies'
Collection of Support Payments (GAO/HEHS-97-11, Oct.  23, 1996). 

Child Support Enforcement:  States and Localities Move to Privatized
Services (GAO/HEHS-96-43FS, Nov.  20, 1995). 

Child Support Enforcement:  Opportunity to Reduce Federal and State
Costs (GAO/T-HEHS-95-181, June 13, 1995). 

Child Support Enforcement:  Families Could Benefit From Stronger
Enforcement Program (GAO/HEHS-95-24, Dec.  27, 1994). 

Child Support Enforcement:  Federal Efforts Have Not Kept Pace With
Expanding Program (GAO/T-HEHS-94-209, July 20, 1994). 

Child Support Enforcement:  Credit Bureau Reporting Shows Promise
(GAO/HEHS-94-175, June 3, 1994). 

Child Support Assurance:  Effect of Applying State Guidelines to
Determine Fathers' Payments (GAO/HRD-93-26, Jan.  23, 1993). 

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

Medicaid:  Ensuring That Noncustodial Parents Provide Health
Insurance Can Save Costs (GAO/HRD-92-80, June 17, 1992). 

Interstate Child Support:  Wage Withholding Not Fulfilling
Expectations (GAO/HRD-92-65BR, Feb.  25, 1992). 

Interstate Child Support:  Mothers Report Less Support From
Out-of-State Fathers (GAO/HRD-92-39FS, Jan.  9, 1992). 


*** End of document. ***