Welfare Reform: Challenges in Maintaining a Federal-State Fiscal 
Partnership (10-AUG-01, GAO-01-828).				 
								 
The upcoming reauthorization of the Temporary Assistance for	 
Needy Families (TANF) block grant represents an opportunity to	 
re-examine the fiscal balance between the federal government and 
the states in providing services to needy families. In the five  
years since the enactment of federal welfare reform, there has	 
been much discussion of the fiscal implications of these sweeping
changes in national welfare policy. A particularly contentious	 
issue has been the extent to which states have replaced, rather  
than supplemented, their own spending with federal TANF dollars  
thereby freeing up state funds for other budget priorities. This 
report reviews the (1) degree to which states have used the	 
flexibility afforded in the federal TANF grant to supplant,	 
rather than supplement, state spending for low-income families,  
(2) changes that have occurred in states' use of different	 
funding sources (including their TANF funds) on programs that	 
help the poor, (3) effects of state funding choices have had on  
the amounts of TANF funds they have left unspent at the U.S.	 
Treasury, and (4) measures states are taking to save a portion of
the TANF grant or set aside state funds for a rainy day.	 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-01-828 					        
    ACCNO:   A01582						        
  TITLE:     Welfare Reform: Challenges in Maintaining a Federal-State
             Fiscal Partnership                                               
     DATE:   08/10/2001 
  SUBJECT:   Federal/state relations				 
	     Reporting requirements				 
	     State-administered programs			 
	     Block grants					 
	     Public assistance programs 			 
	     Funds management					 
	     Unexpended budget balances 			 
	     Aid to Families with Dependent Children		 
	     Program						 
								 
	     Contingency Fund for State Welfare 		 
	     Programs						 
								 
	     Federal Loan Fund for State Welfare		 
	     Programs						 
								 
	     HHS Temporary Assistance for Needy 		 
	     Families Block Grant				 
								 
	     Temporary Assistance for Needy Families		 
	     Program						 
								 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Testimony.                                               **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-01-828
     
A

Report to Congressional Requesters

August 2001 WELFARE REFORM Challenges in Maintaining a Federal- State Fiscal
Partnership

GAO- 01- 828

Letter 5 Background 6 Scope and Methodology 9 Results in Brief 12 State
Budget Actions Shift Federal and State Support for

Specific Activities 16 Net Fiscal Effects of Funding Shifts Varied Among
States 24 States? Contingency Plans Are Undeveloped 34 Design of Federal
Contingency Mechanisms Is Complex and

Restrictive 40 Concluding Observations 43 Recommendations for Executive
Action 51 Agency Comments 52

Appendixes Appendix I: Scope and Methodology 54 Appendix II: California 61
Appendix III: Colorado 70 Appendix IV: Connecticut 78 Appendix V: Louisiana
85 Appendix VI: Maryland 92 Appendix VII: Michigan 99 Appendix VIII: New
York 108 Appendix IX: Oregon 117 Appendix X: Texas 126 Appendix XI:
Wisconsin 135 Appendix XII: Survey Instrument 144 Appendix XIII: Comments
From the Department of Health

and Human Services 154 Tables Table 1: Changes in Basic Welfare Services 28

Table 2: Category 1- State Expenditures on Basic Welfare Services (State
Fiscal Years 1995 and 2000) 29 Table 3: Category 2- Expenditures on Programs
Supporting the Working Poor (State Fiscal Years 1995 and 2000) 30

Table 4: Category 3- Expenditures on Programs Supporting the Working Poor
and Related Health Care Costs (State Fiscal Years 1995 and 2000) 32 Table 5:
Category 4- Expenditures on TANF- Related Social Services for Low- Income
Families and the Working

Poor (State Fiscal Years 1995 and 2000) 33 Table 6: California?s TANF-
Related Expenditures (Categories 1 & 2) 65 Table 7: California?s TANF-
Related Expenditures (Categories 3 & 4) 66 Table 8: Unspent TANF Funds in
California as of September 30,

2000 67 Table 9: Colorado?s TANF- Related Expenditures (Categories 1 & 2) 74
Table 10: Colorado?s TANF- Related Expenditures (Categories 3 & 4) 75 Table
11: Unspent TANF Funds in Colorado as of September 30,

2000 76 Table 12: Connecticut?s TANF- Related Expenditures (Categories 1 &
2) 82 Table 13: Connecticut?s TANF- Related Expenditures (Categories 3 & 4)
83 Table 14: Unspent TANF Funds in Connecticut as of September 30,

2000 84 Table 15: Louisiana?s TANF- Related Expenditures (Categories 1 & 2)
88 Table 16: Louisiana?s TANF- Related Expenditures (Categories 3 & 4) 89
Table 17: Unspent TANF Funds in Louisiana as of September 30,

2000 90 Table 18: Maryland?s TANF- Related Expenditures (Categories 1 & 2)
95 Table 19: Maryland?s TANF- Related Expenditures (Categories 3 & 4) 96
Table 20: Unspent TANF Funds in Maryland as of September 30,

2000 97 Table 21: Michigan?s TANF- Related Expenditures (Categories 1 & 2)
103 Table 22: Michigan?s TANF- Related Expenditures (Categories 3 & 4) 104
Table 23: Unspent TANF Funds in Michigan as of September 30,

2000 106 Table 24: New York?s TANF- Related Expenditures (Categories 1 & 2)
113 Table 25: New York?s TANF- Related Expenditures (Categories 3 & 4) 114
Table 26: Unspent TANF Funds in New York as of September 30,

2000 116 Table 27: Oregon?s TANF- Related Expenditures (Categories 1 & 2)
122 Table 28: Oregon?s TANF- Related Expenditures (Categories 3 & 4) 123
Table 29: Unspent TANF Funds in Oregon as of September 30,

2000 125 Table 30: Texas? TANF- Related Expenditures (Categories 1 & 2) 131
Table 31: Texas? TANF- Related Expenditures (Categories 3 & 4) 132 Table 32:
Unspent TANF Funds in Texas as of September 30,

2000 134

Table 33: Wisconsin?s TANF- Related Expenditures (Categories 1 & 2) 139
Table 34: Wisconsin?s TANF- Related Expenditures (Categories 3 & 4) 141
Table 35: Unspent TANF Funds in Wisconsin as of September 30,

2000 142 Figures Figure 1: Four Categories of Welfare- Related Spending 11

Figure 2: TANF- Related Social Service Expenditures (Category 4) in 10
States for State Fiscal Years 1995 and 2000 13 Figure 3: The Contingency
Fund?s Annual Reconciliation Process 43 Figure 4: California?s TANF- Related
Expenditures for State Fiscal

Years 1995 and 2000 64 Figure 5: Colorado?s TANF- Related Expenditures for
State Fiscal Years 1995 and 2000 72

Figure 6: Connecticut?s TANF- Related Expenditures for State Fiscal Years
1995 and 2000 81 Figure 7: Louisiana?s TANF- Related Expenditures for State
Fiscal

Years 1995 and 2000 87 Figure 8: Maryland?s TANF- Related Expenditures for
State Fiscal Years 1995 and 2000 94

Figure 9: Michigan?s TANF- Related Expenditures for State Fiscal Years 1995
and 2000 102 Figure 10: New York?s TANF- Related Expenditures for State
Fiscal Years 1995 and 2000 112 Figure 11: Oregon?s TANF- Related
Expenditures for State Fiscal Years 1995 and 2000 121 Figure 12: Texas?
TANF- Related Expenditures for State Fiscal Years 1995 and 2000 130 Figure
13: Wisconsin?s TANF- Related Expenditures for State Fiscal

Years 1995 and 2000 138

Lett er

August 10, 2001 The Honorable Nancy L. Johnson Chairwoman, Subcommittee on
Health Committee on Ways and Means House of Representatives

The Honorable Wally Herger Chairman, Subcommittee on Human Resources
Committee on Ways and Means House of Representatives

The upcoming reauthorization of the Temporary Assistance for Needy Families
(TANF) block grant presents an opportunity to re- examine the fiscal balance
between the federal government and the states in providing services to needy
families. TANF was created by the Personal

Responsibility and Work Opportunity Reconciliation Act of 1996 1 (PRWORA, or
the act) which ended families? entitlement to cash assistance by replacing
the Aid to Families with Dependent Children (AFDC) program with the
Temporary Assistance for Needy Families (TANF) block grant, a $16.5 billion
per- year fixed federal funding stream to the states. PRWORA coupled the
block grant with a strong maintenance- of- effort (MOE) requirement 2 in
order to ensure that states remain a strong fiscal partner. The MOE requires
states to maintain a significant portion of their own

historic financial commitment to their welfare programs as a condition of
receiving their full TANF allotments. Since both the amount of federal TANF
funds and the required MOE remain fixed regardless of the number of people
states chose to serve with these funds, and because the number of families
that receive basic cash assistance has declined dramatically since the mid-
1990s (by more than 50 percent nationally), states have additional budgetary
resources available for use in a variety of ways to help 1 Most
appropriations for TANF expire at the end of fiscal year 2002 (Public Law
104- 193, Title I).

2 States? MOE requirements are based on their own spending in federal fiscal
year 1994 on AFDC, Job Opportunities and Basic Skills (JOBS), Emergency
Assistance (EA), related administrative costs and AFDC- related child care
programs: AFDC/ JOBS child care program, Transitional Child Care, and At-
Risk Child Care programs. A state that does not meet the act?s work
participation rates must maintain at least 80 percent MOE. A state that
meets its work participation rate must maintain at least 75 percent of its
MOE. For more

information see Welfare Reform: Early Fiscal Effects of the TANF Block Grant
(GAO/ AIMD98- 137, August 18, 1998).

people avoid dependency on public assistance. At the same time, the fixed
nature of the federal contribution means that states now bear most of the
program?s fiscal risks in the event of an economic downturn or increase in
caseload.

In the 5 years since enactment of federal welfare reform, there has been
much discussion of the fiscal implications of these sweeping changes in
national welfare policy. A particularly contentious issue has been the
extent to which states have replaced, rather than supplemented, their own
spending with federal TANF dollars thereby freeing up state funds for other
budget priorities. This practice is often referred to as ?supplantation? or

?substitution.? 3 Other important intergovernmental fiscal issues
surrounding PRWORA include whether states are setting aside funds for costs
associated with increased caseloads they may experience during a recession
and why some states are leaving large reserves of unspent TANF funds at the
U. S. Treasury. 4

Concerns about how states were spending their TANF funds prompted you to ask
us to examine:

 the degree to which states have used the flexibility afforded in the
federal TANF grant to supplant, rather than supplement, state spending for
low- income families;  the changes that have occurred in states? use of
different funding

sources (including their TANF funds) on programs that help the poor;

 the effect state funding choices have had on the amounts of TANF funds
they have left unspent at the U. S. Treasury; and

 the measures states are taking to save a portion of the TANF grant or set
aside state funds for a rainy day.

Background The change in the federal role in financing state welfare
programs has provided both greater flexibility and greater uncertainty for
the states. They have greater flexibility in how to administer and implement
their welfare programs; however, they also face increased fiscal risk.
Before 3 For more information, see Federal Grants: Design Improvements Could
Help Federal Resources Go Further (GAO/ AIMD- 97- 7, December 18, 1996).

4 For more information, see Welfare Reform: Challenges in Saving for a
?Rainy Day?( GAO01- 674T, April 26, 2001).

1996, under the AFDC program, any increased costs were shared by the federal
government and the states. Under TANF, however, if costs rise, states face
most of the burden of financing the unexpected costs. States must handle
these costs in the context of any budgetary limitations- including those
imposed by legislation, balanced budget requirements, and scrutiny of the
municipal bond market- on their ability to increase spending, especially in
times of fiscal stress. This uncertainty highlights the

importance of contingency planning, or ?saving for a rainy day.? At the same
time, when caseloads drop, as they have since the mid- 1990s, states can
reap large fiscal benefits. A particularly contentious issue has been the
extent to which states used this flexibility to replace, rather than
supplement, state spending with federal TANF dollars thereby freeing up
state funds for other budget priorities- a practice allowable under TANF. At
the same time Congress consolidated the funding streams from the previous
welfare programs into the TANF block grant, it broadened the programs goals
and gave states flexibility to use federal funds on a wider

array of programs. Specifically, states can use TANF funds on programs and
activities to

 provide assistance to needy families so that children may be cared for in
their homes or in the homes of relatives;

 end the dependence of needy parents on government benefits by promoting
job preparation, work, and marriage;

 prevent and reduce the incidence of out- of- wedlock pregnancies; and

 encourage the formation and maintenance of two- parent families. Many
states are in a position to supplant state spending because given large,
recent cash assistance caseload declines, the fixed TANF funding formula
provides them with more budgetary resources for their welfare programs than
they would have had under prior law. However, TANF?s

MOE limits states? ability to supplant funds because states face strict
penalties unless they largely maintain their prior fiscal commitment to
lowincome families. Therefore, states? ability to divert state funds to
other priorities is somewhat constrained.

In theory, federal grants serve purposes beyond simply returning and
reallocating resources to taxpayers in the form of state services. As we
have previously reported, they have also served as vehicles through which
the federal government attempted to achieve a variety of national goals by
providing funding to other levels of government to carry out specific
national priorities. Grants can also serve as a tool to encourage states to
spend federal funds for nationally important activities for which they
otherwise would have spent less. The amount of additional spending is
affected by the degree to which federal grants actually supplement, rather

than supplant, state spending on these activities. 5 5 Federal Grants:
Design Improvements Could Help Federal Resources Go Further (GAO/ AIMD- 97-
7, December 18, 1996).

Paradoxically, at the same time questions are being raised as to whether
states have replaced their own spending with federal funds, the states have
been under considerable scrutiny for not spending their TANF funds more
quickly. There has also been much discussion about whether states are

?saving for a rainy day.? 6 If states choose to engage in this sort of
contingency budgeting, they have several options. They can (1) use general-
purpose state ?rainy day funds,? (2) establish a state reserve specifically
intended for welfare programs, (3) plan on using PRWORA?s two safety net
mechanisms- the $2 billion Contingency Fund for State Welfare Programs
(Contingency Fund) and the $1.7 billion Federal Loan Fund for State Welfare
Programs (Loan Fund), 7 or (4) save an unlimited amount of their TANF
allocation for use in later years. 8 If states choose to save all or part of
their unspent TANF balances, laws governing the

exchange of funds between the federal government and the states require that
these resources be left at the U. S. Treasury until they are needed. 9 Scope
and To address the four objectives of this report, we studied TANF- related

Methodology budget decisions and collected data on social service
expenditures in 10 states -California, Connecticut, Colorado, Louisiana,
Maryland, Michigan,

New York, Oregon, Texas, and Wisconsin. These 10 states were the same states
we studied in 1998, thereby providing a continuing perspective on fiscal
issues among a common group of states as the program unfolds. These 10
states were selected because they represent a diverse array of

socioeconomic characteristics, geographic locations, experiences with state
welfare initiatives, and state fiscal and budget issues. 6 See Welfare
Reform: Challenges in ?Saving for a Rainy Day? (GAO- 01- 674T, April 26,
2001). 7 The Adoption and Safe Families Act of 1997 reduced the Contingency
Fund by $40 million over the remaining life of the fund. (P. L. 105- 89,
sect.404, 111 Stat. 2134.) Loans from the Loan Fund are to have a maturity of no
more than 3 years at an interest rate comparable to the average market yield
on outstanding marketable federal obligations with comparable periods to
maturity.

8 HHS regulations restrict the states? use of these prior year funds to
benefits and basic services to help families meet on- going needs. 9 Federal
regulations prohibit grantees from drawing down funds in advance of
expenditure in order to assure compliance with the Cash Management
Improvement Act of 1990 (CMIA). CMIA settled a long- standing dispute
between the federal government and the states over disbursement of funds for
federal programs administered by the state; it helps to ensure that neither
party incurs unnecessary interest costs in the course of federal grant
disbursements.

 In order to gauge the extent to which a state used TANF funds to free up
state funds, we analyzed state budgets and supporting budget documents and
interviewed key program budget officials, state budget

officers, state auditors and legislative staff.

 To measure post- PRWORA changes in the mix of federal and state funding
across TANF- related social services, we asked the 10 states to provide
expenditure data on programs that provide social services to

the needy. To do this, we adapted a fiscal survey developed by the Nelson A.
Rockefeller Institute of Government. 10 Specifically, we asked for state and
federal expenditures on a variety of programs that targeted assistance to
low- income families and individuals in state fiscal year 1994- 1995 and
estimated expenditures in state fiscal year 1999- 2000, and we verified the
data they reported to us against their own accounting records. To improve
the alignment of our survey data with TANF?s goal of ending the dependence
of needy parents on government benefits, we excluded expenditures on the
elderly and those services that were not transitional in nature. In our
analysis, we aggregated state expenditure

data by program category, allowing us to group programs into four
categories- Basic Welfare Services (Category 1), Support for the Working
Poor (Category 2), Health Care (Category 3), and Other TANFrelated Social
Services (Category 4). The analysis is cumulative; that is, all expenditures
in Category 1 are included in Category 2, and so on. See figure 1 for a
description of programs that are included in our spending categories.

10 ?Changes in State Spending on Social Services Since the Implementation of
Welfare Reform: A Preliminary Report.? Ellwood and Boyd, February 2000.

Figure 1: Four Categories of Welfare- Related Spending Category 4:

All expenditures in Categories 1, 2, and 3 plus:

TANF- related social services

child welfare, juvenile justice, substance abuse prevention and treatment,
services for the developmentally disabled, and mental health services.

Category 3:

Category 1 and 2

Health care

expenditures plus: health care expenditures for low- income families and
nonelderly poor adults. This

Category 2:

All category 1 category excludes

Support

expenditures plus: all long- term and

for the

child- support institutional care

working poor

passthroughs, costs.

emergency assistance, housing assistance, states? supplemental SSI payments,
state earned income tax credits (EITCs), transportation or wage subsidies,
child care,

Category 1:

head start, pre- K

Basic welfare services

programs for lowincome This is a prewelfare reform

children; model of helping families move

family formation and off welfare. This category

pregnancy prevention includes cash assistance and

programs. employment and training programs. It also includes one- time cash
payments such as diversion payments.

 To describe the measures states are taking to ?save for a rainy day? and
to analyze the impact these actions have on TANF reserve balances at the U.
S. Treasury, we analyzed the 10 states? budgets. To obtain a

broader view of the levels of unspent TANF funds in all 50 states, we
reviewed financial data reported by the states to the U. S. Department of
Health and Human Services (HHS). (States report key TANF financial
information on their quarterly ACF- 196 reports.)

Results in Brief The combination of a precipitous drop in caseloads and
fixed grant funding levels provided states with additional budgetary
resources they could use

to finance their new welfare policies. States faced many challenges
determining the appropriate mix among several policy options: saving for the
future, investing in programs serving current beneficiaries, and

addressing fiscal pressures in the states? budgets. Examining states?
choices in allocating federal and state funds for specific activities is an
important first step in understanding financing changes ushered in by the
TANF program. While addressing whether federal funds replaced or

?supplanted? state funds for specific activities- which is legal under TANF-
is one key issue, it is equally important to understand how states used this
new- found fiscal flexibility to support the broader goals of the TANF
program. For example, if a state replaced its own funds with TANF funds and
then shifted the freed- up state funds into other programs that serve the
working poor, the net impact of these shifts would be the same as if they
used the federal funds to fund these other programs. Therefore, it is

more informative to view the impact of supplantation in a broader fiscal
context consistent with the new objectives and goals of TANF.

In examining specific state funding decisions, we found that supplanting was
a common budget practice among the 10 states in our study. At the same time,
looking at the broadest level of TANF- related social services shows that
over time most states have maintained or even increased their own investment
to address the overall needs of low- income families. The

effect of state budgetary decisions on the fiscal balance between federal
and state governments differed depending on the range of programs we
analyzed. The greatest impact was felt in basic welfare services- most
states reduced their own spending and shifted more federal funds into those
programs. However, since welfare reform, the effect of states? decisions has
been to shift resources from this narrow category to support the broadest
array of programs addressing the goals of the TANF program;

nearly all states increased total funding in real dollars for these
programs, including health care and child welfare services. (See figure 2.)

Figure 2: TANF- Related Social Service Expenditures (Category 4) in 10
States for State Fiscal Years 1995 and 2000 25, 000 1995 dollars in millions

20, 000 15, 000 10, 000

5, 000 0

California Colorado

Connecticut Louisi

a na

Maryland Michigan

New York Oregon

Texas Wisconsin

Federal State

Source: GAO survey and analysis of state expenditure data.

An analysis of state budgets shows that the degree to which states
supplanted state spending varied considerably, ranging from about 25 percent
of the annual block grant in Connecticut and Oregon to 5 percent or less in
Colorado, Louisiana, and Michigan. State officials noted that supplanting
had the effect of allowing them to use federal and state funds more flexibly
to support a wide range of programs supporting needy families. Moreover,
state officials noted that supplanting enabled the states that had initiated
major social welfare programs prior to TANF to use federal funds to support
these efforts. However, in many cases, the MOE requirement limited the
extent to which states could use their federal funds to replace state funds.
This MOE requirement has led to a situation in

which many state officials say they are spending more than might be expected
in the face of the large caseload drop.

Some states report that their decisions on how to use federal TANF funds
were influenced by their concerns about the vulnerability of accumulated

unspent TANF balances to rescissions at the federal level. While many view
the balances held in reserve at the U. S. Treasury as the de facto safety
net for states to draw on in the event welfare costs rise, the potential
risk of losing these balances to competing priorities at the federal level
prompted

some states to draw down these balances for current program needs or fiscal
relief. As of September 30, 2000, all 50 states reported leaving a combined
total of $8. 6 billion in TANF funds unspent at the U. S. Treasury, or 13.5
percent of the total TANF funds that have been made available since 1997.
Little is known about whether these accumulating balances represent (1)
actual commitments states made but have not yet liquidated, (2) balances the
states are holding in reserve as rainy day funds, or (3) funds the states do
not need. This uncertainty is largely the result of variations in the way
states report allocations to subgrantees and counties. The lack of
transparency regarding states? plans for their unspent TANF

funds continues to weaken the effectiveness of congressional oversight over
TANF funding issues.

Our review suggests widespread disparities among all 50 states in their use
of TANF funds to prepare for the impact of an economic downturn on this
program. Some states report spending all their federal funds- essentially

holding nothing in reserve- while others report accumulated reserves
totaling more than their annual block grant. The level of these reserves was
not determined through a fiscal planning process analyzing potential

future draws on program resources caused by a downturn in the economy.
Furthermore, the design of PRWORA?s two safety net mechanisms for states-
the $2 billion Contingency Fund for State Welfare Programs (Contingency
Fund) and the $1. 7 billion Federal Loan Fund for State Welfare Programs
(Loan Fund)- is complex and restrictive. Therefore, neither one is likely to
be used by states in a fiscal crisis to obtain more resources for their
welfare programs.

The upcoming reauthorization of TANF presents Congress with an opportunity
to re- examine the new federal- state fiscal partnership it established with
enactment of PRWORA. In particular, Congress has the opportunity to consider
the design of the block grant including (1) how it has affected the balance
of fiscal responsibilities between federal and state governments to address
the goals of the program and (2) whether the program promotes the savings of
both federal and state funds to address a future ?rainy day.? TANF provided
states with increased flexibility; this flexibility combined with falling
caseloads has allowed many states to obtain fiscal relief while maintaining
or increasing their support for the broad purposes of the act. The addition
of a nonsupplant provision might

help ensure that all federal funds supplement existing state spending, but
it could also have the adverse effect of penalizing those states that took
the initiative to begin new programs before TANF, and it might limit the
intended breadth of the block grant by locking states into pre- established
funding priorities. From this perspective, a broad- based MOE calling for
states to maintain spending across a wide range of relevant programs might
limit substitution while preserving state discretion.

The shift from the open- ended match of AFDC to PRWORA?s fixed- dollar TANF
block grant also increased the states? exposure to fiscal risks in the event
of an economic downturn or increased program costs. Recognizing that both
the federal government and the states have an interest in contingency
planning, we have testified on ways to improve the federal contingency
mechanism as well as strengthen states? incentives to ?save for

a rainy day.? For example, Congress could realign the contingency fund MOE
and eliminate the monthly payment limitation on contingency fund access.
There are other options that might strengthen states? incentives to save.
Congress could amend the law to (1) allow states to count rainy day

funds toward a portion of their MOE and/ or (2) allow states to draw down
their entire TANF grant and save these funds in their own treasuries. There
are pluses and minuses to each of these ideas- the exact design and

implementation will be important in determining the actual impact of any
change. In keeping with the intent of the block grant, federal oversight
needs to focus not only on the use of TANF funds specifically, but also on
how states use multiple federal and state funding streams to support the
broadsweeping goals of the TANF program. Going forward, the overall fiscal
balance portrayed in this report may shift over time, particularly under
different economic conditions. Accordingly, we have recommended that the
Secretary of Health and Human Services consult with the states to initiate
periodic reporting on the entire federal- state fiscal effort relevant to

goals of the program.

State Budget Actions Much of the debate over whether- and if so, to what
extent- states used Shift Federal and State

federal funds to replace their own funds has hinged on the definition of
supplantation. Under our most limited definition of supplantation- Support
for Specific

replacing state dollars with TANF dollars for specific activities that are
Activities

allowable uses of TANF dollars- we found that all 10 of the states in our
study except California used TANF funds to replace some state funds from
existing programs. 11 However, the extent of the supplantation varied

considerably across states, as did the reasons given for supplantation.
Moreover, as discussed in a later section, replacement of state funds with
federal funds for specific activities was often accompanied by an increase
in state support for the broadest array of programs addressing the goals of

the TANF program. These fiscal actions took place within a framework of
state budget negotiations, which normally requires some trade- offs among
competing priorities. Moreover, the decisions were made at a time when
caseloads had

fallen dramatically and many states had significant levels of unspent TANF
funds. In allocating federal and state funds, states had to balance a number
of competing pressures. Among these were tensions between the need to draw
down TANF balances, save for future needs, and a concern about building
potentially new permanent state commitments that might outlast the federal
funds themselves. From this perspective, a supplantation

strategy allowed states to draw down their TANF balances to spend on current
needs or save for future contingencies without building permanent new state
commitments.

The programmatic and fiscal flexibility afforded states through the block
grant increases the opportunity to use federal funds to replace state
funding in welfare- related programs. The shift from a matching grant to a
block grant was in part intended to provide states with more flexibility.
Given the complex and evolving relationship between the federal and state

governments and their shared responsibilities for most domestic programs, it
is understandable that observers will have different views of 11 California
has not yet enacted a budget that contains apparent supplantation of federal

TANF funds. However, the Governor?s budget proposal for California?s state
fiscal year 20012002 proposed using $154 million in federal TANF funds to
replace state funds California already spent on its TANF program in 1997.
This retroactive claiming of federal funds was made possible because the
state recently certified that it had met its work participation rate in
federal fiscal year 1997. In so doing, California?s MOE requirement was
reduced in that year from $2. 9 billion to $2. 7 billion and the governor is
proposing to use unspent federal TANF funds to reimburse the state for those
?unnecessary? state expenditures.

supplantation. For example, advocates for low- income families lament the
failure to increase resources available to their constituents as a result of
supplantation. They note that when state resources are withdrawn and
replaced with federal funds fewer total funds are available for low- income
families than would be the case if federal funds were added. State officials
note that the flexibility to use federal funds instead of state funds in
some programs gives them opportunities to use freed- up state funds as
leverage to enact new or expanded programs that also benefit the low- income

population. In this debate, it is interesting to note that some states
prohibit their counties from supplanting at the local level. In Colorado and
New York, for example, counties are prohibited from using any of their TANF
funds passed through from the state to replace their county- generated
spending on social services.

Supplantation Varies Widely Budgets are the tool that state policymakers use
in deciding how to Among States distribute limited resources among different
policy choices. Once a state develops its priorities, decides which programs
it needs, and develops estimates of how many people will be served by the
programs, it can begin

to develop program budgets. In general, the extent to which states can use
federal funds to finance policy priorities is an important part of the
budget debate- maximizing the use of federal funds means that state revenues
go farther. There are certain limitations, however, on how states can
utilize federal funds. For example, states must spend federal grant money on

purposes specified by the federal law authorizing the grant. Furthermore,
grants may also require a form of cost sharing- either a maintenance-
ofeffort or a state match- in order to use federal funds. These features can
help ensure that states maintain their fiscal commitment to programs for
which they receive federal funds.

Although states are free under the law to replace state funds with TANF, in
fact two conditions facilitated supplanting: the state must have surplus
TANF funds available- in addition to those it needs to maintain program
levels- and it must have programs in place that are traditionally funded
with state dollars over and above the level necessary to meet its MOE
requirement. Without the former, it could run short of the funds it needs to
meet program demands and therefore have to find state general funds to

pay for those costs; without the latter, a state would run the risk of
fiscal penalties for not meeting its MOE requirement. During recent budget
deliberations, 9 of the 10 states we reviewed used federal TANF funds to
replace state funds in existing programs that met the

broad purposes of the act. In so doing, all of these states reduced their
own contributions to these programs. Since money is fungible, we cannot know
conclusively how they used these particular freed- up funds. As will be
discussed below, in some states budget documents clearly revealed that
supplanting enabled states to save their own funds for other programs; in
other states supplanting was viewed as a strategy to shift state funds from

traditional cash assistance to other low- income programs in keeping with
the flexibility in TANF. The extent to which states supplanted funds varied
significantly. In the most recent year for which data is available, five of
the states in our study- Connecticut, Maryland, Oregon, Texas, and
Wisconsin- used between 15 percent and 25 percent of their annual block
grants to finance state programs that had traditionally been financed with
state dollars. Four states- Colorado, Louisiana, Michigan, and New York-
used 10 percent or less of their TANF grants to replace state funding. In
the remaining state, California, no evidence of supplantation exists in its
enacted budgets to date, but this could change with its 2001- 2002 budget.
(See footnote 11.)

Interestingly, the role of local governments has also apparently affected
some states? ability to supplant. California, Colorado, and New York depend
on their counties to define welfare spending priorities. This financing
arrangement has limited the amount of TANF resources these states can

use for other spending priorities because any supplantation done at the
state level would reduce the amount of TANF funds available to the counties.
Intergovernmental dynamics within each state were often powerful enough to
dissuade these states from enacting budgets that reduced TANF funding for
their counties. In both New York and Colorado,

the states prohibit the counties from engaging in supplantation at the
county level. That is, counties are prohibited from using any of their TANF
funds to replace their own spending on social services.

Some States Supplanted to Budget officials cited many reasons for shifting
program financing, and

Address Fiscal Pressures most of the states were able to achieve significant
savings of state general fund dollars. For example, in Connecticut, when
spending in many TANFfunded

programs came in under budget, resulting in significant levels of unspent
TANF funds, state officials found other accounts serving needy families that
had traditionally been funded with state funds and used TANF funds instead.
State budget officials estimated they used about $40 million in each of
state fiscal years 1999 and 2000 to replace state funds with federal funds.
They explained that because the state funds were never

drawn from the state?s general fund accounts, the effect of using the TANF
funds in this manner was to enlarge their end- of- year general funds
surplus. Other states were also unwilling to accumulate large balances of
unspent

TANF funds. State officials told us they were concerned that leaving a
portion of their block grant in reserve might send the signal that the funds
were not needed and risk their rescission by Congress. In state fiscal year
2001, Michigan used about $27 million in TANF funds to pay for a lowincome
housing tax credit that had been paid for with state funds since 1973.
Michigan state officials said their decisions to eliminate TANF reserves
were in part due to mounting pressure to get the funds ?off the table?
before they could be taken back by Congress. Likewise, a Maryland state
budget official said that Maryland was motivated to supplant chiefly by the
perceived need to draw down its federal funds quickly or risk their
rescission. They were also concerned that if they left these funds unspent
it might affect future funding levels. In state fiscal year 2001, Maryland

identified several program accounts with annual expenditures of state funds
totaling about $30 million that, under the broad and flexible rules
governing TANF expenditures, could be funded with federal funds. In
developing the budget, the state replaced these state funds with federal
funds. Instead of using the freed- up state funds for nonwelfare activities,
however, the state used them to establish a dedicated reserve for its

welfare program. Some States Supplanted in In addition to realizing savings,
supplantation in some states was part of a Order to Increase Flexibility

broader strategy to shift the source of funds supporting state programs for
to Address Needs of LowIncome the working poor. Freeing- up state funds from
basic welfare programs in Families some cases gave states greater latitude
towards meeting their MOE requirements. For example, in 1997 when Texas
first deliberated how it would use the new TANF funds in its welfare
program, state officials

believed that cash assistance and training would be a good place to direct
state resources to count toward the state?s MOE requirement. However, as
caseloads dropped more quickly and to a greater extent than anticipated, it
became clear that state funds were not being expended at levels sufficient
to meet the MOE requirement. Subsequently, Texas decided to substitute
federal for state dollars for job training activities and to direct state
funds to other priorities that would spend out faster but that would still
count toward the MOE. Child welfare programs were seen as a priority because
of growing foster care caseloads and a state court ruling that called for
heightened investigation and prevention of child abuse and neglect.

Because spending on these programs was growing, these expenditures

were viewed as a better source of MOE financing than spending on basic
welfare services. The legislature increased appropriations for an expanded
child welfare program and directed the state agency to count the state?s
share of the increase towards it MOE. This was not the end of the matter,
however. After reviewing final TANF regulations, state agency officials
found that certain foster care expenditures included in the child welfare
expansion could not be counted toward the MOE requirement, which left them
short again. In the end, state officials identified as MOE a portion of
state aid used to support preschool programs. Since the state aid is

allocated based on the number of needy children in the district, a portion
of it qualifies as a MOE expenditure. The state spent nearly $320 million on
this effort in 2000.

States With Significant Prior Some states argue that supplantation, in
effect, enabled them to gain

Commitments Supplanted federal funding for state initiatives that began
prior to the implementation for Fiscal Relief

of federal reforms. From this perspective, supplantation allowed states to
avoid becoming locked into spending even when federal funds come available,
and also allowed for some federal recognition of their innovation and
progressiveness. Furthermore, since the MOE has prevented states from
reducing their spending as much as they otherwise would have been

able to given recent caseload declines, it is not surprising that states
such as Oregon argue that they deserve some fiscal relief. Oregon began its
welfare reform program in 1992 and, recognizing the importance of employment
and training and employment- related child care programs, invested heavily
in those programs even though the federal match was capped. By the time the
federal reforms were enacted, Oregon?s cash assistance caseload had already
fallen by nearly 40 percent. Oregon officials subsequently decided that they
had ?over- invested? in the federal JOBS program, and since 1997, the state
has withdrawn some of its own

investment in these programs and replaced state funds with federal funds.
Oregon argues that preventing supplantation would penalize it for being
progressive and ?out in front? in terms of welfare reform.

MOE Provision of TANF As we have previously reported, 12 some supplantation
is to be expected in

Limits Extent of any grant. We have previously reported that because of
supplantation, on

Supplantation average, every additional federal grant dollar results in
about 60 cents of supplantation. However, we also noted that certain grant
design features such as a strong MOE provision can limit states? ability to
supplant. 13 Consistent with this, in our study of 10 states, we found that
while most have enacted budgets that authorized some supplantation of
welfare funds, the level did not exceed 25 cents of a state?s block grant
dollar. For

example, program and state budget officials in Louisiana said that one of
the reasons for so little supplantation in Louisiana is the paucity of state
funds in social service programs not already earmarked to meet other federal
program requirements. The TANF MOE did more than limit the

leakage of funds from these programs- it also stimulated state spending
above and beyond what states would otherwise have spent. The following
examples in New York and Wisconsin clearly highlight the significant impact
that TANF?s MOE can have on state spending, particularly at a time when
caseloads were falling so much.

In New York, the level of fiscal relief possible largely depends on
withinstate cost sharing arrangements that were in place under the AFDC
program and continue under TANF. Under AFDC, the federal government paid 50
percent of AFDC costs and the state and local governments shared the
remaining costs equally. Because caseloads dropped and new programs and
services were slow to start for a variety of reasons, spending rates under
TANF were slower than expected. As a result, total expenditures

fell. The state feared it would not meet its MOE requirement if the
expenditure trends continued at the current rate and if the state continued
to reimburse counties with federal and state funds. Since the state operates
on a cost- reimbursement basis with its localities, if the state temporarily
had suspended the federal reimbursements, the effect would

have been to pay for all costs with state and local funds- making it easier
to reach the state?s MOE. To resolve the potential MOE shortfall, the state
halted any reimbursement to the localities for expenditures they had made in
the last quarter of the fiscal year. The state then counted all of the local
expenditures toward the state?s MOE, effectively spending more state and

12 Federal Grants: Design Improvements Could Help Federal Resources Go
Further (GAO/ AIMD- 97- 7, December 18, 1996). 13 See also Block Grants:
Issues in Designing Accountability Provisions (GAO/ AIMD- 95226, September
1, 1995).

local funds on these programs than otherwise would have been the case. The
state now closely monitors total expenditures and adjusts the federal share
of the reimbursement to ensure that the state will meet its MOE each

year. In Wisconsin, MOE also forced the state to spend more state funds than
it otherwise may have. According to legislative members and staff we
interviewed, pre- TANF, Wisconsin was spending less than the new MOE
required it to spend because it had already instituted substantial reforms
which helped lower caseloads, and in turn, expenditures. Because the TANF
MOE was pegged to a period when state spending levels were at an historic
high, Wisconsin had to raise its spending in order to meet its 75 percent
MOE level. During the 1999- 2001 budget deliberations, state budget analysts
projected that the state would fall short of its MOE requirement by about
$40 million because the number of W- 2 families receiving cash assistance
had fallen so dramatically. At the same time, final

TANF regulations were issued and allowed use of TANF funds to establish or
expand state earned income tax credit programs (EITC). This permitted
Wisconsin to substitute federal funds for about $52 million a year in state
funds that it was already spending on its state?s EITC. But because
Wisconsin was short on MOE funds to begin with, further reducing its MOE-
eligible expenditures only served to worsen the looming shortfall. Members
of the state?s Joint Finance Committee- a bicameral legislative committee-
however, realized that part of the freed- up state general revenues could be
used for a property tax cut, and the rest to expand certain programs for the
state?s low- income population. The funds used on

these expanded programs serving the low- income population could then be
used to meet the state?s MOE requirement so there would be no shortfall. The
compromise resulted in increased investment of state funds in a variety of
programs, even though some funds were diverted to a tax cut. The package
presented during the negotiations afforded the opportunity to raise TANF
spending on new programs even further, thereby increasing the

state?s investment on low- income families. Leadership stressed that these
new programs could not have been funded during the 1999- 2001 biennium in
the absence of the compromise arising from the freed- up state general
funds.

?New Spending Test? States? dissatisfaction with the MOE generally arises
because, in order to Augmented the Effect of avoid strict financial
penalties, they are forced to spend more on these

TANF?s MOE programs than they otherwise may have spent. Moreover, the MOE?s

stimulative effect on new state spending is further enhanced by what is
sometimes referred to as the ?new spending test.? The new spending test
allows state spending on eligible programs to be counted toward the TANF MOE
only to the extent that this ?new? spending is on behalf of eligible
families and it exceeds total spending in 1995 on those same programs.
Therefore, a state that had a particular MOE- eligible program in effect
prePRWORA may only count new expenditures on that program toward its MOE
requirement, but a state that implemented a similar program after

PRWORA may count the entire amount spent on the program towards its MOE.

The new spending test was particularly problematic for Michigan. The state
had financed a low- income homestead tax with its own funds since 1973.
However, because of the new spending test, only a portion of the credit
counted towards its MOE requirement. As a result, state budget officials
argue that the state is penalized, when compared to other states, for having
a low income tax credit before TANF was enacted; a state with no tax credit
could enact a new tax credit using only TANF funds.

Not only do some states find the new spending test overly restrictive, but
some states are also concerned that the provision could become difficult to
enforce. Compliance with federal financial requirements is generally
certified annually through the Single State Audit. 14 We spoke to many
auditors who were in the midst of developing audit plans to address
compliance with the new spending test. Several told us that developing these
plans was relatively straightforward; the auditor should simply be able to
establish a baseline for all the MOE expenditures the state was claiming and
then trace those programs back to 1995 and certify that spending claimed for
MOE was indeed new spending. However, these

plans could become more complex if states frequently changed the 14 Rather
than being a detailed review of individual grants or programs, a single
audit is an organizationwide financial and compliance audit that focuses on
accounting and administrative controls. A single audit is designed to advise
federal oversight officials and program managers on whether an
organization?s financial statements are fairly presented and to provide
reasonable assurance that federal financial assistance programs are managed
in accordance with applicable laws and regulations. GAO supports the single
audit concept for auditing federal assistance programs and believes it is
the fundamental mechanism for testing compliance with expenditure provisions
in TANF.

expenditures they were counting from one year to the next (i. e., changed
the programs for which they needed baselines.) For example, auditors in
California said that they would expect the lead TANF agency to certify the
baseline for the new spending test and they would audit the certification.
Because all expenditure data is archived after 5 years, auditing the annual
certification would be especially difficult and time consuming if the state

changes the programs it uses to meet its MOE requirement from year to year.
Officials in Colorado and New York, states that expressly prohibit their
counties from supplanting, also confirmed that enforcement of their own
state prohibitions is weak and that compliance is very difficult to test.
Furthermore, acquiring enough knowledge about every county?s social services
budget in order to establish a reliable baseline would be extremely

costly and might outweigh the benefits. Net Fiscal Effects of Once state
funds are supplanted with federal funds they are ?freed- up? and Funding
Shifts Varied become part of the larger pool of general revenues. While
budgets are a good blueprint to measure policy decisions, looking only at a
specific Among States

change authorized in a budget may be misleading. For example, if a state
supplanted federal funds for state funds, then shifted its own funds into
other programs that serve the working poor, the net impact of these shifts
would be the same as if they used federal funds directly to fund these other

programs. Accordingly, the impact of the funding shifts must be viewed in a
broader context consistent with the new objectives and goals of TANF. In
fact, under our broadest definition of TANF- related social services, eight
states maintained or increased their net investments in constant dollars for

programs that serve low- income families and individuals as they move from
welfare to work or strive to stay off public assistance rolls; only New York
and Wisconsin decreased their net investment in these programs.

To provide a broader context for considering the effects of funding shifts,
we reviewed state expenditures of both federal and state funds on a wide
array of social service programs that target assistance to low- income
individuals and families in order to help them avoid dependency on public
assistance. We surveyed 15 states for their total expenditures on these
programs in 1995 (to establish a prewelfare reform baseline) and for 2000

(to measure any change in total spending). We asked states to break down
their total expenditures by funding source- state funds or federal funds.
Under AFDC, the fiscal partnership between the federal government and

the states was defined by the federal match rate each state received to help
finance its welfare programs. 16 Under TANF, the block grant is fixed-
states decide how much of the federal grant to spend and how much to save.
First, a state decides how much to spend in total on a set of programs.
Second, a state decides the right mix of federal and state funds

(subject to the MOE limitation) to use to finance these programs. This

?mix? is the new fiscal balance and representative of the share of funding
from each partner- the states and the federal government. Essentially the
state controls both the state and federal share. In our analysis, therefore,
federal spending is not defined by the level of the TANF grant allocated to
each state but rather by how much of the grant the state chose to spend.

Further complicating our analysis is the fact that the populations served by
the AFDC program and the TANF program are dramatically different. Under
AFDC, caseload was defined principally by the number of families receiving a
cash assistance payment. Under TANF, however, there is no

single definition of a ?case.? Because states are free to use TANF funds to
finance programs with a variety of different eligibility requirements,
caseload can vary by state, and even by program within a state. This means
that while a family might be eligible for TANF- funded child care subsidies,
the same family might not be eligible for- or request- monthly cash
assistance benefits., Because cash assistance caseloads have fallen in
recent years, a decline in welfare expenditures might be expected. It might

be unreasonable, for example, to expect a state to maintain a higher level
of expenditures on cash assistance while fewer people sought assistance.

15 We adapted a survey developed by Deborah A. Ellwood and Donald J. Boyd of
The Nelson A. Rockefeller Institute of Government for their report ?Changes
in State Spending on Social Services Since the Implementation of Welfare
Reform: A Preliminary Report (for more information on our survey see
appendix I of our report). 16 Each state defined eligibility within federal
guidelines and set benefit levels. The federal government paid from 50
percent to 79 percent of the costs.

However, because TANF is intended to serve a broader purpose than did AFDC,
it is reasonable to expect that states might have used some TANF funds to
invest in programs that serve the goals of TANF rather than use them to
expand programs targeted only to the original AFDC population. Thus, it is
appropriate to consider cumulative categories of welfare

expenditures in order to understand how states financed the transition of
low- income families from dependence to productive employment. To this end,
we grouped individual program spending into four cumulative categories:

 Basic welfare services: cash assistance grants, training programs, and
diversion payments.

 Support for the working poor: category 1 programs plus programs that many
states use to provide transitional assistance to working poor families,
including child care subsidies, transportation subsidies, a state EITC, or
low- income housing assistance.

 Health care: category 2 programs plus health care expenditures for
lowincome families and individuals.  Other TANF- related social services:
category 3 programs plus other TANF- related social service programs such as
child welfare, transitional services for the developmentally disabled,
substance abuse treatment, and mental health services.

In the body of this report, we discuss two key aspects of fiscal trends in
1995 and 2000: (1) state spending in each category in constant 1995 dollars
and (2) the percent change in the state- financed and federal- financed
portions of total expenditures. To observe a net change in spending within a
category the impact of both measures must be assessed. For example, if a
state increases state funding on a set of programs and federal spending
increases by a larger amount (or the state draws down even more federal
funds), the state- financed share of total expenditures would decrease even
though the state?s investment in dollar terms increased. Conversely, if a

state reduced state spending for a particular set of programs on which the
decrease in federal spending was larger, the state- financed share of total
expenditures would have increased even though the state?s investment in
dollar terms decreased. So, in order to observe a net divestiture of state
funds in our analysis, for example, two conditions must be met: (1) a state
must have reduced its own spending in dollar terms from 1995 levels 17 and
(2) the state- financed portion of total spending must have also declined
from 1995.

Category 1: Basic Welfare Under the most narrow definition of welfare
(Category 1)- one in which

Services the definition of a case is similar to that under AFDC prewelfare
reform-

total spending (federal and state combined) generally declined, but somewhat
less than the dramatic decline in the number of cash assistance families.
The correlation between spending and caseload is not perfect even in this
most narrow definition of welfare spending. In large part this is because
the orientation of most states? welfare programs towards jobs skills and job
readiness programs required increased investments in workrelated activities;
as cash assistance budgets were declining, more recipients were required to
participate in work or work- related training programs. Significantly, while
spending dropped in this traditional category under TANF, a comparable
caseload decline under the old openended

matching design would have triggered a more substantial cut in spending
because of the way in which AFDC was financed. In eight states, total
expenditures on these programs declined more than 35 percent in real terms
although caseloads declined by more than 17 We conducted our analysis in
both nominal dollars and constant (real) 1995 dollars. While either basis
leads to similar results, for clarity?s sake, we discuss our findings in
terms of real dollars throughout this report. A discussion in nominal
dollars may be found in state appendixes II- XI.

45 percent over the same period. (See table 1.) The other two states- Texas
and Louisiana- present a different picture. Spending in these two states
with historically low benefit levels did not correlate with the states?
caseload declines. Although caseloads declined by more than 50 percent in
both states, total spending remained constant in Texas and declined by about
20 percent in Louisiana. As a result, spending per family more than

doubled in both states. These two states also increased their monthly cash
assistance grants- in 1999 Texas raised its monthly benefits for the first
time in 15 years while Louisiana increased its benefits in 2000 for the
first time since 1981. Although notable, this probably had little effect on
overall spending trends in this category. A more likely influential factor
was the rapid increase in new spending on employment and training.

Table 1: Changes in Basic Welfare Services Percent change in total

(federal and state) Percent change in families

expenditures receiving cash assistance

(State fiscal years (January 1995 to June

State 1995 - 2000)

2000)

Colorado -64 -72 California -47 -47 Wisconsin -46 -78 Oregon -45 -58
Maryland -44 -64 Michigan -42 -66 Connecticut -42 -55 New York -37 -46
Louisiana -21 -69 Tex as 0 - 54 Note: Percent change in total expenditures
calculated using 1995 dollars. Source: GAO survey and analysis of state
expenditure data.

While all 10 states reduced their own spending on Category 1 programs, we
observe more varied trends in terms of the fiscal partnership. (See table
2.) In 6 of the 10 states (see bolded text in table 2) the state- financed
share of

total expenditures declined relative to 1995 levels. Of the remaining four
states, one maintained its share of spending, and three spent more of their
own funds on these activities than the federal government relative to the

1995 ratios.

As noted above, a state?s ability to alter the mix of federal and state
funds used to finance its welfare program depends on whether or not it has
enough expenditures in other programs- outside the narrow definition of
basic welfare services- to meet its MOE requirement. Although the three
states- New York, California, and Wisconsin- that increased their share of
total expenditures on these activities in 2000 also reduced their own
spending on these programs, they found that the MOE limited their ability to
further divest funds.

Table 2: Category 1- State Expenditures on Basic Welfare Services (State
Fiscal Years 1995 and 2000) Percent change in statefinanced Percent change
in state

portion of total State funds

expenditures Oregon -82 -22 Colorado -76 -14 Maryland -56 -9 Connecticut -49
-6

California -44 +3 Michigan -42 0 New York -34 +4

Louisiana -28 -2 Texas -22 -8

Wisconsin -4 +32 Note: Percent change in state funds calculated using 1995
dollars. Source: GAO survey and analysis of state expenditure data.

Category 2: Support for the While six states reduced the level of state
resources devoted to basic

Working Poor welfare programs (Category 1) in both constant dollar terms and
as a share

of total expenditures, many state officials suggested that states have used
these ?freed- up? resources to invest in new (or expand existing) programs
using a broader definition of assistance to low- income families, such as
child care programs; work- supports, such as a state- EITC; and other
postemployment transitional services- essentially, support for the working
poor (Category 2). When considering states? cumulative funding for this
broader category, five states still divested some of their resources from

these programs in terms of both constant dollars and percentage- share of
total expenditures; Texas was the only state that did not reduce its own

spending in dollar terms from 1995 levels -instead, it actually increased
spending in dollar terms on Category 2 programs. (See table 3.) It is worth
noting, however, that the states? share of total spending in these five
states declined less than those observed in Category 1. Moreover, states?
shares

for this category are cumulative and reflect both their shares in Category 1
and Category 2. When looking at the spending for Category 2 alone, some
states with large reductions in Category 1 programs significantly increased
funds for programs that were added to the analysis in Category 2- i. e.,
programs that more broadly support the working poor. For example, in
Colorado, state spending declined by 76 percent in Category 1 compared to
only a 7- percent decline in Category 2. Because state spending tripled on

child care and other work support programs- programs appearing in our
analysis for the first time in Category 2- the declines observed in Category
1 were largely offset.

Table 3: Category 2- Expenditures on Programs Supporting the Working Poor
(State Fiscal Years 1995 and 2000) Percent change in

state- financed Percent change in

Percent change in portion of total

State total funds state funds expenditures

New York -25 -20 +5

Connecticut -19 -21 -2

California -15 -12 +2

Oregon -12 -32 -7

Wisconsin -11 -11 0

Maryland -9 -17 -4

Michigan -4 -4 0

Colorado +2 -7 -3 Louisiana +2 -7 -2

Tex as +28 +11 - 6 Note: Percent change in total and state funds calculated
using 1995 dollars. Source: GAO survey and analysis of state expenditure
data.

Category 3: Support for the State officials cite many other budget
priorities that have demanded their

Working Poor and Related attention in recent years. Many of these priorities
also benefit low- income

Health Care Costs families but are not included in Categories 1 and 2 of our
analysis of spending under welfare reforms. For example, some states
expanded

health care services for low- income families and their children. During the
time frame of our study, the federal/ state partnership to provide health
insurance to low- income children (The State Children?s Health Insurance

Program- SCHIP) was enacted. In addition, Wisconsin and Oregon began
enhanced Medicaid programs for low- income individuals not eligible for
Medicaid. In Oregon, total health care spending for the low- income working
poor tripled since 1995 while the state- funded share of these costs more
than doubled.

Although TANF funds cannot be used directly to expand access to Medicaid and
there are limitations on the states? ability to use their MOE funds to
expand access to Medicaid, we asked states to report on certain health care
expenditures given the importance of this key benefit in helping people move
into the workforce. Specifically, we asked states to provide expenditure
data on their health care- related costs for low- income individuals but
asked them to exclude any costs associated with their elderly population,
including long- term and institutional care. State

funding for low- income health programs is generally driven by the federal
matching requirements of programs such as Medicaid and SCHIP rather than the
TANF program. Accordingly, from this point forward in this report,

any observable shifts in the federal/ state fiscal balance cannot be solely
attributed to changes brought about through the flexibility of the TANF
block grant or its MOE requirement. Nevertheless, our study still reflects
the use of freed- up funds from welfare programs to finance new or expanded
programs in health care for the same population. Total spending on the
cumulative costs of basic welfare services plus other programs supporting
the working poor, including their health care costs (Category 3), continues
to vary considerably among the states. Nevertheless, as we broaden the scope
of our analysis, states appear more likely to have maintained or increased
their investments in these programs. (See table 4.) Accordingly, only three
states (see bolded text in table 4)- as compared to six states in Categories
1 and 2- appear to have reduced their

investments in both constant dollar terms and as a percentage share of total
funds. Three states have increased their investment in these programs, two
(Michigan and California) have maintained the fiscal balance, and while the
remaining two states (Maryland and Texas) shifted the fiscal balance
slightly in their favor, they nevertheless increased their own investment in
these programs in dollar terms.

Table 4: Category 3- Expenditures on Programs Supporting the Working Poor
and Related Health Care Costs (State Fiscal Years 1995 and 2000)

Percent change in state- financed Percent change in

Percent change in share of total

State total funds state funds expenditures

Connecticut -10 -11 -1

California -6 -7 0

New York -2 -9 -5

Michigan -1 -1 0

Wisconsin +2 -3 -2

Louisiana +2 +8 +1 Texas +12 +11 -1 Maryland +16 +12 -1 Colorado +46 +56 +3
Oregon +72 +81 +2 Note: Percent change in total and state funds calculated
using 1995 dollars. Source: GAO survey and analysis of state expenditure
data.

Because money is fungible, this analysis cannot determine with any certainty
whether the particular state funds freed up by TANF funds were used to
finance expansions in health care for low- income individuals or families.
However, it does show that in five states state funding increased in

dollar terms for these programs over the study period. In fact, in Oregon
and Colorado, total expenditures on health care increased dramatically and
the state share of total expenditures slightly increased relative to the
federal share. Only three states (see bolded text in table 4) reduced their
investments in these programs in constant dollar terms as well as in terms
of the state- financed share of total expenditures.

Category 4: TANF- Related We further expanded the definition of assistance
to test whether states may

Social Services have used their freed- up funds to expand other social
service programs for

low- income families by asking states to report their expenditures in 1995
constant dollars on a wide range of services, including child welfare,
mental health services, substance abuse programs, and programs designed to
help the developmentally disabled become more self sufficient (Category 4-
Other TANF- related Social Services). Again, we asked states to exclude any
costs that were not transitional in nature, such as costs associated with
long- term institutional care.

As expected, states continue to be more likely to have maintained or
increased their fiscal commitments to Category 4 programs- the most macro
level of our analysis. (See table 5.) Only two states (see bolded text in
table 5)- compared to three states in Category 3- appear to have reduced the
amount of funds invested in these programs in terms of

constant dollars and percentage share of total funds. Eight states increased
their investment in this broad program area. In terms of the overall fiscal
balance between the federal government and these eight states, most
maintained or increased the state- financed share of total expenditures
relative to 1995 levels. In some states the analysis provides a dramatically
different picture as we consider broader definitions of spending categories.
For example, in Category 1 programs, of the 10 states in our study, Oregon
and Colorado

displayed the most severe spending reductions in both constant dollar terms
as well as percentage share of total expenditures: -82 percent (dollars) and
-22 percent (share); and -76 percent (dollars) and -14 percent

(share), respectively. In Category 4, however, these same states show the
most dramatic increases in investments in these programs: 42 percent
(dollars) and 52 percent (share) in Colorado; and 61 percent (dollars) and
65 percent (share) in Oregon. Table 5: Category 4- Expenditures on TANF-
Related Social Services for LowIncome

Families and the Working Poor (State Fiscal Years 1995 and 2000) Percent
change in

Percent change in Percent change in

State total funds state funds state share New York -4 -11 -5

Connecticut -1 +2 +2 Louisiana +5 +13 +2 California +5 +5 0

Wisconsin +7 -1 -4

Michigan +13 +15 +1 Maryland +17 +11 -3 Tex as +19 +17 - 1 Colorado +42 +52
+3 Oregon +61 +65 0 Note: Percent change in total and state funds calculated
using 1995 dollars. Source: GAO survey and analysis of state expenditure
data.

States? Contingency Paradoxically, while much attention has been placed on
whether states Plans Are Undeveloped

have replaced their own spending with federal funds, states have also been
under considerable scrutiny for not spending their TANF funds more quickly.
Finding the right balance between saving for future needs and investing
resources today in programs to help people make the transition from welfare
to work continues to be one of the main challenges states face as they
develop strategies to address the needs of low- income families. If states
wish to set aside reserves for future welfare costs, they have two

options: they can save federal TANF funds or they can save their own funds.
If they elect to save federal TANF funds, they must keep these funds in the
U. S. Treasury until the funds are needed.

As of September 30, 2000, states reported leaving $8. 6 billion in unspent
TANF funds at the U. S. Treasury; this amounts to 13. 5 percent of the total
TANF funds awarded since 1997. HHS reporting requirements do not ask states
to signal their intentions for these funds. As a result, congressional
decisionmakers lack reliable information on whether states are saving
adequately for a ?rainy day.? However, the disparity in the reported
balances among the states suggests that some states may not be well prepared
to address any increased demand in their welfare programs while

others may have saved substantially more than they might need. Although 5 of
the 10 states in our study consider a portion of their unspent funds to be

?rainy day funds,? it does not appear that the amounts designated as
reserved were calculated based on any sort of contingency planning or
analysis.

Impact of External Factors As we previously reported, it is unclear what
impact a major economic on TANF Caseloads Is

downturn or recession will have on welfare participation given the
significant reforms in national welfare policy. 18 While research literature
Uncertain generally suggests that caseloads will rise in an economic
downturn, there is substantial uncertainty regarding the extent of the
increase. However, any increase in caseload will almost certainly mean
higher costs. During a fiscal crisis, state policy makers face difficult
choices regarding

whom to serve, for how long, and with what services. Each of these ?hard
choices? must be financed in the context of fiscal limitations- including
legislative restrictions, constitutional balanced budget mandates, or 18
Welfare Reform: Challenges in Saving for a ?Rainy Day? (GAO- 01- 674T, April
26, 2001).

conditions imposed by the bond market- on a state?s ability to increase
spending, especially in times of fiscal stress. For example, revenues may
come in lower than expected during an economic downturn and a state?s
enacted budget can fall into deficit. State balanced budget requirements
often motivate states to both reallocate resources within their budgets and

cut program spending or increase taxes during recessions. This can
exacerbate budgetary stress caused by caseload volatility.

For these reasons prudent fiscal planning, especially contingency budgeting
for a fiscal rainy day, becomes particularly important. In general, during
any sort of fiscal crisis, a state?s need to cut spending or increase
revenues can be alleviated if it has accumulated surplus balances in rainy
day funds- these surpluses may be used to cover a given year?s deficit. 19
However, unless there are reserves specifically earmarked for low- income

families, welfare programs will have to compete with other state priorities
for any of the rainy day funds. Since the fiscal risks of welfare programs
have now largely shifted to the states, it is reasonable for Congress to
inquire as to whether states are prepared for these risks.

Data on State Reserves Although many might view the unspent balances left at
the U. S. Treasury as Have Significant Limitations a ?de facto rainy day
fund? for future welfare costs, states do not report

unspent balances in a consistent manner. This makes it difficult to
ascertain how much of these balances are truly uncommitted and therefore
available for future contingencies. As a result, federal policymakers lack
complete, reliable information to help assess states? plans for economic
contingencies, including whether the levels of available funds are adequate,
and whether all states have access to these funds.

Current Reporting Requirements Reporting requirements should enable
collection of data that will assist Do Not Clarify the Nature of
policymakers in their oversight responsibilities. However, largely because
State Reserves

welfare administrative structures differ dramatically across states,
information on unspent TANF balances is not compiled in the same way in all
states. This makes it difficult to ascertain how much of these balances are
truly uncommitted and therefore available for future contingencies. As

a result, the information on unspent balances is not comparable, and
provides little useful information on the nature and adequacy of states?
resources for future fiscal risks.

19 For additional information, see Budgeting for Emergencies: State
Practices and Federal Implications (GAO/ AIMD- 99- 250, September 30, 1999).

HHS regulations regarding unobligated and unliquidated balances also
contribute to a lack of clarity. 20 Under HHS regulations, if a state has
allocated a portion of its TANF grant to a rainy day fund, the state must
report these balances as unobligated. However, unobligated balances include
more than these rainy day fund allocations. These balances can also include
funds for which the state has made no spending plans or funds the state has
reserved for activities in future years. As a result, it is impossible to
tell from the unobligated balances alone whether a state has

contingency plans. For example, in developing a budget for a new child care
program, officials in Wisconsin assumed that once the program was fully
subscribed it would require all available resources- including any

unobligated TANF funds from previous fiscal years. State officials said that
even though at the end of federal fiscal year 2000 the state reported $40
million TANF funds as unobligated, the state has programmed these funds to
pay child care subsidies to low- income families in future reporting

periods. Unspent TANF funds may also be reported as unliquidated
obligations. These are funds for which, to varying degrees, an underlying
commitment exists for the funds- either through a contract for services for
eligible clients or to a county for expenses it will incur in operating a
countyadministered

welfare program. However, it is unclear how much of these unspent balances
are actually committed for future needs. For example, both California and
Colorado have county- administered welfare systems. These states pass most
of their annual block grants directly to the counties. As caseloads have
continued to decline in both states, the budgets overestimated

actual expenditures leaving considerable balances. Although these funds
remain in the U. S. Treasury until the county needs to spend them, they
remain as unliquidated obligations committed to the counties. California
reports that it has over $1.6 billion in unliquidated TANF

obligations but only about $2.5 million in unobligated balances, implying
that nearly all these funds are earmarked. Recently, California amended its
state statute to allow the state to re- obligate some of these funds, if
necessary, among counties. Colorado, which reported about $95 million in

unliquidated obligations as of September 30, 2000, and passes virtually all
TANF resources to the counties, did likewise. As of June 30, 2000 the state

20 Under HHS regulations the states use two categories to report on the
status of their unspent TANF funds: (1) unobligated balances represent funds
not yet committed for a specific expenditure by a state and (2) unliquidated
obligations represent funds states have committed but not yet spent.

estimated that counties hold about $67 million in reserves, or about 70
percent of the total unliquidated obligations, for future contingencies. As
the above examples illustrate, the difference between unobligated

balances and unliquidated obligations is often unclear, and varies by state.
A significant portion of California?s and Colorado?s unspent funds are not
yet actually committed for a specific expenditure but these facts cannot be
determined based on the aggregate data- in part because of the way HHS
requires states to report funds. Reporting a significant share of their
unspent balances as unliquidated obligations implies that there is an
underlying commitment on these funds when, in fact, these funds are no

more committed than the funds Wisconsin must report in its unobligated
balances but which are budgeted for expected outlays in Wisconsin?s child
care subsidy program. California Holds Reserves in In addition, we found
that in California, the confusion regarding the proper Counties

reporting of certain expenditures further blurs the aggregate levels of
unspent TANF balances. California has transferred more than $1 billion of
the state?s TANF dollars to its counties in the form of incentive payments-

yet these funds remain unspent and are on deposit in county bank accounts.
As noted previously, California relies on its counties to define welfare
priorities at the local level, and California transfers most of its

federal funds to the counties to finance programs that help make the
transition from welfare to work. For most services the counties receive
federal funds as they need them and until that time the funds remain in the
federal Treasury. However, one state program pays counties incentive
payments that are designed, in part, to motivate the counties to help the
state meet certain TANF performance goals. Specifically, the state makes
awards to its counties based on the number of welfare cases that have left
the rolls for employment and the number of cases the counties have

diverted from the rolls and are employed. To date the counties have received
awards totaling more than $1 billion. These funds are federal TANF funds
that have been drawn from the U. S. Treasury and deposited into county bank
accounts. According to state law, counties must spend these funds on TANF-
related purposes. The state has also issued guidance to the counties that
any interest earned while these funds are on deposit must be reinvested into
the TANF program. Apart from the aforementioned guidance, state statute
defers to the counties regarding how best to use the funds. While state
officials have said that these incentive payments were not intended for

counties to use for contingency budgeting, counties are not prohibited from
using these funds to cover unexpected costs if needed. In February 2001,
state officials reported that counties have spent only $46 million- about 4
percent- of these funds. This means counties in California have about $1
billion in unspent TANF funds on account that are

not included in the ?unobligated balances? totals reported to HHS. HHS
believes that California could be in violation of CMIA which seeks to limit
the interest costs to both federal and state governments by coordinating the
timing of a states? draw down from the Treasury with the timing of the
state?s expenditure needs. This coordination ensures that the timing of a
state?s draws do not favor either party and neither partner incurs
unnecessary interest costs. California maintains that these payments are
valid expenditures to service providers- their counties- under a

performance- based agreement to administer welfare services. Improved
Reporting

The lack of transparency regarding states? plans for their unspent TANF
Requirements Could

funds prompted us in 1998 to recommend that HHS and the states work Improve
Federal Oversight

together to explore options for enhancing the information available
regarding these balances. 21 Although HHS, the National Governor?s and
Provide States With

Association (NGA), and the National Conference of State Legislatures
Incentives to Save (NCSL) all agreed with us that more information regarding
unspent TANF balances would be useful, little progress has been made in
implementing this recommendation, and HHS? final regulations, issued on
April 12, 1999, did not address the issue. States were already concerned
that the TANF reporting requirements would pose a substantial burden on
state program administration and argued that adding another reporting
requirement to allow states to signal their intentions for their unspent
balances would only

add to those burdens. However, the lack of useful information on these
balances continues to weaken the effectiveness of congressional oversight
over TANF funding issues, including how well prepared states may be to
address a fiscal downturn.

21 Welfare Reform: Early Fiscal Effects of the TANF Block Grant (GAO/ AIMD-
98- 137, August 18, 1998).

Our 1998 recommendation proposed a strategy that state and federal officials
had tried before and found to be successful. In 1981, a number of
categorical grants were block granted to states to provide maximum
flexibility in developing and managing programs, along the same lines that
TANF was designed in 1996. However, due to variations in the way states

reported information to the federal government on activities funded by some
of these block grants, Congress had no national picture of the grants?
impact. States and some national organizations recognized that these
aggregate data were important and developed their own strategies to collect
the data. 22 We found that a cooperative data collection approach was easier
to implement when (1) there was federal funding to support

data collection activities, (2) national- level staff worked with state
officials, and (3) state officials helped in systems design. We continue to
believe that better information on the status of these unspent balances is
crucial to

effective oversight and could even enhance states? incentives to save some
of their TANF funds. Absent credible information on balances, there may be a
greater risk that Congress could take action to recoup TANF funds- a
prospect that has prompted some states to draw down and spend their TANF
funds rather than leave them in the Treasury.

Contingency Planning As we have previously mentioned, states can save for a
rainy day in two Receives Little Attention as

ways: they can save federal TANF funds and/ or they can save their own
States Cite Few Incentives funds. However, states have noted significant
disincentives with both to Save

options. As previously noted, state officials voiced concern that
accumulating unspent federal TANF balances might signal that these funds are
not needed and that they have been under considerable pressure to spend
their TANF grants more quickly. Similarly, although some states have
accumulated a portion of their own funds in general purpose state rainy day
funds, these funds are not a guaranteed safety net- welfare would have to
compete with other programs for these dollars in the event funds are needed.
Moreover, dedicated state- funded welfare reserves are expensive because
these funds cannot count towards a state?s MOE.

Although some states consider portions of their unobligated TANF balances to
be rainy day funds, it does not appear that the amounts reserved were based
on any kind of contingency planning or analysis by the

22 Block Grants: Federal- State Cooperation in Developing National Data
Collection Strategies (GAO/ HRD- 89- 2, November 29, 1989) and Block Grants:
Federal Data Collection Provisions (GAO/ HRD- 87- 59FS, February 24, 1987).

state. According to NGA, few states have engaged in a systematic fiscal
planning process to project their needs under a variety of economic
scenarios. Our review of 10 states supports NGA?s findings. Five of the 10
states we studied told us that they consider a portion of the funds left at
the

U. S. Treasury to be rainy day funds for unanticipated program needs. But
the levels of the reserves established in those five states do not appear to
have been determined through a fiscal planning process that reflects
budgetary assumptions about projected future needs. Instead, these states
merely designate residual TANF funds not already appropriated by the state
legislature for other purposes as constituting the state?s welfare rainy day
fund. Although many states have healthy general- purpose rainy day funds for
which all programs would compete for funds during times of fiscal stress,
only one of the states in our review, Maryland, has earmarked state funds in
a reserve specifically for contingencies in its welfare program. Reserving
funds specifically for welfare requires state decisionmakers to make
tradeoffs among competing needs. Any funds a state sets aside for future
welfare contingencies cannot count toward a states? maintenance of effort in
the year they are reserved- in order to qualify as MOE, the funds must be

spent. As a result, it is a very expensive proposition indeed for a state to
budget both for a welfare reserve and to meet its MOE. In state fiscal year
2001, Maryland identified nine program accounts with annual expenditures of
state funds totaling about $30 million that, under the broad and flexible
rules governing TANF expenditures, could be funded with federal funds. In
developing the budget, the state replaced these state funds with federal
funds. Instead of using the ?freed- up? state funds for nonwelfare
activities the state used them to establish a dedicated reserve for its
welfare program. In effect, Maryland was able to draw down federal funds and
save them for future needs rather than use those funds to expand existing or
create new programs.

Design of Federal While the ability to carry forward TANF balances is likely
viewed as the

Contingency principle mechanism by which states can prepare for a rainy day,
PRWORA also created two safety- net mechanisms for states to access
additional Mechanisms Is federal resources in the event of a recession or
other emergency- the

Complex and $2 billion Contingency Fund for State Welfare Programs
(Contingency

Restrictive Fund) and the $1. 7 billion Federal Loan Fund for State Welfare
Programs

(Loan Fund).

The Contingency Fund is authorized through 2001, at which time it expires.
The President?s FY 2002 budget proposal did not include a request to
reauthorize the Contingency Fund. Because of a provision in the Adoption and
Safe Families Act of 1997 that reduced the TANF Contingency Fund by $40
million, the current balance in the Contingency Fund is $1.96 billion. 23
States are deemed ?needy? and eligible to receive funds from the

Contingency Fund if they trigger one of two criteria: (1) the state?s
unemployment rate exceeds 6.5 percent for 3 months and is equal to at least
110 percent of its rate in the same period of the previous year or (2) its
average monthly food stamp caseload for the most recent 3- month period is
equal to at least 110 percent of the average monthly caseload from the same

3- month period in fiscal year 1994 or 1995. Once eligible, a state must
certify that it has increased its own current spending to prewelfare reform
levels before it can gain access to the fund. Requiring states to increase
their own financial stake in their welfare programs before giving them
additional federal funds is, in principle, a reasonable approach that seeks
to balance both the federal government?s interest in ensuring that states in
trouble have access to additional funds and its interest in ensuring that
states have done everything possible to address the shortfalls before
turning to the federal treasury. However, this could prove to be too
demanding. Not only does the statute require states

to bring their spending up to the prewelfare reform levels at a time when
states are experiencing fiscal stress, but PRWORA establishes a different
and more challenging base for the Contingency Fund?s MOE. While a state?s
MOE requirement under the basic TANF program can include state

funds expended under certain state programs (known as ?separate state
programs?) and child care expenditures, the MOE requirement for the
Contingency Fund does not include these items. Because states spend a
significant share of their MOE funds on activities that do not qualify as
Contingency Fund MOE expenditures, state budget officials told us that,

rather than shifting their spending priorities to meet the Contingency Fund
MOE, they would find other ways to manage deficits in their TANF budgets
before they would consider turning to the Contingency Fund.

23 The Adoption and Safe Families Act of 1997 reduced the contingency fund
for state welfare programs by $40 million over four years (P. L. No. 105-
89, sect.404, 111 Stat. 2134).

In 1997 seven states 24 and the District of Columbia qualified for
contingency funds. However, only two states requested and were awarded
contingency funds- North Carolina and New Mexico. In the end, only New
Mexico complied with the Fund?s requirements and accepted $2 million. No
state has used the Fund since 1997.

Equally important as the requirement that states raise their own financial
commitment in order to gain access to additional federal funds is a
requirement that states share in all additional program costs- even beyond
the MOE requirements. Requiring a match encourages states to be more cost-
conscious than if the costs of an expanding caseload were covered

only with federal dollars. While the Contingency Fund requires states to
match all federal dollars at the states? FMAP rate, 25 the statute goes a
step further. The statute limits the monthly draws to one- twelfth of 20
percent of a state?s annual block grant. This limitation requires a complex
annual reconciliation process to certify that the state meets its matching
requirement but also that it did not receive more than its monthly

proportional share of contingency funds. (See figure 3.) 26 Prorating a
state?s draws from the Contingency Fund- especially if the state qualifies
for a period that spans 2 federal fiscal years- reduces the share of federal
funds to which it is entitled. This effectively increases the matching
requirement (even higher than required under AFDC), thus raising the state?s
costs for gaining access to the funds.

24 These states are Alaska, California, Hawaii, New Mexico, New York, North
Carolina, and Washington. 25 Under AFDC, state spending was matched at a
rate based on each state?s per capita income. This rate, FMAP, is also used
for other federal- state matching programs such as Medicaid. It ranges from
50 percent for wealthy states to 83 percent of poorer states.

26 For more information see Welfare Reform: Early Fiscal Effects of the TANF
Block Grant.

Figure 3: The Contingency Fund?s Annual Reconciliation Process

As currently structured, the reconciliation process favors states that are
?needy? within a single federal fiscal year compared with those that are
?needy? in months that overlap consecutive federal fiscal years. A state
that is needy for all 12 months during a federal fiscal year would have to
match all funds drawn at its applicable fiscal year FMAP rate with no
adjustments for the number of months it was eligible because it was needy
throughout the year. However, a state that is ?needy? for 12 consecutive
months that span 2 federal fiscal years (e. g., 6 months in each year) with
an identical FMAP rate will see its federal match rate reduced by half
because of the adjustment made for the number of months the state was needy
in each year.

To illustrate, the state that was needy for an entire federal fiscal year
and was eligible for and had drawn $20 million of contingency funds would be
able to retain these funds, provided the state had spent the necessary
matching funds. In contrast, the state that qualified as needy for the same
number of months and was eligible for the same amount from the contingency
fund but overlapping 2 fiscal years would initially obtain $10 million in
each year, reflecting its 6 months of eligibility in each year, but then the
state would have to remit half of these funds after each year?s
reconciliation. This latter reduction is the result of prorating the state?s
grant by the number of months it was eligible for contingency funds, even
though the state?s initial claim for each year was already based on the
number of months of eligibility. As a result, the second state would be
allowed to retain a total of $5 million of federal funds in that fiscal
year, $5 million of federal funds in the next fiscal year- a total of $10
million even though its eligibility over these 2 years was the same as the
state receiving $20 million. In addition, the second state would have to
meet the contingency fund MOE in both years.

Unlike the Contingency Fund, the Loan Fund does not have triggers. Instead,
states that have not incurred penalties for improper use of TANF funds are
eligible for loans from the Loan Fund. Such loans are to have a maturity of
no more than 3 years at an interest rate comparable to the current average
market yield on outstanding marketable obligations of the U. S. Treasury
with comparable maturities. Some state officials told us that they are
eligible for better financing terms in the tax- exempt municipal

bond market. More important, officials in some states indicated that
borrowing specifically for social welfare programs in times of fiscal stress
would not receive popular support.

In summary, neither the Contingency Fund- as currently designed- nor the
Loan Fund is likely to be used by states in a fiscal crisis to obtain more
resources for their welfare programs. The Loan Fund is most likely the wrong
mechanism to provide assistance to states in a fiscal crisis. However, if
the Contingency Fund is reauthorized, Congress could also contemplate
improvements to enhance its usefulness in addressing

budgetary shortfalls in states? welfare programs that, at the same time,
could provide stronger incentives for states to save for a rainy day.

Concluding While addressing the needs of low- income families remains a
joint federalstate Observations

responsibility, TANF changed the nature of the fiscal relationship between
federal and state governments in this area. As we have reported in the past,
block grants prompt trade- offs between federal and state control

over program finances, accountability, and administration. The challenge is
to strike a balance between providing states with the reasonable flexibility
and discretion needed both to achieve efficiencies and to adapt federal
programs to differing local needs while at the same time retaining

accountability for achieving certain broad national goals. While
policymakers should use data on state spending and performance when
designing any grant program, the design of block grant programs ultimately
involves philosophical judgments about the proper roles and relationships
among

levels of government in our system. Within this context, Congress will
consider trade- offs in the fiscal design of the TANF program as it moves to
reauthorize the program. The design of the grant- its fiscal requirements,
allocation provisions, funding flexibility, and accountability provisions-
has an important bearing on the federalstate

fiscal balance. By shifting from an open- ended matching to a closedended
block grant program, TANF shifted greater fiscal responsibilities to the
states for allocating both federal and state resources across activities and
for bearing the fiscal risks associated with the impacts of an economic
downturn on program needs. The design of the TANF grant enabled the

states to reduce their spending to a moderate extent as caseloads declined,
with a floor established by the MOE requirement. The grant level was also
high enough to permit states to plan ahead and perhaps create reserves for
use should case levels increase in less robust economic periods. The higher
grant levels arguably should inspire states to create reserves for the
future in order to avoid fiscal stress during less robust economic times.
Reflecting on the states? fiscal activities in the past 5 years, our study
can help provide insights on key fiscal design questions for the
reauthorization

debate:

 How has the fiscal design of the program affected the balance of fiscal
responsibilities between federal and state governments in terms of
addressing the goals of the program; what effects might proposed changes
have on this balance?

 Does the program enable or promote the saving of reserves at both federal
and state levels to address a future ?rainy day??

Federal/ State Fiscal Balance In a previous study on the design of federal
grant programs, we noted that federal grants do not necessarily have the
effect of augmenting spending in aided areas on a dollar- for- dollar basis,
since a portion of the federal funds

are often used by states to replace some of their own funds. We observed
that several factors influenced the potential for fiscal substitution
including (1) the breadth of activities for which the grant can be used, (2)
the presence and design of MOE and/ or nonsupplant provisions, and (3) the
extent of states? prior investment on the aided activities. Under TANF, each
of these factors has, in fact, played an important role in defining the new
fiscal balance between federal and state governments. 27 Our study suggests
that states did reduce their support for basic welfare services in the face
of declining caseloads, with the federal government picking up a greater
share of the costs for this narrow category of activity.

Replacing state funds with TANF funds was a common budgetary practice in all
10 of the states, and entirely legal under the federal program. At the same
time, when looking at the broadest level of TANF- related social services,
it is apparent that most states have maintained or even increased their own
investment over time to address the overall needs of low- income

families. The design of the grant influenced these results. The increased
flexibility in TANF increased the potential for states to use federal funds
in areas of traditional and long- standing state support. However, that same
flexibility also enabled states to focus their efforts on TANF?s broader
goals of supporting work and self- sufficiency for low- income families. The
absence of a nonsupplant provision for TANF funds- similar to the ?new
spending test? on state MOE funds- permitted direct substitution of federal
for state funds. However, even with this substitution, the range of
activities eligible for TANF funding was broad enough to enable states to
identify enough state- funded activities to meet their MOE. Unlike a
nonsupplant provision, the MOE was not designed to stop reallocation of

state funds within TANF- eligible services. Instead, it was intended to
limit the extent to which states could use federal funds for fiscal relief.
Indeed, it appeared to play a strong role in limiting the degree to which
states could reduce their spending to fully reflect the sharp drop in cash
assistance caseloads. 27 Federal Grants: Design Improvements Could Help
Federal Resources Go Further (GAO/ AIMD- 97- 7, December 18, 1996).

If states used TANF funds to replace state funds counted for MOE, they had
to increase their own spending on other low- income programs to satisfy the
MOE requirement. Moreover, even though MOE permitted states to reduce their
spending by 20- 25 percent from the base year, it can be argued that the
requirement caused some states to spend more than they otherwise would have
spent in light of the more than 50 percent drop in caseloads. Given the
impact that MOE appears to have had on state spending to date, any changes
in the level and design of this requirement could have major consequences
for the federal- state fiscal balance for this program.

That different observers and actors will have different views of
supplantation is to be expected. Some would argue that allowing
supplantation of any kind increases the likelihood that states could weaken
their commitment to the core services (e. g., cash assistance) on which
lowincome families depend most and shift more of the financing for these
services to the federal government under the guise of program flexibility.
Potentially, supplantation permits states to convert federal TANF funds into
a kind of general revenue sharing program with very little incremental
impact on poverty. Although most states did supplant within the narrowest

category of spending in our analysis, given the stated purposes of PROWRA,
this is too narrow a perspective. Since TANF is intended to serve the
broader purpose of helping low income families achieve greater levels of
self sufficiency, the focus should be not only on how specific grant funds
were spent but rather on how states allocated federal and state funds more

broadly to address this outcome. Accordingly, focusing on the act of
supplantation alone in the narrowest category of spending provides a
misleading perspective of the fiscal impact of the program. A more complete
analysis would address the impact of the states? broader fiscal actions on
the goals of TANF.

These differing perspectives inform the debate over whether or not to add a
nonsupplant provision to the TANF program during reauthorization. Some argue
that creating a nonsupplant provision similar to that found in other federal
grant programs would help ensure that all federal funds supplement existing
state spending. While the program?s MOE provision places a floor under state
spending for the broad area of assistance to low- income families, a
nonsupplant provision would be aimed at preventing states from

using federal funds to replace state funds for specific activities. Some
have argued that the absence of a nonsupplant provision undermines the goal
of the ?new spending test? because this test applies to activities states
count toward MOE but not to activities on which federal TANF funds are
spent.

However, any consideration of a nonsupplant provision should recognize that
such a provision could have significant adverse effects as well. Such a
provision would penalize states that took the initiative to start programs
before the federal TANF program became available. The states that began

programs using state funds before the federal grant would be locked in to
maintaining their own spending commitments to these activities while other
states without such prior commitments could fund such activities entirely
with federal dollars. Thus, a nonsupplant provision not only would appear to
reward states for inaction but also could serve to induce states to defer
program innovation until federal funds become available. In this context it
is worth noting that in our study, supplanting tended to be

concentrated in states with prior investments in programs supporting the
working poor; states with relatively low levels of state effort did not have
the kind of investment that prompted supplanting in the first place. Given
this perspective, supplantation could be seen as rewarding these states for
their past fiscal initiative without further reducing benefits in low-
effort states.

Once accountability shifts to the broader goals supported by the block grant
program, many would argue that states should have the flexibility to shift
funds across activities within the broader umbrella of social services for
lower income families to best meet the broad goals of the program. Indeed,
states in our study took advantage of this flexibility by using TANF

funds to free up state funds to fund broader activities supporting lowincome
families. Just as the absence of a nonsupplant provision creates the
potential for the grant to become a general fiscal relief program, the
presence of such a provision might limit the intended breadth of the block

grant by locking states into pre- established funding priorities. From this
perspective, a broad- based maintenance- of- effort provision- such as the
TANF MOE requirement- calling for states to maintain spending across a wide
range of relevant programs might both limit substitution while providing
greater state discretion than a traditional nonsupplant requirement.

Whether or not Congress imposes a nonsupplant provision regarding TANF
funds, given the intent of the block grant, federal oversight should focus
on the broader fiscal balance, i. e., the way states use multiple federal
and state funding streams to support the overall goals of the program. As we
have said, focusing on more narrow categories of spending can be misleading
for broad purpose grants; 28 federal funds for any one specific program tend
to lose their fiscal identity and become integrated with funds from related
federal and state programs. Furthermore, in the case of TANF, the
performance outcomes call for states to bring to bear a wide range of
related federal, state, and local funding streams to make a difference for

low- income families. Accordingly, periodic reporting for the entire
federalstate fiscal effort is important. As we have said, we found that
states generally maintained or increased their support for the broadest
category

of assistance for low- income families. However, it will be important for
the Department of Health and Human Services periodically to gather
information on these broader measurements of the federal- state fiscal
balance so that Congress can monitor this important indicator in the future,

particularly as the economy changes and potential caseload demands for
programs and services increase. HHS should undertake this effort in concert
with states to ensure that the reports accurately reflect a broader fiscal
picture. Although such reporting may constitute an additional burden to the
states, following a select sample of states over a period of years and
relying on independent researchers to gather the data across consistent
categories every few years could mitigate such burdens. Further, states have
an interest in maintaining congressional confidence in the quality of this
information. The past history of block grants at the federal level suggests
that nationally consistent information is often necessary to maintain
support within Congress for continued funding

and flexibility. In the absence of uniform information, policymakers can
lose confidence in state implementation and are often pressed to intervene
in block grant programs based on examples that may not be representative of
states? overall track record.

Planning for Future Both the states and the federal government have a
significant interest in Contingencies preparing TANF to meet challenges in
times of fiscal distress. While the shift from the open- ended match of AFDC
to the fixed- dollar TANF block

28 For more information, see Grant Programs: Design Features Shape
Flexibility, Accountability, and Performance Information (GAO/ GGD- 98- 137,
June 22, 1998)

grant under PRWORA provided states with increased program flexibility, and
sometimes, fiscal relief, it also increased the states? exposure to fiscal
risks in the event of an economic downturn or increased program costs.
Clearly, this gives states an interest in planning for future needs.
Similarly, the fact that the federal government recognized its continued
federal role in assisting low- income families, coupled with the fact that
those states that fail to engage in contingency planning could, when faced
with fiscal pressures, turn to the federal government for help creates a
federal interest in the viability of TANF in times of economic stress.

In testimony this past April we offered for congressional consideration
several options that might improve the federal contingency mechanism as well
as strengthen states? incentives to ?save for a rainy day.? 29 Options to
Improve the Federal The Contingency Fund, as currently designed, has not
proven to be an

Contingency Mechanism inviting option to the states that have actually
experienced fiscal stress to

date. Should Congress decide to reauthorize the Contingency Fund,
consideration could be given to approaches that could both improve the
usefulness of the fund for hard pressed states as well as ensure that states
contribute their fair share to future welfare costs. Such approaches could
include (1) eliminating the more restrictive the Contingency Fund- MOE and
substituting the more flexible basic TANF- MOE and (2) eliminating the
Monthly Payment Limitation (MPL) on the amount of contingency funds to

which each state has access. These actions could help strengthen the role of
the Contingency Fund in state contingency budgeting.

Realigning the MOE and eliminating the MPL would make the Fund more
accessible and, therefore, more responsive. If states had better access to
federal contingency funds, they might be more likely to use the money when
needed. However, greater accessibility must be balanced by fiscal
responsibility. It is important to be mindful of this balance so as not to
make it too easy for states to access federal contingency funds because

they might be less likely to save for a rainy day on their own, which could
pose risks to the federal Treasury. The changes discussed above would still
require states to increase their own spending to pre- TANF levels (i. e.,
meet a 100- percent MOE) to gain access to the Contingency Fund- a higher
level

than they must maintain for the regular TANF program- as well as provide a
matching share for the additional federal funds. By broadening the fiscal
base that states can draw upon to meet this higher MOE, these changes 29
Welfare Reform: Challenges in Saving for a ?Rainy Day? (GAO- 01- 674T, April
26, 2001).

might not only make the fund more accessible in times of need but prompt
states to save their own funds in anticipation of accessing the federal
funds. Options to Increase States? There are other options that could
strengthen states? incentives to save. For Incentives to Save

example, Congress could (1) allow states to count rainy day funds towards a
portion of their MOE and (2) allow states to draw down their entire TANF
grants and save these funds in their own treasuries. There are pluses and
minuses to each- and the exact design and implementation will be important
in determining the actual impact of any change.

Allowing states to count rainy day funds towards part of their MOE would
give them a greater incentive to create such funds. However, ?maintenance

of effort? implies an actual expenditure and has played a critical role in
limiting supplantation. If states save their own funds instead of spending
them, they might be more likely to draw down all of their TANF dollars now
to replace the state dollars- thus resulting in no net increase in spending
in this area. This effect could be mitigated by limiting the amount of rainy
day

funds that could count towards a state?s MOE. In addition, as we suggested
in our April testimony, states could be required both to certify that state
rainy day funds were auditable and to establish criteria for the size of the
fund and the release of funds from it.

Some state officials have said that requiring states to keep unspent TANF
funds in the U. S. Treasury provides a disincentive to save these funds. As
noted earlier, they believe that these balances could be attractive targets

for rescission by Congress. If states could keep unspent TANF funds in state
accounts instead, these officials suggest, states would have more of an
incentive to save for contingencies. Such a provision would mean outlays
would be recognized at the time of the grant award rather than when the
money is spent for a program need. There are significant issues associated
with this proposal. First, regardless of where these federal funds are
?stored,? states are accountable for the use of these funds in accordance
with the law. Congress would still need consistent, reliable, and auditable
information on the funds. Second, such a

proposal would affect the measured federal surplus. Since outlays would be
recognized when the states received their grants, more outlays would be
recorded sooner. If states draw down all unspent balances in a single year,
the outlays recorded for the TANF program would shift and the budget surplus
drop by that amount in the near term.

In addition, the federal government would incur interest costs while states
could earn interest on grant money drawn down from the U. S. Treasury but
not yet spent by the states. This would violate CMIA, which prohibits states
from drawing down funds until those funds are needed and so helps ensure
that neither the states nor the federal government incur unnecessary
interest costs (or forgoes interest income) in the course of federal grant
disbursement. CMIA also promotes transparency since states? unspent

balances remain visible in the federal Treasury rather than spread through
the various state treasuries- although given the measurement problems
discussed earlier, this may overstate the quality of current information.

To provide TANF an exemption from a governmentwide grant policy that settled
years of intergovernmental conflicts between federal and state
administrators would require justification. The permanent nature of the
appropriation to each state and the significant devolution of
responsibilities and fiscal risk to states may argue for an exemption from
CMIA. If Congress wished to provide this exemption without abandoning fiscal
neutrality between the federal and state governments, it could require
states to reimburse the U. S. Treasury for any interest the states earn on
TANF drawdowns. Under such a provision the states would hold the funds, but
the earnings on the funds would remain as if the funds were in the U. S.
Treasury. Similarly, requiring better reporting on states? expenditures and
reserves could mitigate the damage to transparency.

Recommendations for Given the breadth of the goals for the TANF program,
understanding the Executive Action

overall federal- state fiscal balance covering the broad range of
TANFrelated social services is important for national oversight of the
program. However, little information is currently gathered systematically on

expenditures on the broad array of social services we surveyed for this
report. In order to inform decisionmakers on the status of the federal-
state fiscal balance, the Secretary of Health and Human Services should
consult with the states to explore cost- effective ways to periodically
gather data revealing the federal- state fiscal balance relevant to
achieving the broader

programmatic goals of the TANF program. The categories used in this report
could constitute a basis for these periodic reports. We also reiterate our
prior recommendation that the Secretary work with the states to provide for
more transparent reporting by the states of their plans for the unspent TANF
balances. Such plans can enhance congressional oversight and provide the
states with an opportunity to better demonstrate their plans and needs for
federal funding.

Agency Comments We received comments from HHS, which are reprinted in
appendix XIII. In addition, each of appendixes II through XI was reviewed by
officials in the relevant states and their comments were incorporated as
appropriate. We also provided copies of the draft report to the National
Governor?s Association (NGA) and the National Conference of State
Legislatures (NCSL). NCSL provided official comments. NGA did not provide
official comments, but senior analysts provided us with substantive informal
comments on the draft. We incorporated these comments, NCSL?s comments, and
technical comments from HHS as appropriate. NGA staff and NCSL both
commented that the report was balanced. NCSL noted that the report ?clearly
puts TANF spending in context of how states make policy and budget decisions
and will enhance the upcoming debate over

welfare reform reauthorization.? At the same time both NGA staff and NCSL
expressed concern that some readers could use the data in this report to
criticize states? decisions to shift state funds from Category 1 to broader
TANF- related purposes. Our report acknowledges that these shifts were
prompted by a number of factors including the declining caseloads

and the breadth of TANF?s goals. These funding shifts are a result of policy
choices made by the states- we do not impose a value judgement on the
appropriateness of these choices. Moreover, we point out that these shifts
are not only legal, but are encouraged by TANF law. We continue to believe

analysis across the full range of categories is important to providing a
more complete picture of state choices.

With regard to our recommendation to explore ways periodically to gather
data revealing the federal- state fiscal balance covering the broad range of
TANF- related services, HHS agreed that ?comprehensive, timely data from

the states will be essential during TANF reauthorization,? but suggested
that unspecified legislative restrictions and cost constraints could hamper
its ability to collect this essential information. We believe that all data
collection should be cost- effective. We note that the Department agrees

that this information is essential for TANF reauthorization; at this stage
we suggest that the necessary data might be obtained through incremental
adjustments to and leveraging of the current reporting requirements and
existing research. In addition, as part of the reauthorization debate all
data collection requirements might be examined with an eye to focusing
reporting requirements on the data that Congress, the states, and the

Department consider most useful in assessing progress towards meeting the
goals of welfare reform. As part of that re- examination, the Department
could note any legislative changes necessary to facilitate this process.

As agreed with your offices, we are sending copies of this report to the
Secretary of Health and Human Services, appropriate congressional
committees, and other interested parties. We will also make copies available
to others upon request.

If you or your staff have any questions about this report, please contact me
on (202) 512- 9573 or Thomas M. James on (202) 512- 2996. Individuals making
key contributions to this report included Bill J. Keller, Patricia L.

Elston, Jacqueline M. Nowicki, Raymond G. Hendren, and Marcus G. Melton.

Paul L. Posner Managing Director Federal Budget Issues and Intergovernmental
Relations

Appendi Appendi xes x I

Scope and Methodology Scope To meet our objectives, we studied TANF- related
budget decisions and collected data on social service expenditures in 10
states- California, Colorado, Connecticut, Louisiana, Maryland, Michigan,
New York, Oregon, Texas, and Wisconsin. These 10 states- the same states we
studied in 1998- were chosen because they represent a diverse array of
socioeconomic characteristics, geographic locations, experiences with state
welfare initiatives, and state fiscal and budget issues. 1 Although
differences in the way states present their social services budgets and the
way they finance their programs make any generalizations difficult, our case
study analysis of these 10 states, with their diverse approaches to
financing welfare reforms, provides opportunities to highlight critical
financing issues states face implementing their welfare reforms. Methodology
To provide background information and context to our discussion on how

states are using federal and state funds to finance programs to help the
poor, we reviewed and summarized state welfare plans, program documents, and
other supporting documentation. We discussed state welfare programs with
state program and budget officials, legislative analysts, and local welfare
advocacy groups. We reported on features of state welfare programs in place
by state fiscal year 2001.

1 Welfare Reform: Early Fiscal Effects of the TANF Block Grant (GAO/ AIMD-
98- 137, August 18, 1998).

Supplantation in the State A discussion of whether a state uses federal TANF
funds to supplant state

Budget funds depends in large part on the definition of supplantation. Much
of the

controversy over whether states have used federal funds to replace their own
spending, and the extent to which they have done so, hinges on this
definition of supplantation. In this analysis we used a rather limiting
definition of supplantation, one used by the former Chair of the Human
Resources Subcommittee, Committee on Ways & Means, House of Representatives
in a letter to Governors dated March 15, 2000, which defines supplantation
as the act of ?replacing state dollars with TANF dollars on activities that
are legal uses of TANF funding.? A limiting definition is justified because
all funds are interchangeable. The effect of any state decision to use
federal funds to supplant its own funds is first to enlarge the pool of
state general fund resources available to finance any

state program, how the state then establishes priorities for funding new and
existing programs- essentially, how it will use its freed- up funds- must be
viewed apart from its decision to supplant. 2 In order to gauge the extent
to which a state used TANF funds to free up

state funds, we analyzed state budgets and supporting budget documentation.
We identified instances where the state changed how it financed certain
programs by replacing state funds from its currentservices baseline 3 budget
with federal TANF funds through state fiscal year 2001. We also gathered
information on the general fiscal environment that existed in each state as
these budget choices were made, other spending priorities that competed for
state funds, any fiscal crises in the states, and

interactions with other federal programs. While it is difficult to tell with
certainty how a state uses its freed- up funds, the decision to supplant is
often made through the state?s appropriation or 2 Most academic studies that
seek to assess the extent to which states use federal funds to supplant
their own define supplantation as a decrease in state spending from what the
state

would have spent if more federal funds had not been made available. These
studies gather historical data on state spending, control for certain
demographic variables such as caseload changes, identify trends in state
spending, and then extrapolate those trends. The studies compare the trends
with actual expenditure data and explain any variance from the trends as
supplantation. 3 A current- services baseline budget assumes the
continuation of current policies and reflects anticipated costs of ongoing
programs and activities without policy changes. It generally includes
allowances for inflation and changes in caseload. Proposed policy changes
that would affect the costs of programs are compared to the current-
services

baseline to estimate the budgetary impact.

budget process, which is often accompanied by analyses of different policy
options for the use of any freed- up funds and can influence policymakers?
decisions to approve the switch in financing. In reviewing these analyses it
is possible to understand the choices facing decisionmakers as they decide
among the different ways to use federal and state funds, to finance

programs that benefit low- income families and other needy individuals. To
describe the motivation for states to use federal TANF funds to free up
state funds, we interviewed key staff from the state welfare agencies, state
budget offices, legislative fiscal analysts offices, as well as some state
legislators with oversight responsibilities for welfare reform.

Spending Trends From 1995- While budgets are a good blueprint to measure
policy choices, looking only

2000 at a specific change authorized in a budget may be misleading. For
example, if a state supplanted federal funds for state funds, then shifted
its

own funds into other programs that serve the working poor, the net impact of
these shifts would be the same as if it used federal funds directly to fund
these other programs. Accordingly, the impact of the funding shifts must be
viewed in a broader context consistent with new objectives and goals of
TANF; which are to

 provide assistance to needy families so that children may be cared for in
their homes or in the homes of relatives,

 end the dependence of needy parents on government benefits by promoting
job preparation, work, and marriage,

 prevent and reduce the incidence of out- of- wedlock pregnancies, and

 encourage the formation and maintenance of two- parent families. To
provide this broader context for considering the effects of funding shifts,
we reviewed state expenditures of both federal and state funds on a wide
array of social service programs that meet these objectives. In this
analysis, a federal expenditure is not defined by the level of the TANF
grant

allocated to each state but rather by how much of the grant the state chose
to spend. Essentially, the states control both the state and federal share
and thus control the ?fiscal balance.? Under AFDC, the fiscal partnership
between the federal government and the states was defined by the federal
match rate each state received to help finance its welfare programs. Under
TANF, the block grant is fixed- states decide how much of the federal

grant to spend and how much to save.

This ?fiscal balance? is key to understanding how the partnership between
the federal government and the state in financing welfare services may have
changed. In order to measure post- PRWORA changes in the mix of federal and
state funding across TANF- related social services, we asked the 10 states
to provide expenditure data on programs that provide social

services to the needy. To facilitate gathering this information, we adapted
a fiscal survey developed by the Nelson A. Rockefeller Institute of
Government. 4 Specifically, we asked for state and federal expenditures on

a variety of programs that targeted assistance to low- income families and
individuals in state fiscal years 1994- 1995 and 1999- 2000, and we verified
the data they reported to us against their own accounting records. To
improve the alignment of our survey data with TANF?s goals, we excluded
social services expenditures targeted to the elderly and those services that

were not transitional in nature. We used the survey as a tool to encourage
states to view these programs broadly, and not to limit their survey
responses just to those programs that received TANF funding. The survey
allowed us take a comprehensive look

at state social service program budgets. Generally, we asked the State
Budget offices to help us coordinate the survey responses. In their
oversight role of the entire state budget, these offices had a broader view
of state funding priorities than the State Social Services departments, or
any one specific department, which focus primarily on their own budgets. We
asked state budget and program officials to identify programs that met our
survey criteria, whether or not these programs were funded with TANF or

MOE dollars (see appendix XII for our survey criteria instructions), and to
provide us with expenditures on these programs for their state fiscal years
1995 and 2000. Our focus on the state fiscal year as opposed to the federal

fiscal year allowed states to provide us with data directly from their
accounting records without having to make adjustments to the federal fiscal
year. 5 Since our objective was to examine changes in the way states have
used federal and state funds to finance programs that help the poor, this
focus on the state fiscal year is adequate. Moreover, since state

decisions to supplant federal funds for their own were made in the context 4
?Changes in State Spending on Social Services Since Implementation of
Welfare Reform: A Preliminary Report.? Ellwood and Boyd, February 2000. 5
Seven of the 10 states in our review begin their fiscal year on July 1 and
end on June 30. The other three states are all different; in New York it
begins on April 1 and ends on March 31, in Texas it begins on September 1
and ends on August 31, and in Michigan it begins on October 1 and ends on
September 30.

of their budget deliberations, the survey?s focus on the state fiscal year
complemented our analysis of state budget decisions.

Once the state program and budget officials identified the programs to
include in the survey, we verified through program documentation and
discussions with these state officials that the program descriptions,
targeted beneficiaries, and program goals met the survey criteria. We worked
closely with state officials to complete the survey instrument (see appendix
XII) with data from state accounting records. 6 We verified with officials
from the State Auditor?s offices that state expenditure system

controls had been tested and verified. As noted previously, a federal
expenditure, for the purposes of this analysis, is not defined by the
federal allocation to the state, rather by how much the state chooses- or in
some cases is required- to spend on the aided activity. For this reason, we
did

not consider a number of 100- percent federally funded programs that do not
flow through the state budget. For example, the Food Stamps program is
administrated by the state and the shared administrative costs are included
in the survey but the value of the food stamp coupons disbursed in the
fiscal year, borne 100 percent by the federal government, is not.

In our analysis we aggregated state expenditure data by program category,
allowing us to group programs into four categories- Basic Welfare Services
(Category 1), Support for the Working Poor (Category 2), Health Care
(Category 3), and Other TANF- related Social Services (Category 4).

The analysis is cumulative; that is, all expenditures in Category 1 are
included in Category 2, and so on. See figure 1 for a description of the
types of programs that are included in our spending categories. We
aggregated program spending in order to make comparisons among states in how
they design and fund their strategies to address TANF?s goals. State

budget structures are very different; some states place all their needs-
based employment and training programs within their social services budgets;
other states place these programs in their economic development agencies.

These differences make comparisons of state budgets and expenditures
difficult. In asking states to report spending on individual programs,
regardless of which state agency oversaw these programs, and then
aggregating the programs into the same categories for each state, we were
able to compare state expenditure trends across all states. Given the

6 In some cases state accounting records were not closed out when we
completed our survey instruments. In those cases, working with state program
and budget analysts, we estimated state fiscal year 2000 expenditures based
on actuals to date or historical spending patterns.

differing views of supplantation, our categories were designed to provide
decisionmakers with information on shifts in spending on the widely
different- and increasingly broad- strategies employed by the states to
achieve the goals of the TANF program. Our first category includes programs
that are linked closely with the goals

and objectives of the former welfare programs, namely cash assistance
payments and some job training programs. Our second category adds programs
that many view as supplements to those basic welfare services. The additions
include programs that help families make the transition from welfare to work
or avoid welfare altogether. For example, child care subsidies and rental
assistance payments can help parents remain working even if they are working
in low- wage jobs; family formation and pregnancy prevention programs are
also included in this category as they address specific TANF goals. Our
third category adds health care costs to the types

of programs already discussed. The inability of many low- income families to
gain access to health insurance once they leave the welfare rolls has been
recognized as a major barrier hindering their ability to attain
selfsufficiency. PRWORA delinked the AFDC program from the Medicaid program,
but in an effort to safeguard access to health insurance for eligible low-
income families, the act required states to implement a separate Medicaid
eligibility category which ensured that eligible lowincome families could
qualify for Medicaid even if they were no longer receiving cash assistance.
Improving access to health insurance, especially for low- income
individuals, has been in the forefront of this country?s

domestic policy debate for the last few years. During this time frame,
states were encouraged to participate in a new federal health insurance
program for low- income children- the States? Children?s Health Insurance
Program (SCHIP)- who were not eligible for Medicaid. Moreover, some states
expanded access to health insurance to other low- income individuals not
eligible for Medicaid. Adding states? health care expenditures to our
analysis recognizes the costs of these programs and states? investments in
strategies to help families make the transition from welfare to work.

Finally, our last category recognizes the full breadth of programs that some
states use to develop strategies to achieve TANF?s goals. These program
include child welfare programs, substance abuse programs, mental health
programs, and programs that help the developmentally disabled attain a

level of self- sufficiency. While many of these state programs do not have
income standards to determine eligibility, many states are able to certify
that some of the participants in these programs meet the eligibility
requirements set forth in their TANF plans and can claim TANF funds for
these expenditures. As such, states are able to supplement, or supplant

from, these existing programs- or create new programs- as part of their
strategy to achieve their TANF goals.

State Contingency Plans To describe the measures states are taking to ?save
for a rainy day? and to analyze the impact these actions have on TANF
reserve balances at the U. S. Treasury, we analyzed the ten states? budgets
as well as the quarterly financial reports they are required to file with
HHS on the status of TANF funds they have been awarded to date. In addition,
we interviewed program and budget officials to determine the states? plans
for their unspent federal TANF funds. In particular, we asked states that
have set aside some of their federal funds for contingencies for information
on how the size of these reserved funds were determined. We also interviewed

officials in all 10 states to determine how they might mitigate the impact
of an economic downturn on their program budgets if changes in the economy
caused welfare rolls to swell and costs to rise. To obtain information on
the levels of unspent TANF funds in all 50 states we

reviewed financial data reported by the states to HHS. We requested comments
on a draft of this report from HHS, NGA and NCSL. These comments are
discussed in the letter and HHS? comments are reprinted in appendix XIII. In
addition, we provided drafts of each state appendix to the appropriate state
officials for technical review and have incorporated their comments where
appropriate.

Appendi x II

California Background

State fiscal year: July 1- June 30 Budget cycle: Annual TANF grant: a $3.
734 billion

75% MOE: $2. 727 billion Cash assistance caseload: b 926,000 families in
1995

489,000 families in 2000 47 percent decline a California received an
additional $45 million in federal TANF funds in 2000 and $36 million in 2001
as bonuses for successful outcomes. b Caseloads as reported to HHS for
January 1995 and June 2000 and rounded to the nearest thousand. Caseload
data for 2000 do not include 63,000 families receiving assistance through a

separate state program.

California implemented its welfare reform program called California Work
Opportunity and Responsibility to Kids (CalWORKS) in January 1998. The
program provides cash aid and services to eligible needy California
families, with the goal of ensuring that individuals who work are better off
financially than if they do not work.

To be eligible for cash aid, generally adults in families must engage in
workrelated activities or lose their eligibility. Most are assigned
immediately to job search activities, and if unsuccessful after 4 weeks, may
be assigned to other work- related activities such as job training,
education, counseling, or substance abuse treatment. Cash assistance
includes the monthly assistance payment, and diversion payments made in lieu
of monthly

assistance, where appropriate and in accord with county policies. The
maximum monthly assistance payment was increased in 2000 and ranges from
$626 for a family of three in high- cost counties to $596 in low- cost
counties. In addition to cash assistance, single parents may receive the
first $50 of the current month?s child support collected on the families?
behalf. 1 Families are eligible for cash assistance if their adjusted
incomes are less than the payment level, and they remain eligible for up to
5 years as long as their adjusted incomes do not exceed the payment level.
After that time, the children may continue to receive assistance.

1 The state and federal governments may retain collections as reimbursement
for cash assistance payments made to these families. California has chosen
instead to give its full share to CalWORKS families.

While they are participating in program activities, CalWORKS families may
receive help with child care, transportation, and related expenses.
Depending on the policy of the county in which they reside, families who
leave assistance for employment may continue to receive casemanagement and
supportive services for 1 year. Transitional medical assistance is available
for 1 year and to the extent funding is available, subsidized child care can
be authorized for up to 2 years or longer if the

families? incomes remain below 75 percent of the state median income. 2 Two-
parent families and recent legal immigrants not eligible under TANF are in
separate programs that are funded entirely with state funds and used to
count towards the state?s MOE.

California administers its welfare program through its 58 counties. The
counties pay cash assistance directly to welfare families and have
considerable policy and programmatic discretion in managing cases and
providing services. The state allocates its federal TANF and state MOE

funds to the counties in the form of block grants and, in turn, requires
that counties invest a portion of their own funds toward the MOE. About $141
million in county funds was counted toward the MOE in 1999. Counties have
received additional federal TANF funds from the states as performance
incentive payments for successful program outcomes and are

permitted to use these funds for any TANF purpose. Supplantation in the

In California, since passage of TANF, it appears that the Governor and State
Budget

legislature have enacted budgets that fully support the state MOE and have
managed to balance competing priorities with no evidence of supplantation.
For the most part, we found little evidence that the state had engaged in
major shifts in the methods of financing its welfare programs. The state has
devolved the funding for program development and implementation to the
counties, leaving few opportunities for the state to shift federal funds
into budget accounts that had previously been funded with state funds.
Further, when asked about possible supplantation by the counties, state
officials responded that their monitoring has not yielded any indication
that counties have used federal TANF or state MOE funds to free up their own
county funds. State officials also considered county- level supplantation
unlikely, because few counties operate programs with their 2 According to
state officials families, who actually receive child care subsidies are at
much lower income levels, usually between 35 and 50 percent of SMI.

own funds that could be considered allowable candidates for TANF funds,
leaving few opportunities for supplantation.

At several points in their budget deliberations, California?s Governor and
legislature considered the opportunities afforded by lower caseloads and
fixed federal funding to free up state general funds. For example, the state
recently learned that it met 1997 work participation rates, which qualified
the state for a 75 percent, rather than 80 percent, MOE. As a result, the
state can retroactively claim some of its unspent TANF balances in lieu of
the state funds that it spent in 1997. In his budget proposal for state
fiscal year 2001- 2002, the Governor recommended using some of the state?s

unspent federal TANF balances to reimburse the state for $154 million it
spent in 1997 to meet its $2. 9 billion MOE requirement. In another example,
California assumed financial responsibility after 1995 for providing
nonemergency health care, primarily preventive in nature, to recent legal
immigrants who are not eligible for TANF. Legislative analysts have
suggested that the state count its investment in this program as MOE, reduce
the state?s investment in CalWORKS by an equivalent amount, and use TANF
funds to make up the difference. CalWORKS would remain intact and fully
funded. One analyst explained that the proposal would help reduce reserves
of unspent TANF balances, yield state savings, and, at the same time,
effectively restore federal funds that were previously available for this
service.

Spending Trends From 1995 to 2000

Figure 4: California?s TANF- Related Expenditures for State Fiscal Years
1995 and 2000

1995 dollars in millions 25, 000

20, 000 15, 000 10, 000

5, 000 0

95 00 95 00 95 00 95 00 Basic welfare Support for Support for TANF- related

services working poor working poor +

social services health care

Federal State

To assess how state budget decisions might have affected spending, we
collected data on California?s use of state and federal funds. The results
are grouped into four categories, as shown in figure 4 and described further
in appendix I. Each category is identified on the graph by two bars- one for
spending that occurred in state fiscal year 1995 and one for spending that
occurred in state fiscal year 2000. From left to right, the categories are

cumulative. Expenditures for basic welfare services, such as cash assistance
and job training, are included in expenditures for support of the working
poor. These expenditures in turn are included in expenditures for

support of the working poor plus health care. Finally, all expenditures are
included in social services expenditures. It is clear from figure 4 that
total spending in California has declined in real terms for the first three
categories, but has increased in the last category which includes all social
services. The majority of California?s MOE requirement is met through its
cash assistance payments. As caseloads have fallen in California, total
spending on basic welfare services declined from $6.3 billion to $3. 6
billion in nominal dollars from state fiscal years

1995 through 2000. (See table 6.) In real terms, this represents a drop of
47 percent. Although total cash assistance payments have declined, the state
still must meet a fixed MOE requirement. Given the state?s strategy in

meeting its MOE, the state?s contributions to these programs has increased
from 48 to 51 percent of the total. As a result, California now finances a
greater share of the total program costs for basic welfare services than it
did prior to welfare reform.

Table 6: California?s TANF- Related Expenditures

Dollars in millions

Real Nominal (expressed in 1995 dollars) 1995 2000 1995 2000

Basic welfare services Total funds $6, 252 $3, 618 $6, 252 $3, 301 State
funds $2, 998 $1, 829 $2, 998 $1, 669 Support for the working poor Total
funds $9, 962 $9, 283 $9, 962 $8, 470 State funds $6, 037 $5, 843 $6, 037
$5, 331 Source: GAO survey and analysis of state expenditure data.

Combined spending on support for the working poor which includes basic
welfare services as well as other benefits, employment services, and child
care for both welfare recipients and the working poor, has declined slightly
from $10 billion to $9. 3 billion in nominal dollars from state fiscal years
1995 to 2000; this represents a real decline of about 15 percent. Although
less is being spent, the state share of program spending rose from 61

percent in 1995 to 63 percent in 2000.

When health care expenditures are added to the totals described above,
combined spending increased slightly in nominal dollars, from $17. 1 billion
to $17.6 billion from state fiscal years 1995 through 2000, but not enough
to keep up with inflation (see table 7). Instead, in real terms, total
spending declined by 6 percent. During this time, the state?s share has
remained constant at 56 percent of total expenditures, with federal funds
making up

the remaining 44 percent.

Table 7: California?s TANF- Related Expenditures

Dollars in millions

Real Nominal (expressed in 1995 dollars) 1995 2000 1995 2000

Support for the working poor plus health care Total funds $17, 146 $17,586
$17, 146 $16, 046 State funds $9, 579 $9, 756 $9, 579 $8, 902 TANF- related
social services Total funds $20, 634 $23,809 $20, 634 $21, 724 State funds
$11, 525 $13,289 $11, 525 $12, 125 Source: GAO survey and analysis of state
expenditure data.

However, when expenditures for child welfare and other social services, such
as substance abuse and mental health treatment are included, total
expenditures increased in both nominal and real terms from state fiscal
years 1995 through 2000. Countering the decline in spending on basic welfare
services, total child welfare expenditures increased by two- thirds and
other social services expenditures doubled. In this area, the state
maintained its investment at 56 percent of total expenditures from 1995
through 2000.

Additional federal safety- net programs provide food assistance and
supplemental security income for needy families and children in California.
For example, the federal government provided about $2. 5 billion in food
stamp benefits to low- income individuals in California in 1995 and $1. 8
billion in 2000. In addition, from 1995 through 2000, federal supplemental
security income payments increased from $3.4 billion to $4 billion, with
about 65 percent of these benefits going to individuals under 65 years of
age.

Rainy Day Funds California budgets nearly all of its annual block grant,
reserving very little for future contingencies. It had no TANF reserve set
aside until state fiscal year 2000- 01 when it appropriated about $50
million in federal TANF funds to a reserve account. 3 Even though the state
does not have significant reserves for contingency purposes, the state does
have substantial reserves of unspent funds. According to the financial
reports filed by the state with HHS as of September 30, 2000, the state
reports $1.6 billion in total unspent TANF funds, almost half of its TANF
grant for any given year. (See table 8.)

Table 8: Unspent TANF Funds in California as of September 30, 2000

Dollars in millions

Unspent funds as a Unobligated

Unliquidated obligations Tot al unspent

percent of annual TANF Funds of TANF funds funds

TANF grant

$3 $1, 637 a $1, 640 44% a Subsequent amended reports submitted by
California reduced unliquidated obligations for this period to $612 million.
Source: HHS? Administration for Children and Families.

Most of the state?s unliquidated funds have been obligated to counties in
their annual single allocation. In its second year under TANF, California
kept tight control of funds, issuing separate allocations to counties for
each

expenditure category. Because this process limited counties? flexibility and
made it difficult for the counties to reform their welfare programs in a
timely fashion, the state changed its strategy in its second year under
TANF. Now the state allocates the majority of program funds to counties in
broadpurpose single allocations that are viewed as block grants. These
allocations represent the maximum amount of reimbursement that the counties
can expect from the state for their basic CalWORKS expenditures. When the
state accounts for its TANF block grants on its quarterly financial reports
to HHS, the state reports its allocations to counties as obligations in
accordance with federal regulations. The allocations are typically made in
July after the state has enacted its budget. Therefore, California?s TANF

obligations typically spike in the last quarter of the federal fiscal year
as a result of its allocation process.

3 TANF funds held in reserve must be left in the U. S. Treasury until
expended.

Counties operate on a cash accounting basis, submitting claims for
reimbursement of expenditures that they have made for CalWORKS from their
block grant allocations. When the state receives county claims for
reimbursement, the state reports these as expenditures on federal forms.
Unliquidated obligations represent the difference between what the state has
allocated to the counties and what the counties have spent on actual cases.
Because spending on the TANF programs has declined proportionally with the
decline in caseloads, the state continues to report high levels of
unliquidated obligations. However, state officials maintain that as the
counties make progress implementing welfare reforms they are liquidating
their TANF obligations more quickly so that each year the level of reported
unliquidated obligations has decreased. Based on current

expenditure projections, these officials do not expect high levels of
unliquidated obligations to continue. Recently, the state amended its
statute so that these funds can be reallocated from counties where they may
not be needed to counties that are running short. Because the state has the
discretion to reobligate these unspent funds from counties with ample
reserves to counties with greater unmet needs, these unspent funds-
currently reported as ?unliquidated

obligations?- are a potential source of funds that state can use as long as
some counties continue to carry- forward large unspent balances.

California also has another significant source of unspent TANF funds that
can be used by the counties to finance increased costs in the event of
economic contingencies. California has transferred more than $1 billion of
the state?s TANF dollars to its counties in the form of incentive payments-
yet most of these funds remain unspent and are on deposit in county bank
accounts. This state program pays counties incentive payments that are
designed, in part, to motivate the counties to help the state meet certain

TANF performance goals. Specifically, the state makes awards to its counties
based on the number of welfare cases that have left the rolls for employment
and the number of cases the counties have diverted from the

rolls and are employed. To date the counties have received awards totaling
about $1. 1 billion. These funds are federal TANF funds that have been drawn
from the U. S. Treasury and deposited into county bank accounts. According
to state law, counties must spend these funds on TANF- related purposes. The
state has also issued guidance to the counties that any interest earned
while these funds are on deposit must be reinvested into the TANF program.
Apart from the aforementioned guidance, state statute defers to the counties
regarding how best to use the funds. The program

was authorized in 1997 as part of the legislation that established CalWORKS
and the first awards were paid in 1999. While state officials have said that
these incentive payments were not intended for counties to

use for contingency budgeting, counties are not prohibited from using these
funds to cover unexpected costs if needed. In fact, some counties regard the
incentives as their reserve funds, according to one state official.

In February 2001, state officials reported that counties have spent only $46
million- about 4 percent- of these funds. This means counties in California
have about $1 billion in unspent TANF funds on account that are

not included in the ?unobligated balances? totals reported to HHS. HHS
believes that California is in violation of CMIA, which seeks to limit the
interest costs to both federal and state governments by coordinating the
timing of a states? draw downs from the Treasury with the timing of the
states? expenditure needs. This coordination ensures that the timing of a
state?s draws do not favor either party and neither partner incurs
unnecessary interest costs. California maintains that these payments are
valid expenditures to service providers- their counties- under a

performance- based agreement to administer welfare services.

Appendi x II I Colorado Background

State fiscal year: July 1- June 30 Budget cycle: Annual TANF grant: a $136
million 75% MOE: $83 million b

Cash assistance caseload: c 39, 000 families in 1995 11, 000 families in
2000 72 percent decline a Colorado receives between $3 million and $10
million in annual supplements because the state?s preTANF population growth
exceeded 10 percent. b HHS found that in 1999 not enough of Colorado?s two-
parent caseload was engaged in work activities

to meet the higher participation rates required for that group and approved
the state?s Corrective Action Plan on April 27, 2001. Because the state did
not meet the requirement, it will be held to an 80percent MOE of about $88
million. c Caseloads as reported to HHS for January 1995 and June 2000 and
rounded to the nearest thousand.

Colorado implemented its welfare reform program called Colorado Works in
July 1997 with the goal of promoting personal responsibility and enabling
needy families to become self- sufficient. To achieve this goal, the state
has given each county social services agency the flexibility to select an
approach to program design and service delivery that the county agency
considers likely to yield success. Under the program, cash assistance
caseloads have declined steadily through mid- 2000.

Colorado contracts with its 63 counties for the operation of Colorado Works
through a memorandum of understanding and allocates most of its federal TANF
and state MOE funds- about $140 million in state fiscal year 2000- in the
form of block grants. The counties are required to invest a

portion of their own funds to help make up the required state MOE,
equivalent to 20 percent of the county?s 1995 expenditures on these
programs.

To receive cash assistance, applicants must sign individual responsibility
contracts or treatment plans immediately and participate in work related
activities within 24 months, or earlier in accordance with county
requirements. Counties determine the type of work activities that are
available and whether other activities, such as mental health and substance
abuse treatment, can be considered work- related. The state has set the
minimum standard benefit amount at $357 per month for a family of three,
virtually unchanged since before TANF, but individual counties may set

higher levels. Needy families with children whose adjusted incomes do not
exceed the state?s cash assistance levels are eligible for the program.
Those who participate may also receive other benefits, such as
transportation, clothing, or education payments. Families may receive cash
assistance for no more than 5 years, in accordance with TANF law. In lieu of
ongoing cash assistance, eligible families and, at county option,

other needy families may receive single, lump- sum diversion payments.
Recent immigrants not eligible for TANF are served in a separately funded
program. Families who succeed in finding jobs may achieve earnings that make
them ineligible for Colorado Works. For example, a single parent with two
children-- earning the minimum wage for a 40 hour work week would no longer
qualify for cash assistance. However, the family would qualify for 12 months
of additional medical assistance. When they exhaust those

benefits, the families may continue to receive health care for their
children, because the state has streamlined the application and enrollment
process and replaced the Children?s Health Insurance Program (CHIP) monthly
premiums with nominal enrollment fees to cover families with incomes

between 151 and 185 percent of poverty. Also families with incomes that do
not exceed $75,000 may be eligible for services and benefits such as
emergency aid and family preservation services as described in individual
county plans. Supplantation in the

We found little evidence in Colorado of the state?s use of federal funds to
State Budget replace state funds. State officials contend that supplantation
is less likely in Colorado than might be the case in other states for two
reasons. First,

counties wield a great deal of political power in Colorado. Since most of
the TANF funds the state receives are immediately allocated to the counties,
the state retains very little budgetary control over TANF funds. Most state
and county officials we spoke with agreed that politically it would be very
difficult if the state were to use TANF funds to replace state

funding in other programs, because it would mean that there would be fewer
TANF funds to appropriate to the counties. Second, the counties are less
likely to supplant funds because, according to a state official, most have
not had significant budgetary resources allocated to TANF- related program
areas from which they could supplant. In addition, in its state welfare
reform legislation, the state imposed an explicit prohibition against
counties using their block grants to supplant county general funds.

Although most of the state?s block grant is passed through to the counties,
the state has kept two modest reserves of TANF funds for use at the state?s
discretion. (See ?Rainy Day Funds? section of this appendix for more
information.) In state fiscal year 2000, the state reassessed the way these
two reserve funds were financed as they had accumulated significant balances
and had never been drawn upon. The counties were amassing significant
reserves as well since actual expenditure rates were lower than their annual
allocations. In reassessing the financing mechanism for the state reserves,
the state had about $20 million that it could use for other purposes. Of
this total amount, the state used about $7 million to refinance a number of
state- funded programs which freed- up a like amount of state funds. These
?freed- up? state funds were used to finance a new reading initiative for
elementary school children.

Spending Trends From 1995 to 2000

Figure 5: Colorado?s TANF- Related Expenditures for State Fiscal Years 1995
and 2000 1995 dollars in millions 1, 800

1, 600 1, 400 1, 200 1, 000

800 600 400 200

0 95 00

95 00 95 00 95 00 Basic welfare Support for Support for

TANF- related services

working poor working poor +

social services health care

Federal State

To assess how state budget decisions might have affected spending we
collected data on Colorado?s use of state and federal funds. The results are
grouped into four categories, as shown in figure 5 and described further in
appendix I. Each category is identified on the graph by two bars-- one for
spending that occurred in state fiscal year 1995 and one for spending that
occurred in state fiscal year 2000. From left to right, the categories are

cumulative. Expenditures for basic welfare services such as cash assistance
and job training are included in expenditures for support of the working
poor. These expenditures in turn are included in expenditures for

support of the working poor plus health care. Finally, all expenditures are
included in social services expenditures. States prepare, authorize, and
execute their budgets differently and, as a result, it is very difficult to
compare state budgets as they are presented by each state. One of the
challenges in our methodology was to ensure, to the extent possible, that
comparable program expenditures were presented

similarly for each state. For the most part, expenditures for social
services are made out of states? general funds and federal funds- as opposed
to states? capital funds which finance large scale construction and
infrastructure budgets. In Colorado, constitutional and statutory
restrictions limit expenditure growth and the size of the state?s rainy day
fund. Thus, in any given year if state revenues exceed expenditures, the
state rebates the surplus to taxpayers. A number of the rebates the state
offered in state fiscal year 2000 were targeted to low- income taxpayers,
but these rebates were authorized on a onetime basis only and contingent on
certification by the State Revenue Department that a state general fund
surplus occurred. Because these expenditures were never factored into the
states? baseline budget we did not consider them in our analysis of state

spending trends. Essentially we believe that these rebates did not present
decisionmakers with the same choices presented decisionmakers in other
states and should not be considered in our analysis of whether states used
TANF funds to supplant their state funds. Nevertheless, these targeted
rebates were considerable. In state fiscal year 2000, the state temporarily
expanded its EITC from 8. 5 percent of the federal credit to 10 percent at a
cost of about $32 million. It also temporarily increased the child care tax
credit to targeted to families with incomes less than $64, 000 at a cost of
about $61 million. If we had included

the rebates in our analysis, these expenditures would have appeared first in
our Category 2- Support for the Working Poor and cumulatively throughout
Categories 3 and 4.

In line with the caseload declines, total spending on basic welfare services
declined dramatically between 1995 and 2000- from $155 million to about $61
million. (See table 9.) This represents a drop of about 64 percent in

real terms, but state spending declined even further- by 76 percent in real
terms- from $70 million to about $17 million. Accordingly, the share of the
programs now financed with state funds declined from about 45 percent of the
total in 1995 to about 31 percent in 2000.

Tabl e 9: Col orado?s TANF- Related Expenditures

Dollars in millions

Real Nominal (expressed in 1995 dollars) 1995 2000 1995 2000

Basic welfare services Total funds $155 $61 $155 $55 State funds $70 $19 $70
$17 Support for the working poor Total funds $340 $380 $340 $347 State funds
$127 $129 $127 $118 Source: GAO survey and analysis of state expenditure
data.

However, for expenditures on a broader array of programs that support the
working poor, total expenditures increased 2 percent in real terms. This
trend was driven by increased spending of federal funds in child care and
child development programs. Despite the increase in total spending on this
category, the share of the expenditures financed with state funds declined
by about 7 percent in real terms. (See table 9.) However, as noted

previously, the state targeted $93 million in rebates of the state?s surplus
in 2000 to low- income families.

Table 10: Colorado?s TANF- Related Expenditures

Dollars in millions

Real Nominal (expressed in 1995 dollars) 1995 2000 1995 2000

Support for the working poor plus health care Total funds $894 $1,436 $894
$1, 309 State funds $385 $657 $385 $600 TANF- related social services Total
funds $1,188 $1, 842 $1,188 $1, 681 State funds $555 $924 $555 $843 Source:
GAO survey and analysis of state expenditure data.

By expanding our expenditure analysis to include the costs of health care
for the working poor, we find that the state has increased both the total
level of spending as well as the share of the total it finances from state
funds. Total spending on programs supporting the working poor and their
related health care costs rose by about 46 percent in real terms from 1995

to 2000. The share of those expenditures the state financed out of state
funds rose even further, however, by about 56 percent. Accordingly, the
share financed by the state increased since 1995 when it financed about 43
percent of total costs; in 2000, the state?s share was 46 percent.

Under the broadest definition of TANF- related social services including
child welfare programs, transitional services for the developmentally
disabled, mental health services, and substance abuse programs- cumulative
expenditures on those programs as well as those previously discussed in this
analysis rose by 42 percent in inflation adjusted terms.

The state?s share of the total rose even more, however, and as a result the
state now finances a larger share of the total costs. In 1995, the state
financed 47 percent of the total and, in 2000, it financed 50 percent of the
total.

Additional federal safety- net programs provide food assistance and
supplemental security income for needy families with children in Colorado,
but are not shown in figure 5. The federal government provided $213 million
in food stamp benefits to low- income individuals in the state in 1995, and
$130 million in 2000. In addition, from 1995 to 2000, federal supplemental
security income payments increased from $216 million to

about $239 million in nominal dollars- about 82 percent of these benefits
went to individuals under 65 years of age. Rainy Day Funds Colorado has
established reserves of TANF funds; a short- term emergency

fund for the state Department of Human Services to use to address
contingencies, a long- term reserve that could be used in the event of an
economic downturn, and a county reserve fund. Since Colorado allocates

most of its TANF block grant to counties, some counties have been able to
amass significant reserves. While these reserves remain in the U. S Treasury
until the counties need them, these funds remain under county control. The
fundamental purpose of the county reserves is to provide for the maximum
amount of flexibility to counties to implement the provisions of welfare
reform to meet local needs. However, many county officials

view these reserves as the first line of defense for the counties to turn to
in the event of a recession.

Because TANF funds are allocated to the counties under the memorandum of
understanding, the state records all TANF funds including the county
reserves as an obligation by the state even though the county might not have
made specific decisions regarding their plans for these funds. As of June
30, 2000, the end of the state?s fiscal year, the state estimated that the

counties held about $67 million in reserves. The state?s TANF financial
reports for the end of the federal fiscal year 2000- September 30, 2000-
however, record no unobligated balances but $94 million in unliquidated
obligations. (See table 11).

Table 11: Unspent TANF Funds in Colorado as of September 30, 2000

Dollars in millions

Unspent funds as a Unobligated

Unliquidated Tot al unspent

percent of annual balance obligations funds

TANF grant

0 $94. 2 $94.2 69% Source: HHS? Administration for Children and Families.

In addition to the county reserves, the state maintains both its short- term
and long- term reserves. The ?Short- term Emergency Works Fund? receives an
annual appropriation of $3 million of federal TANF funds. These funds are
provided to the state Department of Human Services to use in response

to counties for emergencies. Depending on the circumstances, the counties
are required to deplete their allocation and their reserve before gaining
access to these emergency funds. At the end of the state fiscal year,

the balance in the fund is reduced to zero and any funds not expended are
transferred to the ?Long- term Works Reserve Fund.?

The long- term reserve was designed to provide the state and the counties
with enough funds to continue the state?s welfare program in the event of an
economic downturn. The state?s Department of Human Services (DHS) developed
a model to forecast how the state?s welfare costs might respond under a
variety of economic scenarios. Although DHS presented the results of the
model to the legislature as a basis for determining the appropriate size for
the Long- term Works Reserve Fund, a legislative analyst told us the model
was flawed. While the model forecasts changes in caseloads, it multiples the
higher caseloads by a factor that is less than

the minimum monthly assistance grant for a family of three. The model does
not factor in other costs for employment and training programs, child care
subsidies, or any other transitional service. The state funds the long- term
reserve a number of ways. First, at the end of the state fiscal year, it
assesses each county?s expenditures against its allocation. Counties retain
the first 20 percent of their unspent balance

and fifty percent of the remaining unspent balance. The remainder reverts to
the state?s Long- term Reserve Fund. In addition, each year the balance from
the state?s short- term reserve reverts to the long- term reserve, as well
as any federal TANF funds made available to the state- e. g., through the

high performance bonus- that have not otherwise been appropriated or spent.
While the funds in the long- term reserve are federal TANF funds and can
only be used for the Colorado Works program, there is no statutory
prohibition against the legislature making appropriations from these funds

for a number of TANF- related purposes, in addition to an economic
contingency.

Appendi x V I Connecticut Background

State fiscal year: July 1- June 30 Budget cycle Biennial TANF grant: a $267
million 75% MOE: $183 million $183 million Cash assistance caseload: b 61,
000 families in 1995

27, 000 families in 2000 55 percent decline a Connecticut received $2.5
million in 2000 and $2.6 million in 2001 in federal TANF funds as bonuses
for successful program outcomes.

b Caseload as reported to HHS for January 1995 and June 2000 and rounded to
the nearest thousand.

Connecticut launched its statewide welfare program, Jobs First, in January
1996, based on earlier demonstration programs operated under federal waivers
that began in 1995 and continue through September 2001. The purpose of Jobs
First is to encourage able- bodied individuals to actively seek, obtain, and
retain employment.

To receive cash assistance, job- ready adults in families must look for work
immediately, in accordance with the program?s work- first philosophy. As an
incentive, families may continue to receive their full monthly cash
assistance payments while they are working as long as their earnings are
below the federal poverty level and they have not exceeded the program?s
time limits on cash receipt. Families in Jobs First may also retain up to
$100 in child support collected on their behalf by the state agency without
any reduction in their cash assistance, and they are eligible for medical
assistance. They may also receive help with transportation and child care,
if necessary to facilitate their employment or job search.

The maximum cash assistance payment is $543 a month for a family of three,
an amount unchanged since the enactment of TANF. However, eligibility for
assistance has eased, as permitted under TANF, making families eligible with
incomes at or below 75 percent of the state?s median income, which
approximates 335 percent of the federal poverty rate.

Alternatively, eligible applicants may choose diversion payments equivalent
to 3 months of cash assistance and forgo assistance under the Jobs First
program. Receipt of cash assistance under Jobs First is limited by state law
to 21 months for those considered job ready but may be extended in

increments of 6 months for families with incomes below the monthly cash
assistance payment level. Families who lose assistance due to time limits
and who do not have

income at or above the cash assistance payment standard are referred to the
state- funded safety net program called Worksteps for counseling, food,
clothing, and up to 12 months of vouchers for rental assistance. Working
families who leave Jobs First may receive transitional medical assistance
for 2 years after they leave the program. Subsidized health care is also
available to children of the working poor whose incomes are at or below 185
percent of the poverty level. Working families may receive subsidized
childcare if their incomes are below 75 percent of the state median income.

Needy two- parent families who meet eligibility guidelines are served under
a separate, state program with policies similar to Jobs First. Recent legal
immigrants who are not eligible for TANF may receive cash and medical
assistance under a separate program supported with state funds. Both these
programs count towards state MOE.

The state operates Jobs First directly through its social service offices
located in five regions that determine eligibility and arrange for
electronic payments of cash assistance. Since 1998, job training and
employment services are provided through the state?s Department of Labor and
its regional workforce boards and subcontractors. In addition to Jobs First
and related programs, TANF funds are used to support child welfare and

substance abuse services as well as teenage pregnancy prevention pilots in
eight urban and one rural area. Supplantation in the

Connecticut?s policy is to draw down its entire TANF grant, leaving nothing
State Budget

in reserve. Historically, most of Connecticut?s TANF grant has been expended
through traditional welfare programs within the control of the state social
services agency. As caseloads have fallen, it has become increasingly
difficult for the agency to spend the entire grant and at the end of federal
fiscal year 1999, the state reported nearly $40.7 million of unspent TANF
funds. Budget officials said they asked the agency to pursue options that
would allow the state to use TANF funds instead of

state funds in programs operated outside the agency. Using TANF funds, the
agency hired a contractor to help identify such program expenditures, and in
June 2000, retroactively drew down $40 million of its TANF balances for
expenditures the state made in 1999. In addition to these direct- TANF
claims, by maximizing the amount of TANF funds it transfers to the Social

Services Block Grant, the state also frees up its funds from the budgets of
a variety of other state- funded programs. The state has reported saving
about $24 million a year in state funds through this transfer mechanism.

The state initiated expansions in child welfare programs and early childhood
development programs with state funds between 1995 and 2000; once
significant TANF- savings were achieved as a result of the large decreases
in cash assistance caseloads, state budget officials contend that these
savings were then reinvested into those child welfare and early

childhood development programs. State officials said that the expenditures
the contractor helped them identify were- in large measure- from these child
welfare and early childhood development programs.

Given current caseload levels, state budget officials said that in limiting
the use of TANF funds to only those programs operated by the state social
services agency it would continue to be difficult to spend the state?s
entire

TANF grant. As a result, the state will most likely continue to claim the
full TANF grant each year in order to cover expenditures in other
departments which had previously been paid for with state funds State budget
officials said that program expansion is difficult in Connecticut because of
a spending cap that applies to both federal funds and state funds and
therefore limits the growth of new programs and new spending regardless of
the source of financing.

State Spending Trends From 1995 to 2000

Figure 6: Connecticut?s TANF- Related Expenditures for State Fiscal Years
1995 and 2000 1995 dollars in millions 4, 000

3, 500 3, 000 2, 500 2, 000 1, 500 1, 000

500 0

95 00 95 00 95 00 95 00 Basic welfare Support for

Support for TANF- related

services working poor working poor +

social services health care

Federal State

To assess how state budget decisions might have affected spending, we
collected data on how Connecticut uses both state and federal funds in its
TANF- related social services programs. The results are grouped into four
categories, as shown in figure 6 and described further in appendix I. Each
category is identified on the graph by two bars-- one for spending that
occurred in state fiscal year 1995 and one for spending that occurred in

state fiscal year 2000. From left to right, the categories are cumulative.
Expenditures for cash assistance and employment training are included in
expenditures for support of the working poor. These expenditures in turn are
included in expenditures for support of the working poor plus health care.
Finally, all expenditures are included in social services expenditures.

Spending on cash assistance payments in Connecticut has gone down by about
$250 million in nominal dollars from state fiscal year 1995 through

2000, as caseloads have declined and the monthly assistance level has
remained constant. (See table 12.) At the same time, spending on job
training has increased by about $40 million so that it nearly equals the
amount spent on cash assistance. Adjusted for inflation, overall spending on
cash assistance and job training has declined 42 percent. The state?s

share of cash assistance and job training costs has fallen relative to the
federal share. Where the state paid about 52 percent of these costs in 1995,
it paid only 46 percent in 2000.

Table 12: Connecticut?s TANF- Related Expenditures

Dollars in millions

Real Nominal (expressed in 1995 dollars) 1995 2000 1995 2000

Basic welfare services Total funds $574 $365 $574 $333 State funds $300 $168
$300 $153 Support for the working poor Total funds $1, 087 $967 $1, 087 $883
State funds $618 $534 $618 $488 Source: GAO survey and analysis of state
expenditure data.

Looking at the broader picture of expenditures for welfare recipients and
the working poor in Connecticut that includes the totals above and child
care shows less steep declines. (See table 12.) Combined expenditures fell
by $120 million from 1995 through 2000 in nominal dollars, a drop of 19

percent in real terms. The decline is moderated because the aforementioned
reductions in cash assistance are offset to some extent by increases
elsewhere. Spending on child care accounts for the largest

increase, more than doubling to $181 million in 2000. Budget officials note
that some of the additional funds targeted for child care remain unspent; it
is unclear whether this is because demand has leveled off or available child
care is difficult to access. The state paid 57 percent of all these costs in
1995; by 2000 its share had dropped by 2 percent relative to the federal
share. Total spending on health care costs for low- income individuals and
families increased by nearly $100 million, but the increase was not enough
to offset the reductions in programs to support the working poor and basic
welfare

services. (See table 13.) When healthcare expenditures are added to the
totals described above, combined spending shows a slight decline of $24
million in nominal dollars from state fiscal years 1995 through 2000,
representing a decline of 10 percent in real dollars. However, the state?s
share of funding has remained stable- about 56 percent- since 1995.

Table 13: Connecticut?s TANF- Related Expenditures

Dollars in millions

Real Nominal (expressed in 1995 dollars) 1995 2000 1995 2000

Support for the working poor plus health care Total funds $2, 567 $ 2,543
$2, 567 $2, 320 State funds $1, 452 $1, 417 $1, 452 $1, 293 TANF- related
social services Total Funds $3, 371 $3, 664 $3, 371 $3, 343 State Funds $2,
060 $2, 229 $2, 060 $2, 097 Source: GAO survey and analysis of state
expenditure data.

In contrast, a different trend appears when costs of child welfare and other
social services such as services for those with developmental disabilities,
substance abuse problems, and mental health needs are included in the totals
above. Expenditures for all social services declined slightly in real terms
from 1995 through 2000 (see table 13), but state spending shows a small
increase. Costs of these services have been borne primarily by the

state, but some shifts occurred between 1995 and 2000, such that the federal
share of child welfare increased modestly- as evidenced by the state?s use
of TANF funds in place of state funds in these programs- while the state
share of other services increased modestly. Overall for TANFrelated
expenditures on social services that include basic welfare services, support
for the working poor, child care and health care, the state?s

investment has increased relative to the federal investment; the state share
rose from 61 percent in 1995 to about 63 percent in 2000.

Additional federal safety- net programs provide food assistance and
supplemental security income for needy families and children in Connecticut.
For example, the federal government provided $204 million

in food stamp benefits to low income individuals in Connecticut in 1995 and
$197 million in 2000. In addition, from 1995 to 2000, federal

supplemental security income payments increased from $176 billion to $223
billion with about 82 percent of these benefits going to individuals under
65 years of age. Rainy Day Funds According to budget officials, Connecticut
employs a unique system of budgeting they term ?gross budgeting.? An agency
receives an

appropriation deemed sufficient to cover its priorities for the budget year
without regard to the source of funding. The agency is expected to allocate
its expenditures to the appropriate federal or state funding source and
document all claims to federal funds. This claiming process is revisited at
year- end to ensure that the agency maximizes the use of federal funds

available. As noted previously, Connecticut?s budgetary policy is to spend
its entire TANF grant, leaving no balances at the U. S. Treasury. State
officials said their budget strategy is to provide the agency with the
financing it needs to accomplish its objectives and to require the agency to
find a way to claim all available federal funding. In addition, state budget
officials explained

that there was concern that any balances left might be rescinded by the
Congress.

In line with its policy, Connecticut reports no unspent TANF funds through
September 30, 2000. (See table 14.)

Table 14: Unspent TANF Funds in Connecticut as of September 30, 2000 Unspent
funds as a Unobligated

Unliquidated obligations Total unspent

percent of TANF TANF funds of TANF funds funds

annual grant

0 000% Source: HHS? Administration for Children and Families.

Accordingly, Connecticut does not have a TANF rainy day fund. The state does
not have a formal mechanism or fund to prepare for shortfalls in funding for
its future social service needs. Budget officials said that the state has
always placed a priority on serving the needy and historically has found a
way to finance these programs.

Appendi x V

Louisiana Background

State fiscal year: July 1- June 30 Budget cycle: Annual TANF grant: a $164
million 75% MOE: $55 million Case assistance caseload: b 82, 000 families in
1995

26, 000 families in 2000 69 percent decline a Louisiana receives from $4
million to $13 million in annual TANF supplements because the state?s pre-
TANF expenditures per poor person were less than 35 percent of the national
average. In addition, Louisiana received a bonus of $3.7 million in 2000 for
positive program outcomes.

b Caseload as reported to HHS for January 1995 and June 2000 and rounded to
nearest thousand.

The welfare reform that took effect in Louisiana in January 1997 consists of
two components: (1) a cash assistance component called the Family
Independence Temporary Assistance Program (FITAP) and (2) an employment and
training component called FIND Work. The overall goal is to decrease long-
term dependency on cash assistance through the promotion of job preparation
and work. Although work is the goal, the program is not based on a strict
work- first approach. Instead, families may receive cash assistance for up
to 2 years, the limit set in TANF law, before they are required to
participate in work activities or lose their eligibility for continued
assistance. Those who are

completing education or training programs, actively seeking work, employed a
short time with limited earnings, or otherwise exempt may continue to
receive cash payments beyond the 2 years. Those who are not exempt and who
have received cash assistance for 2 years must wait another 3 years before
they are eligible for cash assistance again. After the equivalent of 5 years
of cash receipt, they are ineligible, in accordance with TANF. The maximum
cash payment for a family of three is $240 a month, an

increase from the $190 a month that was in effect from 1981 through 2000.
Needy families with children whose adjusted incomes are below the payment
level are eligible for cash assistance. State policy requires that
applicants be screened for drug abuse and, where appropriate, assessed,
tested, and treated. To facilitate their participation in activities under
the FIND Work component, families may receive transportation subsidies up to
$60 a month, child care, and case management. Subsidized jobs may be

provided where unemployment is high in order to help meet work participation
requirements. To provide a financial incentive to work, families who obtain
unsubsidized jobs may continue to receive their full cash assistance
payments for 6 months, if their earnings are within certain limits. When
their earnings make them ineligible for monthly cash

assistance, they may receive 12 months of transitional medical and child
care assistance. For needy children who are not in the care of their
parents, Louisiana

instituted a kinship care program in March 2000. Under this program, needy
children are eligible for a monthly cash payment if they are legally in the
care of relatives other than their parents, and if the relatives? income is
less than 150 percent of the federal poverty level. In July 2000, the state

expanded its teen pregnancy prevention efforts, in line with the new TANF
goals.

A single state agency in Louisiana administers both FITAP and FIND Work.
State staffs working out of offices in the local parishes 1 throughout the
state process applications and determine eligibility for cash assistance
payments that are made electronically. Work- related activities and other
services are provided under contracts with the state agency.

Supplantation in the We found little evidence of supplantation in
Louisiana?s budget since TANF

State Budget was enacted. A state- funded teenage pregnancy prevention
program was

moved to the state social services agency in state fiscal year 2000,
expanded statewide, and funded with $3. 5 million in TANF funds instead of
state funds. The freed- up state funds were used on other state priorities.

According to senior budget officials, the lead TANF agency in the state- the
Department of Social Services (DSS)- controls the use of the federal TANF
funds in the budget process. Senior DSS officials said that they have been
very concerned about the perception federal oversight authorities

might get of Louisiana if the stateLouisiana used the TANF funds to replace
state funds- even if the rules allowed it. The executive budget office and
the legislature have largely deferred to the agency when budgeting for TANF.

1 Parishes are similar to counties in that they are a subunit of government.

State budget officials said that state general funds for social services are
relatively scarce, and there are very few programs financed with state funds
that are not already being used to match federal funds in other programs or
to meet the TANF maintenance- of- effort requirement. Given the span of
control of DSS, there is little incentive for the agency to look outside its
own budget for programs that could be financed with TANF funds. Thus, the
cautious approach to replacing state funds with TANF funds is due, in part,
to the limited role played by the state budget office and the legislature in
directing the use of the TANF funds to other areas within the state budget.

Spending Trends From 1995 to 2000

Figure 7: Louisiana?s TANF- Related Expenditures for State Fiscal Years 1995
and 2000 1995 dollars in millions 3, 000

2, 500 2, 000 1, 500 1, 000

500 0

95 00 95 00 95 00 95 00 Basic welfare Support for Support for

TANF- related services

working poor working poor +

social services healthcare

Federal State

To assess how state budget decisions might have affected spending, we
collected data on how Louisiana uses state and federal funds to finance its
TANF- related social services expenditures. The results are grouped into
four categories, as shown in figure 7 and described further in appendix I.

Each category is identified on the graph by two bars- one for spending that
occurred in state fiscal year 1995 and one for spending that occurred in
state fiscal year 2000. From left to right, the categories are cumulative.
Expenditures for cash assistance and employment training are included in
expenditures for support of the working poor. These expenditures in turn

are included in expenditures for support of the working poor plus health
care. Finally, all expenditures are included in social services
expenditures.

Spending of combined state and federal funds for cash assistance and job
training in Louisiana has declined by 21 percent in real terms from 1995
through 2000 as the number of families receiving cash assistance has fallen.
The state?s investment has declined from $69 million to $55 million- a real
drop of 28 percent. The state?s share of total expenditures slipped declined

from 23 percent in state fiscal year 1995 to 21 percent in 2000. Federal
funds make up the difference and continue to support most of the costs of
cash assistance and job training in Louisiana, as is evident in table 15.

Table 15: Louisiana?s TANF- Related Expenditures

Dollars in millions

Real Nominal (expressed in 1995 dollars) 1995 2000 1995 2000

Basic welfare services Total funds $305 $263 $305 $240 State funds $69 $55
$69 $50 Support for the working poor Total funds $522 $581 $522 $530 State
funds $117 $120 $117 $109 Source: GAO survey and analysis of state
expenditure data.

Real gains are shown in total spending on programs that support the working
poor by helping low- income families transition- or stay- off public
assistance. In addition to FITAP, these services include child care,
transportation subsidies, and employment services. Total spending rose by
nearly $60 million- a real gain of about 2 percent. Driving this trend is
increased spending on child care, which more than tripled from 1995 through
2000. However, the state used federal funds to finance most of this
increase. While state spending remained virtually unchanged in nominal
terms, it declined by about 7 percent in real terms. Again, the state?s
share

of total expenditures has slipped declined from 23 percent in state fiscal
year 1995 to 21 percent in 2000. Total spending on traditional welfare
programs, supportive services for the working poor, and their related health
care costs grew by about 2 percent in real terms from 1995 through 2000. In
nominal dollars total expenditures increased from $2. 1 billion to $2.3
billion, as shown in table 16. State spending rose by nearly $100 million-
or 8 percent in real terms. Federal spending stayed relatively constant in
real terms. The relative shares of the costs of these programs changed as a
result. The state now finances about 28 percent of these programs, but in
1995 it financed about 26 percent. Table 16: Louisiana?s TANF- Related
Expenditures

Dollars in Millions

Real Nominal (expressed in 1995 dollars) 1995 2000 1995 2000

Support for the working poor plus health care Total funds $2,083 $2,339 $2,
083 $2, 134 State funds $543 $643 $543 $587 TANF- related social services
Total funds $2,338 $2,681 $2, 338 $2, 446 State funds $663 $821 $663 $749
Source: GAO survey and analysis of state expenditure data.

An upward trend is also evident when costs of child welfare and other social
services such as those for substance abuse, mental health, and developmental
disabilities, are included in the totals above. Total expenditures for all
social services increased from $2.3 billion to $2. 7 billion in nominal
dollars from state fiscal years 1995 through 2000-

an increase of 5 percent in real terms. State spending increased by about 13
percent in real terms. The state share of all these programs costs rose from
about 28 percent in 1995 to about 31 percent in 2000.

Additional federal safety- net programs provide food assistance and
supplemental security income for needy families and children in Louisiana.
The federal government provided $634 million in food stamp benefits to low-
income individuals in Louisiana in 1995 and $452 million in 2000. In
addition, from 1995 to 2000, federal supplemental security income

payments increased from $715 million to $745 million, about 85 percent of
which went to individuals under the age of 65.

Rainy Day Funds Louisiana has left a large share of its TANF funds unspent.
As of September 30, 2000 Louisiana reported about $169 million in unspent
TANF balances, more than the state?s annual block grant award and about 26
percent of the cumulative TANF funds awarded to the state since 1996. (See
table 17.) While these funds must remain in the U. S. Treasury until they
are needed, state program officials said that they consider these

balances to be a contingency fund. They added that they would use the funds
to cover unexpected program costs in the event the appropriations estimates
were too low. Budget officials told us that no study had been conducted on
how much the state would need to maintain in a reserve in order to be able
to finance its welfare program during an economic downturn. DSS budget
officials said they were concerned about the impact a recession might have
on the program?s budget and sought to set aside a

significant TANF reserve to ensure that the state had access to additional
program funds. While DSS did not establish a specific balance for its
reserve, budget officials believe that now that it has reached the level of
the state?s annual block grant they might have a better chance of making
their case in the legislature for using more of the future grants to create
new

programs or expand existing programs.

Table 17: Unspent TANF Funds in Louisiana as of September 30, 2000

Dollars in millions

Unliquidated Unspent funds as a

Unobligated obligations of

Total unspent percent of annual

TANF funds TANF funds funds

TANF grant

$169 0 $169 103. 1% Source: HHS? Administration for Children and Families.

State program and budget officials acknowledge that the budget climate in
the state makes it very difficult to raise spending for social services even
if the state can use federal funds to finance the programs. Once new
programs are enacted they create new constituencies that will continue to
demand services. These officials expressed concerns that if federal funding
were ever reduced or some restrictions placed on the block grant

that affected the state?s flexibility, the state would have to find a way to
finance the program with its own funds.

State program officials said that they have known for some time that in
order to spend more of the TANF funds that were available on a declining
caseload, they would have to expand programs or create new programs and
services to help FITAP families transition into the workforce. They also
said that it is very difficult to get the legislature to act on these
requests. For example, they said it took 3 years to negotiate an increase in
the monthly cash grant for a family of three from $190 to $240. This

increase, which took place in state fiscal year 2000 was the first increase
in grant levels since 1981. Based on current caseload projections this
increase in the cash grant was expected to cost an additional $18.6 million
per year and was funded entirely with federal TANF funds. During the same
fiscal year the state expanded access to its subsidized child care program-
at a cost of $24 million, financed entirely with TANF funds- that was
expected to serve an additional 8,000 eligible children.

Although legislative analysts disagreed with the department?s assessment
that the legislature has opposed expanding the use of TANF funds, they
acknowledged that there is a conservative political environment in the
state. They added, however, that they plan to begin pressing DSS to explore
more options for spending the state?s TANF resources and for slowing the
growth in unspent balances.

Appendi x VI

Maryland Background

State fiscal year: July 1- June 30 Budget cycle Annual TANF grant: $229
million 75% MOE: $177 million Cash assistance caseload: a 81, 000 families
in 1995

29, 000 families in 2000 64 percent decline a Caseloads as reported to HHS
for January 1995 and June 2000 and rounded to the nearest thousand.

Maryland?s welfare reform effort, known as the Family Investment Program
(FIP), was initiated in October 1996. The program represents a major shift
from an income maintenance focus to one that seeks to promote family
independence through work, personal and family responsibility, and

community involvement. To receive cash assistance under FIP, adults in
eligible families must be willing to seek work immediately. The maximum cash
assistance level for a family of three is $439 a month, an amount that has
increased since passage of TANF to keep pace with inflation. Medical
assistance coverage is automatic for families receiving cash aid. Needy
families with children whose adjusted incomes are below the monthly cash
assistance levels are eligible for FIP. Support services available to FIP
families vary by county and may include

transportation, child care, referral to family planning, and substance abuse
treatment in addition to assessment, job search, and employment services. In
place of monthly cash assistance, diversion payments are available to
eligible families, according to policies established by each county. After 3
months of cash receipt, a working family may receive child care and

medical assistance for up to 12 additional months. Working low- income
families may be eligible for the state?s earned income tax credit. Separate,
state- funded programs are offered for needy twoparent families, legal
immigrants not otherwise eligible for TANF, children in the care of
relatives who are not their parents, and certain other groups. These
programs, and the state?s earned income tax credit (EITC), are used toward
the state?s MOE.

A single Maryland state agency supervises FIP, which is administered at the
local level by county social service agencies and by contract with
nonprofits and other organizations. In addition to FIP, TANF funds are used
to support a broad array of other programs. These include programs that
provide work opportunities for youth, summer activities designed to

lower school drop- out rates, home visits with expectant parents who are at
risk of welfare dependency, family support centers, independent living
skills for teen parents, family preservation, and other child welfare
services.

Supplantation in the In preparing the FIP budget for state fiscal year 2001,
the state sought to State Budget

replace state funding in a variety of program accounts with federal TANF
funds. However, instead of using the freed- up state funds for other
spending priorities Maryland set them aside in the Dedicated Purpose

Account- essentially a rainy day fund that can only be spent on programs for
low- income families. State budget officials suggested that substituting
state funds with TANF funds in this way was a more fiscally prudent approach
for using their unspent TANF balances than spending all savings

on expanding programs and services. They said that their decision was rooted
in concerns that Congress would recategorize the block grant and restrict
state flexibility, or rescind TANF funds left unspent by the states. There
was also concern that the TANF block grant might be reduced in
reauthorization, and if this occurred, the state could be left with high
expectations and demands for services previously funded with federal dollars
that would have to be continued with state dollars. These concerns, coupled
with the desire to prepare for economic contingencies, motivated

the state to consider using unspent TANF funds to replace state funds which
were then set aside in the Dedicated Purpose Account.

Maryland?s fiscal year 2001 budget used $30 million in federal TANF dollars
to continue several programs formerly supported with state funds. A state
budget official said that if the FIP caseload and expenditures remain
stable, then there will be sufficient TANF funds available over the next 2
years to continue to finance these programs and to allow the state to
deposit its freed- up funds in the Dedicated Purpose Account. The state also

retroactively adjusted its state fiscal year 2000 appropriations and
authorized a shift of $22.3 million in state funds out of several child
welfare and pregnancy prevention programs.

Spending Trends From 1995 to 2000

Figure 8: Maryland?s TANF- Related Expenditures for State Fiscal Years 1995
and 2000

1995 dollars in millions 3, 000

2, 500 2, 000 1, 500 1, 000

500 0

95 00 95 00 95 00 95 00 Basic welfare Support for Support for

TANF- related services

working poor working poor +

social services health care

Federal State

To assess how state budget decisions might have affected spending, we
collected data on how Maryland uses both state and federal funds in its
TANF- related social services programs. The results are grouped into four
categories, as shown in figure 8 and described further in appendix I. Each
category is identified on the graph by two bars- one for spending that

occurred in state fiscal year 1995 and one for spending that occurred in
state fiscal year 2000. From left to right, the categories are cumulative.
Expenditures for basic welfare services like cash assistance and

employment training are included in expenditures for support of the working
poor. These expenditures in turn are included in expenditures for support of
the working poor plus health care. Finally, all expenditures are included in
TANF- related social services expenditures.

Spending on basic welfare services in Maryland has been cut from $417
million in state fiscal year 1995 to $257 million in 2000 as a direct result
of the smaller caseload. (See table 18.) In real terms, spending has
declined 44 percent, despite higher monthly cash assistance payment levels
for FIP families. Before welfare reform was enacted, the state contributed

about 41 percent of the expenditures in these areas. Since welfare reform
was enacted, however, the state?s and counties? contributions for cash
assistance and job training has dropped to 32 percent. Table 18: Maryland?s
TANF- Related Expenditures

Dollars in millions

Real Nominal (expressed in 1995 dollars) 1995 2000 1995 2000

Basic welfare services Total funds $417 $257 $417 $ 234 State funds $170 $
81 $170 $74 Support for the working poor Total funds $837 $835 $837 $762
State funds $377 $345 $377 $315 Source: GAO survey and analysis of state
expenditure data.

The spending decline is less steep when programs that provide support for
the working poor are also considered. In this category, spending slipped
from $837 million in 1995 to $835 million in 2000 in nominal dollars.
Adjusted for inflation, this represents a decline of 9 percent. In line with
the state?s decision to make child care more affordable and accessible,
child

care spending has grown by $47 million so that it nearly equals spending on
cash assistance. Since 1995, Maryland?s share of spending on these programs
has declined by about 4 percent. When we include health care expenditures in
our analysis, the trend is reversed- total spending increases from $1. 2
billion to $1.5 billion over the same time period. (See table 19.) This
represents an increase in real terms of about 16 percent. Combined federal-
state spending on health care for the working poor as well as those in FIP
has nearly doubled to $695 million. Since 1995, Maryland?s share of all
spending on basic welfare services,

support for the working poor, and health care has declined only slightly- by
about 1 percent.

Table 19: Maryland?s TANF- Related Expenditures

Dollars in millions

Real Nominal (expressed in 1995 dollars) 1995 2000 1995 2000

Support for the working poor plus health care Total funds $1, 208 $1, 530
$1, 208 $1, 396 State funds $560 $689 $560 $628 TANF- related social
services Total funds $2, 261 $2, 897 $2, 261 $2, 644 State funds $1, 273 $1,
552 $1, 273 $1, 416 Source: GAO survey and analysis of state fiscal survey.

An upward trend in total spending is also evident when costs of child
welfare, juvenile justice, and other social services such as services for
the developmentally disabled, substance abuse treatment, and mental health
services are included in the totals above. Expenditures for all TANFrelated
social services increased from $2. 3 billion to $2. 9 billion in nominal
dollars from state fiscal year 1995 to state fiscal year 2000. In real
terms, this represents an increase of 17 percent. On average for all
services included in our survey, Maryland funds about 54 percent of costs,
down

about 3 percent since welfare reform was enacted. Additional federal safety-
net programs provide food assistance and supplemental security income for
needy families with children in Maryland, but they are not shown in figure
8. The federal government provided $366 million in food stamp benefits to
low- income individuals in Maryland in 1995, and $204 million in 2000. In
addition, from 1995 to 2000, federal supplemental security income payments
increased from $333 million to

$411 million in nominal dollars- about 80 percent of these benefits went to
individuals under 65 years of age. Rainy Day Funds Maryland is the only
state in our study to set aside state funds rather than

federal funds for economic contingencies. As early as 1997, Maryland placed
$15.7 million of its state general funds in the Dedicated Purpose Account
for welfare programs, and in 2001, the state deposited an additional $52
million in state funds into this account. But as previously

noted, Maryland found a way to transfer the costs of saving state funds to
the federal government. As with states that set aside a portion of their
federal TANF block grant funds for a rainy day, the levels of Maryland?s
reserves do not appear to be based on any sort of contingency planning or
analysis. While an economic model which forecasted how caseloads- and
related costs- might respond under a variety of economic scenarios was
presented to the legislature, a senior budget official acknowledged that the
amount deposited in the Dedicated Purpose Account was based on the amount of
state funding that could be replaced with federal TANF funds.

In Maryland there are several safeguards in place to help ensure that these
funds remain in reserve until they are needed to finance budget shortfalls
in welfare programs. The executive branch must get legislative approval for

any appropriation from the fund, and under state statute, funds deposited to
a dedicated purpose reserve must be spent on purposes for which the fund was
established. If, after 4 years, the funds are not spent, they revert to the
state?s general purpose rainy day fund. However, the state can

reauthorize the dedicated reserve fund for an additional four years, as it
did with the original $15.7 million deposited to the fund in 1997. Senior
budget officials said that they expect the state legislature to continue to

appropriate these funds for their designated purpose- the state?s welfare
program.

Although Maryland has shifted TANF funds into various state programs outside
of FIP, the state has accumulated significant reserves of unspent TANF
funds, representing 45 percent of its annual grant. See table 20 for the
TANF funds that the state reported as unspent through September 20,

2000.

Table 20: Unspent TANF Funds in Maryland as of September 30, 2000

Dollars in millions

Unspent funds as a Unobligated

Unliquidated obligations Tot al unspent

percent of TANF TANF funds of TANF funds funds

annual grant

$49. 5 $54. 2 $103. 7 45. 3% Source: HHS? Administration for Children and
Families.

Amounts reported as unobligated are due in large part to delays in getting
the new program up and running and to the larger than expected drop in the
number of families the program is serving, according to a senior budget
official. Funds reported as unliquidated are due primarily to the structure
of the state?s operation and its reliance on county social service agencies
and other contractors for the provision of services. Unlike other states
with county- administered welfare programs, Maryland records obligations of
TANF funds only when the counties enter into contracts for specific services
or have been billed for services already provided. Many local social service
offices were not ready for the significant changes required by welfare
reform, and they were slow to start spending the federal and state funds
that were available. Even after these programs were established, the program
management inexperience of many local social service personnel contributed
to delays in disbursing funds.

In developing the state?s 2001 budget, Maryland included about $160 million
in unspent TANF funds in addition to its annual block grant. The state
dramatically increased spending on child care by raising the income
eligibility requirements and thereby allowing more families access to the
subsidies. State officials said that these budget actions, once fully
implemented, should succeed in drawing down the unspent TANF balances. In
fact, according to budget estimates, the state?s basic TANF grant is not
sufficient to sustain this level of funding in future years. If budget
forecasts prove accurate, some analysts predict the state will have

to cut back some funding from these new programs or increase the levels of
general funds it contributes to maintain the programs? funding levels.

Appendi x VII

Michigan Background

State fiscal year: October 1- September 30 Budget cycle: Annual TANF grant:
a $775 million 75% MOE: $469 million

Cash assistance caseload: b 207,000 families in 1995 71, 000 families in
2000 66 percent decline a Michigan received an additional $3 million in
federal TANF funds in 2000 as a bonus for work- related program outcomes and
$20 million in both 1999 and 2000 for reductions in out- of- wedlock births.
b Caseloads as reported to HHS for January 1995 and June 2000 and rounded to
the nearest thousand.

Under a series of waivers of federal rules, Michigan initiated its welfare
reform program from 1992 through1996 through incremental policy changes
applied to a research and demonstration project. The statewide program that
resulted in October 1996 from these early experiments consists of two
components: a cash assistance component called the Michigan Family
Independence Program (FIP) and an employment and training component called
Work First. As the name implies, the program is guided by work- first
principles, and seeks to help families achieve the maximum possible self-
support and independence.

To receive cash assistance, the adult in the family must participate in work
or related activities within 2 months of enrollment unless he or she meets
established criteria for deferrals from work requirements. The maximum
amount of cash assistance paid has not changed since passage of TANF and is
$459 per month for a family of three living in Wayne County, the state?s

most populous county. Needy families with children whose adjusted incomes
are below the monthly cash assistance levels are eligible for the program.
Once they become eligible, there is no time limit on families? receipt of
cash assistance, because families in need of assistance will be

supported with state funds when federal TANF funds can no longer be used. In
addition to cash assistance, families may receive up to $50 in child support
collected by the state agency on their behalf. 1

1 The remainder is retained by the state and federal governments as
reimbursement for cash assistance payments made to these families.

After attending orientation and signing a personal responsibility plan and
family contract, program participants are assigned to job search, subsidized
work, or other activities. To make it possible for participants to work or
attend other required activities, they may receive case management services,
emergency assistance, transportation, subsidized child care, rental
subsidies for up to 2 years, and individual development accounts for future
education or home ownership. These services and

benefits are supported with TANF funds and are also available to families
with incomes below 200 percent of the federal poverty level who are not
enrolled in the program. When program participants find work and no longer
need cash assistance, they remain eligible for transitional childcare and
medical assistance.

Outside of the program, TANF funds are used for several activities that are
designed to help prevent poverty, promote child welfare by allowing children
to remain in their own homes, and encourage family formation. These include
state?s supplemental security income payments, family preservation and
support services, and family planning and teen pregnancy

prevention efforts. They also include subsidies to families with incomes up
to $60, 000 to allow them to care for their disabled children in their homes
and subsidies to families with incomes under 500 percent of the federal

poverty level who adopt special needs children. In addition, Michigan helps
families with incomes below 80 percent of median income levels obtain
affordable housing by subsidizing down payments and repairs. Michigan?s
program is operated by two state agencies. One agency

provides cash assistance directly to welfare recipients and the other
develops agreements with local providers, called Michigan Works! Agencies,
for employment and training services.

Supplantation in the Since 1998, Michigan has used TANF funds to replace
about $126 million of

State Budget state funds in a variety of state programs. For example,
according to a legislative fiscal analysis, in 1998 the state shifted about
$34 million in general funds out of about 10 program accounts and replaced
these funds

with federal TANF funds, and in state fiscal year 1999 the state found about
$46 million in seven additional programs. In the single largest example of
supplantation, the state began using $27 million in TANF funds in state
fiscal year 2000 to finance a low- income homestead tax credit that had been
financed with state funds since 1973, and the state funds were used to
finance other priorities.

State budget officials said that even though the state has enjoyed healthy
finances, it has been very difficult for the state agency administering FIP
to get state funds beyond what is needed to cover the MOE requirement. As a
result, a number of new welfare- related initiatives are funded with federal
dollars. They added that the state?s policy is to maximize the use of
federal funds whenever possible. Given this budget environment, when
Michigan

has to come up with additional state funds for the state agency, a senior
budget official said they have to find ways to come up with general funds
savings in order to offset the need for additional state resources. In state
fiscal year 2000, Michigan had to finance a $160.7 million supplemental
appropriations bill for the state agency administering FIP. About $80.6
million of this total were state funds which were needed to pay federal
fines relating to its child support enforcement program, to increase food
stamp compliance, and to cover certain child welfare activities funded in
the prior fiscal year with TANF funds but not allowed under the final TANF
rules. State officials said that there was little political appetite in
Michigan to begin new programs that would require outlays of state funds.
The supplemental bill requiring additional state funds helped set the stage

for the state fiscal year 2001 budget debate. Since a provision in TANF law
limits what a state can count toward their MOE requirement, Michigan claimed
only a portion of the homestead credit to count towards its MOE

requirement. This provision, called the ?new spending test,? stipulates that
when states use expenditures on programs that were not part of the old AFDC
programs to count towards its MOE requirements, they can only use those
expenditures that represent an increase over what the state spent on the
programs in the base year of 1995. State budget officials argue that this
penalizes Michigan compared to other states for having a low- income tax

credit before TANF was enacted; a state with no tax credit in existence
could enact one using only TANF funds. As a result, these officials
maintained that it was reasonable to expect Michigan to refinance its tax
credit with TANF funds, putting the state on par with other- less generous-
states.

Spending Trends From 1995 to 2000

Figure 9: Michigan?s TANF- Related Expenditures for State Fiscal Years 1995
and 2000

1995 dollars in millions 10, 000

9, 000 8, 000 7, 000 6, 000 5, 000 4, 000 3, 000 2, 000 1, 000

0 95 00

95 00 95 00 95 00 Basic welfare Support for Support for TANF- related
services

working poor working poor +

social services health care

Federal State

To assess how state budget decisions might have affected spending, we
collected data on how Michigan used their federal and state funds to finance
TANF- related social service expenditures. The results are grouped

into four categories, as shown in figure 9 and described further in appendix
I. Each category is identified on the graph by two bars- one for spending
that occurred in state fiscal year 1995 and one for spending that occurred
in state fiscal year 2000. From left to right, the categories are
cumulative. Expenditures for cash assistance and employment training are
included in expenditures for support of the working poor. These expenditures
in turn

are included in expenditures for support of the working poor plus health
care. Finally, all expenditures are included in social services
expenditures.

For cash assistance and job training, total spending has fallen from $1. 6
billion to $1. 0 billion. (See table 21.) Adjusted for inflation, this
represents a decline of 42 percent. State spending 2 also declined from

$835 million to $527 million in nominal dollars, for a drop of 42 percent in
real terms. As a result, the state share of spending has remained relatively
constant at about 51 percent since state fiscal year 1995.

Table 21: Michigan?s TANF- Related Expenditures

Dollars in millions

Real Nominal (expressed in 1995 dollars)

1995 2000 1995 2000

Basic welfare services Total funds $1, 621 $1,031 $1, 621 $941 State funds
$835 $527 $835 $481 Support for the working poor Total funds $3,401 $3, 562
$3,401 $3, 250 State funds $1, 711 $1,778 $1, 711 $1, 622 Source: GAO survey
and analysis of state expenditure data.

When spending on a broader array of programs that provide support for low-
income families and the working poor is examined, the decline in total
spending is less pronounced. In this broader area, combined federal- state
spending remained relatively stable. (See table 21.) In real terms, this
represents a decrease of 4 percent, but considering that cash assistance
caseloads have fallen by nearly 66 percent, it suggests a strong

commitment of resources to these programs. Notably, total expenditures
increased by nearly $500 million in child care, $131 million in employment
support and job training programs, and $67 million in housing assistance.
Michigan?s share remained stable as well, with the state continuing to
contribute about 50 percent of the costs of these activities.

Spending on health care programs for this population, combined with support
for the working poor, shows little change in either total spending 2 In
referring to state spending in this section, we include county generated
funds that are used to count towards state match and MOE requirement.

or in the composition of spending- state funds vs. federal funds (see table
22). The state continues to finance about 48 percent of these programs.

Table 22: Michigan?s TANF- Related Expenditures

Dollars in millions

Nominal Real 1995 2000 1995 2000

Support for the working poor plus health care Total funds $5,973 $6, 480
$5,973 $5, 912 State funds $2, 853 $3,095 $2, 853 $2, 824 TANF- related
social services Total funds $7,783 $9, 633 $7,783 $8, 790 State funds $4,
006 $5,066 $4, 006 $4, 623 Source: GAO survey and analysis of state
expenditure data.

Our survey also captured spending on a number of other social service
programs that are geared towards meeting the needs of Michigan?s lowincome
families and the working poor. These programs include child welfare, mental
health, and substance abuse programs- all programs that can benefit from
TANF funding. Real spending increases are evident when

these costs are included in the totals above. Total spending increased from
$7. 8 billion to $9. 6 billion in nominal dollars- an increase of 13 percent
in real terms from state fiscal years 1995 to 2000. Total spending on social
services such as developmental disabilities, mental health, and substance

abuse alone has increased by over $1 billion, whereas child welfare and
juvenile justice increased by over $200 million. For all social services
combined, state spending outpaced the federal spending. As a result, the
state now pays a slightly higher share of the costs of these programs- about
53 percent in 2000 compared with 52 percent in 1995.

Additional federal safety- net programs provide food assistance and
supplemental security income for needy families with children in Michigan,
but are not shown in figure 9. The federal government provided $807 million
in food stamp benefits to low income individuals in Michigan in 1995, and
$338 million in 2000. At the same time, federal supplemental security income
payments increased from $878 million to $997 million in nominal dollars-
about 89 percent of these benefits went to individuals under 65 years of
age.

Rainy Day Funds Michigan does not have a TANF- specific rainy day fund.
However, the state does have a large general fund reserve- the Budget
Stabilization Fund. As of September 30, 2000, the fund had a balance of $1.3
billion. This fund is available to finance the state?s spending needs during
an economic downturn when state revenues fall short of expectations,
including any shortfalls in the state?s TANF program. State officials said
that during the recession in the early 1990s, most state departments were
required to cut back spending. The Budget Stabilization Fund was used to
minimize the degree to which cuts were needed. The state welfare agency?s
budget returned to prerecession baseline levels as soon as the fiscal crisis
was over.

Officials told us the state?s decision to forgo a dedicated welfare reserve
was influenced largely by its wish to avoid leaving any TANF funds unspent
at the U. S. Treasury. In state fiscal year 2000- 01 the governor proposed
leaving nearly $85 million unbudgeted- although this is not strictly a rainy
day fund, it would have been a source of funds the state could have used in
a fiscal crisis. During the budget process, the state House of
Representatives recommended establishing a $25 million rainy day fund

reserve and increased spending on other TANF- related services to use the
remaining $60 million the governor had left unbudgeted. However, the state
Senate did not go along with that provision. A senior analyst with the
state?s Senate Fiscal Agency said that concerns that unspent TANF funds
might signal to the U. S. Congress that these funds were not needed prompted
the state Senate to commit virtually all of the available TANF funding. The
supplemental appropriations bill was also designed, in part, to reduce
Michigan?s TANF surplus remaining in the U. S. Treasury- again

because these balances had attained a heightened visibility and raised
concerns that they might be rescinded by the U. S. Congress. Michigan
administers its welfare program directly through its state offices and does
not rely extensively on subgrantees or contractors to provide services.
Those eligible for specific services generally deal directly with

the Family Independence Agencies Offices around the state. As a result, the
state does not record significant levels of TANF obligations that remain
unliquidated for extensive periods. Table 23 shows the TANF funds that
Michigan reported as unspent as of September 30, 2000.

Table 23: Unspent TANF Funds in Michigan as of September 30, 2000

Dollars in millions

Unliquidated Unspent funds as a

Unobligated Obligations of

Total unspent percent of annual

TANF funds TANF funds funds

TANF grant

$200. 0 $14. 1 $214. 1 28% Source: HHS? Administration for Children and
Families.

Even though Michigan reports considerable levels of unobligated balances,
this does not mean either that the state has no plans to use those funds or
that these funds have been set aside in a rainy day fund. In fact, in state
fiscal year 2000- 01, the state appropriated almost its entire block grant
and all TANF funds left unspent from earlier fiscal years. The final
enrolled bill left only $500, 000 unbudgeted out of more than $844 million
in TANF funds available. The budget contains nearly $760 million to finance
ongoing expenditures in the states? welfare program and about $85 million to

finance new projects such as a child care rate increase and teen parent
counseling programs. Although these programs have been enacted and funded
through the state?s appropriation process, the majority of the funds used
will remain unobligated until the programs are established and incurring
expenses. This is because the state will operate these programs directly and
so cannot record an obligation of funds until they are actually expended.

State budget officials said that Michigan should be able to deal with
increased caseloads, even within the existing state agency budget. Many of
the new programs funded from the supplemental were one time initiatives,
which adds flexibility to the state budget process, because if resources
become tight in other programmatic areas the funds used to finance these one
time initiatives could be used elsewhere. The officials added that, even
though the budgets essentially authorized the expenditure of the entire TANF
grant, the state would continue to carry forward some unused TANF

funds which could be canceled and re- appropriated to more pressing needs at
the state agency that operates FIP. In addition, budget analysts from both
the legislative and executive branch

in Michigan suggested that the state typically overestimates the resources
needed for many of these new programs. It is unclear how many people will
enroll in the new programs or even how much they will cost. Historically,
analysts agree the state has not spent all TANF funds it has appropriated

and they expected it is likely with these new programs as well so there will
continue to be some level of unused TANF funds, even with the new
expansions.

Appendi x VI II

New York Background

State fiscal year: April 1- March 31 Budget cycle Annual TANF grant: a $2.
443 billion 75% MOE: $1. 719 billion Cash assistance caseload: b 461,000
families in 1995

248,000 families in 2000 46 percent decline a New York received an
additional $8 million in federal TANF funds in 2000 as a performance bonus.
b Caseloads as reported to HHS for January 1995 and June 2000 and rounded to
nearest thousand.

New York launched its welfare reform program in August 1997 with the passage
of state legislation creating the Family Assistance (FA) program. The goal
of the program is to promote self- sufficiency by providing temporary and
transitional assistance for needy families, along with services designed to
increase employment and job retention. When families apply for FA, the
adults are assessed to determine their job

readiness. Those considered job ready are encouraged to seek work
immediately. Under the state?s policy, families may receive cash assistance
for up to 24 months before adults must obtain employment or participate in
assigned activities, but localities may establish shorter time frames for
job

ready adults. The amount of cash assistance families can receive varies by
location and family size. For example, in one downstate county a family of
three may receive a maximum of $703 while in an upstate county the

maximum is $499. These payment levels have not changed since the enactment
of TANF. Families whose adjusted incomes are below cash assistance levels
are eligible for FA. In addition to cash assistance, a family

may receive up to $50 a month in child support collected on its behalf. 1
Families who receive monthly cash assistance payments for the equivalent of
5 years are not eligible for receipt of TANF cash assistance, in accordance
with the TANF law. Under a separate safety- net program supported by state
and local funds, families who reach the 5- year time limit may continue to
receive assistance, 1 The remainder is retained by the state and federal
governments as reimbursement for cash assistance payments made to these
families.

but the assistance is not paid in cash to the adults in these families.
Instead, the assistance takes the form of payments made directly to
landlords and utility companies to cover costs such as housing and heating.
Recent immigrants who are ineligible for TANF cash assistance under TANF
law, children living apart from adult relatives, and other needy individuals
may also be eligible for the safety- net program. They may receive direct
cash payments each month for up to 2 years, and

subsequently, payments may be made to landlords and utility companies on
their behalf, where deemed appropriate.

Provision of certain services to help families overcome barriers to
selfsufficiency are mandated by the state, while other services vary in
accordance with programs developed at the local level. The state requires

that families be provided information on, and screening for domestic
violence and that adults be screened and, if needed, treated for substance
abuse. Under the state?s policy, family members may look for work for more
than 6 weeks and may pursue 2 years of postsecondary education, although
their participation in these activities does not count toward the work
participation rates that states must achieve under TANF. For working

families receiving cash assistance, subsidized child care is guaranteed, and
it may be provided for other families with incomes up to 200 percent of
poverty, depending on availability of funds and local agency policy.
Families who find work and no longer receive monthly cash payments may
continue to receive 12 months of transitional medical assistance. They may

also be eligible for state and federal earned income tax credits or refunds
and state child care tax credits or refunds.

In addition to supporting the FA program, TANF funds are used to provide
emergency, housing, and transportation assistance. They are also used for
pregnancy prevention and family formation programs, in line with the new
federal welfare goals, and for child welfare activities.

Although the state oversees FA, it has devolved responsibility for program
design and operation to local social service agencies in the state?s 57
counties and New York City. The state supplements the localities? basic FA
programs with an appropriation funded entirely with TANF funds to provide
other supportive services to help families avoid or ease their

transition from public assistance. The localities must submit annual
spending plans for social services for approval by the state agency that
oversees these activities. To carry out their programs, the localities
receive block grants from several state agencies to cover childcare, child
welfare, employment and related services for TANF- eligible family members,
and

transitional services for families with incomes at or below 200 percent of
poverty. Local agencies continue to be required to provide their share of
funds to match state funds for basic cash assistance provided through the FA
program in order to help meet the state?s MOE level.

Supplantation in the In 1997, we reported that New York used federal TANF
funds to provide

State Budget over $344 million in fiscal relief for the state and its
counties. While it is possible for all states to replace state funds with
federal funds once they meet their MOE requirement, in New York the level of
fiscal relief possible depends on cost- sharing arrangements that were in
place under the AFDC program and that continue under TANF. In New York,
social services are

provided by local government social service offices. These local offices
provide services and then submit claims for reimbursement of those services
to the state. Under the AFDC cost- sharing arrangements, the federal
government paid 50 percent of the AFDC costs and the state and local
governments shared the remaining costs equally. The state also had a general
assistance program that provided support to individuals who were not
eligible for the AFDC program, such as those who had no dependent children.
The state and local governments also shared these costs. This

arrangement continues under TANF; both the state and localities contribute
equal shares towards the MOE requirement. Theoretically, once the MOE is met
neither the state nor local governments need to make any further
contributions and can rely on unspent TANF funds to finance the program for
the rest of the year. However, caseloads have dropped, and the

state has struggled to come up with enough expenditures to meet its MOE.
Senior state budget officials said that this aspect of fiscal relief has
been overemphasized in the budget debate and, in many respects, did not
materialize as projected. In calculating the level of fiscal relief the new

financing arrangement might bring, the budgets assumed relatively stable
spending levels. That is, if the counties spent the same total amount, they
could be reimbursed for a higher share of the total claim. New programs and
services that could be offered to the ever- dwindling caseload were slow to
start for a variety of reasons, which affected the new program?s outlay
rates. As a result, total expenditures also fell- in line with a falling
caseload- and the state found itself falling well short of the expenditures
it would need to meet its MOE.

The pressing fiscal issue in New York became one of how to meet the MOE
under its cost- sharing arrangement. The state realized it would not meet
its MOE requirement if expenditure rates continued at the current rate and

it continued to reimburse counties with federal and state funds. Since the
state operates on a cost- reimbursement basis with its localities, if the
state were to have temporarily suspended the federal reimbursements, the
effect would have been to pay for all costs with state and local funds
making it easier to reach the state?s MOE. To resolve the potential MOE
shortfall, the state halted any reimbursement to the localities for
expenditures they

made in the last quarter of the fiscal year. In so doing, the state counted
all of the local expenditures toward the state?s MOE, but provided little in
the way of fiscal relief to the counties. The state now monitors closely
total expenditures and adjusts the federal share of the reimbursement to
ensure that the state will meet its MOE each year. While fiscal relief may
not have materialized as a result of these costsharing arrangements, the
state has been able to achieve significant fiscal relief in other ways. For
example, as permitted by federal law, the state

transfers TANF funds to the Social Services Block Grant (SSBG). State
officials explain that, in the past, most of the SSBG funds were used to
augment federal, state, and local funding for child welfare. In transferring
TANF funds to SSBG for use in the child welfare system, the state has
reduced the nonfederal share of these program costs considerably. Federal
funds represented nearly 54 percent of child welfare program expenditures

in state fiscal year 2000- 2001 whereas the federal funds supported 49
percent of spending on these programs in 1995. The state reports spending
about $1. 5 billion in total on child welfare programs in 2000. By shifting
about 5 percent of those costs to the federal government, it realized
savings of approximately $72 million.

More recently, however, the state has changed its approach to financing
welfare programs. The state will maintain its traditional structure of
reimbursing counties for basic welfare services and counting both the state
and local share towards its MOE requirement. Beginning in state fiscal year
1997- 1998, it created a number of new grants- funded exclusively with TANF
funds- which provide resources to local social service agencies that they
can use to develop new programs to serve the needs of both those still on
the welfare rolls and those transitioning off the rolls. Under this new
financing arrangement, the state and local share of total TANF

program expenditures will decrease as the state maximizes use of federal
funds. Because of fears that counties will take advantage of the flexibility
of the TANF block grant and reduce their own commitment to some social
service programs, the state prohibits counties from using any of the TANF

funds to ?supplant? their own spending on social services for needy
families.

Spending Trends From 1995 to 2000

Figure 10: New York?s TANF- Related Expenditures for State Fiscal Years 1995
and 2000 1995 dollars in millions 30, 000

25, 000 20, 000 15, 000 10, 000

5, 000 0

95 00 95 00 95 00 95 00 Basic welfare Support for

Support for TANF- related

services working poor

working poor + social services

health care

Federal State

To assess how state budget decisions might have affected spending, we
collected data on how New York used state and federal funds to finance TANF-
related social services. The results are grouped into four categories,

as shown in figure 10 and described further in appendix I. Each category is
identified on the graph by two bars-- one for spending that occurred in
state fiscal year 1995 and one for spending that occurred in state fiscal
year 2000. From left to right, the categories are cumulative. Expenditures
for cash assistance and employment training are included in expenditures for
support of the working poor. These expenditures in turn are included in
expenditures for support of the working poor plus health care. Finally, all
expenditures are included in social services expenditures.

Tabl e 24: New York?s TANF- Related Expenditures

Dollars in millions

Real Nominal (expressed in 1995 dollars) 1995 2000 1995 2000

Basic welfare services Total funds $5,188 $3, 590 $5,188 $3, 276 State funds
$3, 483 $2,534 $3, 483 $2, 312 Support for the working poor Total funds
$8,148 $6, 659 $8,148 $6, 076 State funds $5, 380 $4,696 $5, 380 $4, 285
Source: GAO survey and analysis of state expenditure data.

In New York, total expenditures in every category declined between state
fiscal year 1995 and 2000. Spending on cash assistance, both to TANFeligible
families and to those receiving safety- net assistance and enrolled in job
training programs declined from $5. 2 billion to $3.6 billion in nominal
dollars. (See table 24.) Adjusted for inflation, spending on cash assistance

and job training has declined 37 percent- in line with the drop in TANF
caseloads. Although the state now finances a greater share of total
expenditures, 2 this may be largely due to the struggles the state faced in
meeting its MOE requirements. State officials concede that New York was

slow to begin spending on new programs and, therefore, had to spend state
funds first.

There has also been a decline in both federal and state funds spent on
programs supporting the working poor which include child care, employment
services, and other benefits as well as the cash assistance and job training
described above. In state fiscal year 1995, the amount spent was $8.1
billion; by 2000, the amount was $6.7 billion in nominal dollars.

This represents a decline of 25 percent in real dollars. Although less is
being spent overall, the state?s share of expenditures in state fiscal year
2000 is 5 percent higher than it was in state fiscal year 1995. (See table
24.) 2 In referring to state spending in this section, we include county-
generated funds that are used to count towards state match and MOE
requirements.

We also tracked health care expenditures for our target population. (See
table 25.) Including health care spending in our survey data does not change
the picture. Total spending dropped 2 percent when adjusted for inflation.
But the federal share of the costs of these programs increased by

10 percent in real terms while state share decreased by 9 percent. As a
result, the fiscal balance across this range of social services shifted
significantly from 1995 to 2000. The state financed 61 percent of the costs
of all these activities in 1995; in 2000 it financed 56 percent.

Tabl e 25: New York?s TANF- Related Expenditures

Dollars in millions

Real Nominal (expressed in 1995 dollars) 1995 2000 1995 2000

Support for the working poor plus health care Total funds $21, 285 $22,910
$21, 285 $20, 904 State funds $12, 955 $12,893 $12, 955 $11, 794 TANF-
related social services Total funds $26, 133 $27,635 $26, 133 $25, 216 State
funds $16, 405 $16,045 $16, 405 $14, 640 Source: GAO survey and analysis of
state expenditure data.

The same trend holds when we broaden our survey to include other social
services that are eligible for TANF funding- child welfare services, mental
health programs, and substance abuse programs. Total spending declined by
about 4 percent in real terms; state spending declined by about 11 percent.
The state share decreased from 63 percent in 1995 to 58 percent of program
costs in 2000. Additional federal safety- net programs provide food
assistance and supplemental security income for needy families with children
in New York, but are not shown in figure 10. The federal government provided
$2 billion in food stamp benefits to low income individuals in New York in
1995, and $1.4 billion in 2000. In addition, from 1995 to 2000, federal
supplemental security income payments increased from $2.1 billion to $2. 7
billion in nominal dollars- about 73 percent of these benefits went to

individuals under 65 years of age.

Rainy Day Funds For its welfare reform program, the state has a dedicated
reserve account called the New York Works compliance fund. The funds in this
account are

federal TANF dollars that the state has not budgeted for other TANFrelated
purposes. These funds remain in the U. S. Treasury until the state needs
them. Budget officials estimate that the balance in this fund was about $662
million as of September 30, 2000- the end of the last federal fiscal year.
Under the state statute creating this fund, amounts are determined based on
residual funds; it specifies that after making all authorized payments, any
unspent TANF funds may be deposited to the

compliance fund. Agencies that need additional funds can gain access to
these funds with the approval of the state budget director. No legislative
branch approval is required. The statute does not specify what criteria the
state budget officer should use in deciding whether to grant an agency
access to the fund.

In accordance with federal requirements, the TANF balances in the New York
Works compliance fund are held in reserve at the U. S. Treasury. Federal
regulations require states to report any TANF funds held for a rainy

day as unobligated balances. (See table 26.) However, a significant share of
what the state reports as unobligated is not part of the state?s rainy day
fund because even though these funds are appropriated they are not obligated
until the county has submitted a claim for reimbursement. Under federal
regulations, the state must also report funds it will use to reimburse the
counties for basic welfare services as unobligated until the state

actually reimburses the counties for these expenditures. This is
significantly different from the approach used in other county- administered
welfare systems. (See appendix II for information on California and

appendix III for information on Colorado.) In these states, funds are
obligated when the state notifies the counties of their allocations. While
even in these states federal regulations require that TANF funds remain at
the U. S. Treasury until the counties need them for immediate expenditure,
unspent TANF funds are recorded as an unliquidated obligation rather than
simply an unobligated balance.

Table 26: Unspent TANF Funds in New York as of September 30, 2000

Dollars in millions

Unspent funds as a Unobligated

Unliquidated Tot al unspent

percent of annual balance obligations funds

TANF grant

$761 $547 $1, 308 53% Source: HHS? Administration for Children and Families.

New York also reports a significant level of unliquidated obligations. Once
the state determines the annual allocation levels for the new block grants
to the counties, it can enter into an obligation for those funds. As in
other states with county or contractor- administered welfare systems, this
is a way in which the state reduces the level of unobligated TANF balances.

State officials said that, in addition to ensuring that new program spending
was predominately financed with federal funds, they were concerned that if
they continued to report large unobligated balances, Congress might perceive
that those funds were not needed by the state.

Appendi x IX

Oregon Background

State fiscal year: July 1- June 30 Budget cycle: Biennial TANF grant: $168
million 75% state MOE: $92 million

Cash assistance caseload: a 40, 000 families in 1995 17, 000 families in
2000 58 percent decline a Caseloads as reported to HHS for January 1995 and
June 2000 and rounded to the nearest thousand.

Oregon was one of the earliest states to begin welfare reform, starting in
1992 and going statewide with ?Oregon Option? in July 1996 under federal
waivers that remain in effect through June 2003. The program takes a

?work- first? approach based on the state?s premise that work is better than
welfare- which is temporary- and that all welfare recipients are capable of,
and responsible for, making progress toward economic independence and family
stability. To receive monthly cash assistance under the Oregon Option,
adults must participate in work activities immediately. The maximum monthly

assistance is $460 per month for a family of three, a level unchanged since
1991. Needy families with children whose adjusted incomes are below the
monthly cash assistance levels are eligible for the program. The list of
work activities under the state?s waiver is broader than under TANF, and
includes unlimited job search, education, and mandatory substance abuse and
mental heath treatment as appropriate. Subsidized work is available, so some
participants placed in temporary private sector jobs may receive subsidized
wages instead of cash assistance and earn money in an individual account for
their future education. Adults may receive cash assistance for no more than
24 out of 84 months, but under the state?s waiver the time they spend in
work activities does not count toward the 24- month limit. Families in
crisis may receive diversion payments or services such as child care,
transportation, or work tools in lieu of monthly

cash assistance. Program participants who find work and go off cash
assistance may receive subsidized child care, called Employment Related Day
Care (ERDC), and medical services under the Oregon Health Plan if they meet
the eligibility criteria. Both were initiated prior to TANF and serve the
state?s low

income working poor. ERDC is available for families earning less than 185
percent of poverty. Income limits are higher for ERDC than for the Option
Program, and families participating in the state?s child care program are
expected to contribute to the costs on a sliding scale, based on their
incomes. The Oregon Health Plan is a demonstration program operating under a
federal Medicaid waiver from 1994 through 2002. It is available to
individuals with incomes up to 100 percent of the federal poverty level and
to children in families with incomes up to 133 or 170 percent of poverty.
The state human services agency administers the Oregon Option program
through its branch offices and prime contractors in 15 districts. The

branch offices determine eligibility, arrange for electronic payments of
cash assistance, assess families? needs, and refer families to local service
providers. The prime contractors, such as community colleges and job
training partnership agencies, provide training and job services directly
and subcontract with local partner agencies for other services as needed. In
2001 the state human services agency reorganized to provide more integrated
management and services at both the state and local levels. In addition to
supporting the Oregon Option program, TANF funds

continue to be used for child welfare activities, such as family
preservation, emergency care for children removed from their homes, and
special foster care rates. In line with the new federal welfare goals, TANF
funds are also

used for pregnancy prevention and family formation programs statewide.
Supplantation in the

TANF funds used for the Oregon Option program and related activities State
Budget appear to offer a greater variety of benefits and services to a
greater number of families than might have been possible under AFDC. But the

reduced cash assistance caseload also means that fewer TANF funds are needed
to pay those benefits, which frees up more TANF funds that can be allocated
to other state priorities. Since 1997, Oregon has used about

$122 million, or 18 percent, of its total TANF funds, to replace state funds
in a variety of programs that serve low- income individuals and families.
According to state officials, Oregon has increased state funding in other
budget priorities serving these same beneficiaries.

State officials said that Oregon began its welfare reforms early and
invested more state funds in its employment programs and ERDC than required
to obtain the maximum federal matching grants. According to these officials

when the state?s investment paid off and families started to become more
self- sufficient and left the caseloads, less funding was needed to support
the program. For example, in April 1999, the Governor?s recommended budget
was adjusted to take into account lower utilization rates projected for the
state?s EDRC child care program. 1 As a result of the lower

projections, the state found that it had ?over budgeted? the ERDC program by
nearly $40 million. Oregon found that it could use the TANF funds to replace
state general fund dollars in several other programs. In turn, the state
general funds that were freed up helped support a $400 million

increase in Oregon?s support for K- 12 education programs. State officials
emphasize that Oregon is able to meet its MOE requirement, despite shifts in
funding, because the state continues to spend generously on services for
needy families and individuals. Furthermore, even though K- 12 expenditures
are generally not allowable MOE expenditures, these officials argue that by
shifting some state funds away from the Oregon Option program and into
education they are investing in a prevention strategy that will help reduce
poverty in the long run by ensuring that the next generation has the skills
needed to enter the job market. State officials said they have maintained
their commitment to the state?s

poor by continuing to finance the Oregon Option Program that has proven very
successful at reducing dependency on public assistance. Also they have
maintained their commitment to improved access to health care through the
Oregon Health Plan that serves many who would otherwise be ineligible, such
as families not meeting the Medicaid income test, childless couples, and the
elderly. State officials also suggest that in freeing up state funds, Oregon
was able to fund a variety of new programs and enhancements that would not
otherwise have been funded. Examples include a $14.3 million program to
provide services for high- risk youth, an $8 million expansion to the
state?s

alcohol and drug treatment program, and an $8 million expansion to a 1 Funds
for Oregon Option and other state programs and services are appropriated by
the state legislature every 2 years when the legislature meets and acts on
its biennial budget. When the legislature is not in session, adjustments to
the authorized spending levels can be made by the legislature?s Emergency
Board- this process is called rebalancing.

variety of early childhood development programs. The officials acknowledged
that because money is fungible, it is impossible to say precisely which
programs were funded with the freed up funds, but they emphasize that new
state spending would have been very difficult in the

absence of budgetary tradeoffs that reduced state funding of the Oregon
Option program.

Welfare advocates agree that Oregon invested state funds early and on its
own initiative to support the Oregon Option program, the Oregon Health Plan,
and subsidized child care. They disagree, however, about the current level
of funding needed to continue the programs. They believe that continued
reliance on caseload reduction to measure success in the program is
misplaced. They note that although the state?s poverty rate is about the
same as it was in 1995, the state?s population is growing, so the

number of families eligible for cash benefits must be increasing. The
advocates believe the state should focus its efforts on reaching these
families that are eligible but are not receiving assistance. Further, they
note that many of the families who remain in the Oregon Option program are
among the hardest to serve and need more services than are currently
provided.

Spending Trends From 1995 to 2000

Figure 11: Oregon?s TANF- Related Expenditures for State Fiscal Years 1995
and 2000

1995 dollars in millions 12, 000

10, 000 8, 000 6, 000 4, 000 2, 000

0 95 00 95 00 95 00

95 00 Basic welfare Support for Support for TANF- related services

working poor working poor +

social services health care

Federal State

To assess how state budget decisions might have affected spending, we
collected data on how Oregon used state and federal funds to finance TANF-
related social services. The results are grouped into four categories, as
shown in figure 11 in line with a methodology described further in appendix
I. Each category is identified on the graph by two bars- one for

spending that occurred in state fiscal year 1995 and one for spending that
occurred in state fiscal year 2000. From left to right, the categories are
cumulative. Expenditures for cash assistance and employment training are
included in expenditures for support of the working poor. These expenditures
in turn are included in expenditures for support of the

working poor plus health care. Finally, all expenditures are included in
social services expenditures.

Our review of state expenditures shows a marked decline in total spending on
basic welfare services. Total spending on these programs declined about 45
percent in real terms since 1995. (See table 27.) But the state realized
greater savings overall; real state spending declined by 82 percent while
federal spending on these programs declined 27 percent in real terms. In
2000, Oregon financed 10 percent of these programs? costs,

compared to the 32- percent share the state financed before welfare reforms.
Table 27: Oregon?s TANF- Related Expenditures

Dollars in millions

Real Nominal (expressed in 1995 dollars)

1995 2000 1995 2000

Basic welfare services Total funds $255 $155 $255 $141 State funds $81 $16
$81 $15 Support for the working poor Total funds $446 $430 $446 $392 State
funds $130 $97 $130 $88 Source: GAO survey and analysis of the state
expenditure data.

While the cash assistance caseload has declined, programs that provide
transitional welfare services- such as case management and the ERDC program-
have experienced caseload growth, according to state documents. Our
expenditure survey data confirm this observation. Although total spending
declined dramatically (45 percent) in traditional welfare programs, spending
on programs that broadly support the working poor declined by about 12
percent in real terms. However, the state continued to realize substantially
more real savings than the federal

government. As the state continued to shift its resources out of these
programs, it has created a new fiscal balance. Before welfare reform, the
state financed about 29 percent of the costs of programs broadly

supporting the working poor and used federal funds to finance about 71
percent. In state fiscal year 2000, it financed 22 percent of the costs and
uses federal funds to finance the rest. State officials claim that the
successes Oregon has achieved in its Oregon Options program have permitted
it to transfer savings realized from a

smaller caseload to other programs, such as the Oregon Health Plan. Indeed,
looking at expenditures on programs that support the working poor as well as
health care costs related to these same individuals and

families shows a dramatic increase in spending. Total spending rose 72
percent in real terms from 1995 through 2000. (See table 28.) Real state
spending rose about 81 percent, slightly more than federal spending, which
increased 67 percent. State spending rose more than federal because, under
Oregon?s Medicaid waiver, the state is covering individuals who are not
categorically eligible for Medicaid. As a result of this shift, the state
now finances a larger share of the costs associated with programs supporting
the working poor and their related health care costs; the state share was
about 34 percent in 1995 and 36 percent in 2000.

Table 28: Oregon?s TANF- Related Expenditures

Dollars in millions

Real Nominal (expressed in 1995 dollars)

1995 2000 1995 2000

Support for the working poor plus health care Total funds $983 $1,849 $983
$1, 687 State funds $334 $664 $334 $605 TANF- related Social Services Total
funds $1,303 $2, 304 $1,303 $2, 103 State funds $502 $909 $502 $829 Source:
GAO survey and analysis of state expenditure data.

Our survey also captured spending on a number of other social service
programs that are geared towards meeting the needs of states? low- income
families and the working poor. These programs include child welfare, mental
health, and substance abuse programs. When considered along with other
programs supporting the working poor, we see that spending on all these
social service categories has risen about 61 percent since 1995. Driven
largely by the expansions in the Oregon Health Plan, state spending on
social services rose 65 percent and the federal spending rose 59 percent.
Remarkably, despite the significant shifts in the way the state finances
these programs, the federal- state fiscal balance has remained stable since

1995. The state continues to finance about 39 percent of these programs and
the federal government contributes about 61 percent.

Additional federal safety- net programs provide food assistance and
supplemental security income for needy families with children in Oregon, but
are not shown in figure 11. The federal government provided

$250 million in food stamp benefits to low- income individuals in Oregon in
1995, and $174 million in 2000. In addition, from 1995 through 2000, federal
supplemental security income payments increased from $181 million to

$235 million in nominal dollars- about 85 percent of these benefits went to
individuals under 65 years of age. Rainy Day Funds Senior budget officials
in Oregon said that the state?s policy is to spend all

of its annual TANF grant, saving none of the federal moneys for a rainy day.
Generally, state officials believe that cash assistance caseloads can be
kept low if sufficient funds are invested in appropriate services to achieve
families? long- term self- sufficiency. They are committed to helping
recipients find work, retain their jobs, and obtain the skills necessary to
ensure their resiliency, even in a changing economy.

The state?s policy is to draw down all available funds, leaving nothing in
reserve. Generally appropriation levels are set so that the full TANF grant
will be spent. As with most states with biennial budgets, the ?rebalancing

process? in Oregon allows for the shifting of resources among program
accounts outside the regular appropriations cycle. This allows the state to
periodically reassess program needs even when the legislature is not in
session. Oregon requires its agencies to regularly reassess its budgetary
needs- or rebalance its budget- allowing for the transfer of funds into or
out of program accounts. Each agency is required to submit its rebalance

plan every 6 months. Decisions made during Oregon?s budget process highlight
some of the factors considered and the procedures involved in aligning
appropriations, obligations, and expenditures. In March 2000, the state
agency again prepared for its rebalancing process and found that the ERDC
child care utilization rate was lower than expected, which would have
resulted in about $8. 2 million in unspent TANF dollars. To absorb these
funds, the agency reduced the copayments that poor families had to pay and

increased the subsidies, thus increasing the amount of TANF funds expended
per child. In this way, the agency accomplished the state?s goal of spending
all the TANF funds available, speeding the rate at which these TANF funds
are spent and avoiding having to categorize them as unobligated balances.

Despite its lack of any contingency plans, Oregon reports that more than $40
million of cumulative TANF funds designated for the state as of September
30, 2000, remain unspent at the U. S. Treasury. Most of these unspent funds
are shown as obligated but not liquidated because they have been committed
through contracts or other agreements, or they are owed

because services have been provided, although payments have not been made
against these obligations. The remainder of Oregon?s unspent funds are shown
as unobligated from a single year- 1999. (See table 29.) Table 29: Unspent
TANF Funds in Oregon as of September 30, 2000

Dollars in millions

Unliquidated Unspent funds as a

Unobligated obligations of

Total unspent percent of annual

TANF funds TANF funds funds

TANF grant

$4. 8 a $39. 1 $43.9 2. 6% a Subsequent amended reports submitted by Oregon
reduced the unobligated balance to zero.

Source: HHS?/ Administration for Children and Families.

TANF funds dedicated to the Oregon Option program that are not used by the
state agency or its branches are set aside for prime contractors in the 15
districts for specific training and employment services. These funds are
shown as unliquidated obligations because they are tied up in contracts. The
prime contractors in turn execute subcontracts for additional support
services. The system of contracts and subcontracts results in an inevitable
time lag before obligated funds are shown as liquidated. Some time lags

occur because subcontractors and contractors must submit claims, and the
claims must be processed before TANF payments can be made. Other time lags
occur because TANF funds are obligated in contracts with start and end dates
that do not coincide with the federal fiscal year, so that funds obligated
in one quarter may not be paid until the following quarter.

Appendi x X

Texas Background

State fiscal year: September 1- August 31 Budget cycle: Biennial TANF grant:
a $486 million 75% MOE: b $236 million Cash assistance caseload: c 280,000
families in 1995 128,000 families in 2000

54 percent decline a Texas receives between $13 million and $39 million in
annual TANF supplements because the state?s pre- TANF expenditures per poor
person were less than 35 percent of the national average. In addition, Texas
received $16 million in federal TANF funds in 2000 and $24 million in 2001
as bonuses for successful outcomes.

b Texas did not meet required work participation rates for two- parent
families in 1997 and 1998, but subsequently complied with corrective action
plans, thus avoiding any financial penalty or increased MOE. c Caseloads as
reported to HHS for January 1995 and June 2000, rounded to the nearest
thousand.

Texas implemented welfare reform between November 1996 and November 1997
with its employment and training program called CHOICES, an outgrowth of
experiments started earlier under a waiver covering 1996- 2002. CHOICES
follows the work- first model, with the

expectation that all Texans should support themselves and their families and
that they should take the initiative in finding work, while program staff
and services are available to assist them in achieving that goal. Under the
program, the number of families receiving cash assistance declined until
1999 and leveled off. The caseload has declined again over the first few

months of 2001.

To receive cash assistance, applicants must be willing to seek work as soon
as they enroll in the program. Under the Texas waiver, those with young
children have been exempt from TANF work requirements, but the exemptions
are gradually being phased out. In 1999, the state approved the first
increase in the monthly cash assistance level in 15 years; now a family of
three can receive up to $201 a month. Needy families with children whose
adjusted incomes are below the monthly cash assistance levels are eligible.
In addition, a family may receive up to $50 a month in child support
collected by the state agency on the family?s behalf. 1 Texas al so provides
an annual $60 supplement for each TANF- certified child. If they meet
poverty guidelines, grandparents 50 or older who care for TANFeligible

children may receive one- time cash supplements up to $1, 000, but if they
accept these payments they are ineligible for cash assistance for 1 year.
Families in crisis may receive a one- time diversion payment of $1, 000, but
if so, they are ineligible for cash assistance for a year.

CHOICES participants engaged in work or work- related activities may receive
services to facilitate their participation, such as transportation
assistance and help with other work expenses. Subsidized child care is
guaranteed for those in CHOICES who participate in work activities, and

where funding is available, to other families with incomes at or below 85
percent of the state median income. Program participants may receive cash
assistance for 1 to 3 years, depending on their education and prior work
experience, at which time the adults become ineligible for 5 years. Those
who lose cash assistance due to increased earnings or time limits may
receive transitional medical assistance and subsidized child care for 12 to
18 months. To reduce the likelihood that these children will have to depend
on welfare, Texas provides grants toward a college education for recent high
school graduates in families eligible or receiving cash assistance.

Two state agencies cooperate in the administration of the program- one
determines eligibility through its local offices and makes cash assistance
payments electronically and the other provides work- related services,
including child care subsides through contracts with 28 local workforce
boards. The goal of the boards is to provide integrated services at each
location that respond to local needs. The state requires that the local

1 The remainder is retained by the state and federal governments as
reimbursement for cash assistance payments made to these families.

workforce boards provide matching funds in order to receive the maximum in
federal child care funds. In addition to funding CHOICES, TANF funds support
certain child welfare activities that were supported previously with federal
welfare funds, such as child protective services. To prevent dependency and
deal with the

needs of hard- to- serve CHOICES participants, TANF funds are also invested
in family formation, pregnancy prevention, mental health, developmental
disabilities, and substance abuse treatment.

Supplantation in the Texas has used the flexibility afforded by TANF
financing rules to fund a

State Budget number of state priorities for needy families but has also used
these funds to replace a significant amount of state funds. From 1998 to
2001, Texas

budget analyses show that about $275 million of federal TANF funds were used
to replace state funds, including $115 million in the 1998- 1999 biennium
and $160 million in the 2000- 2001 biennium. Texas budget officials and
legislative analysts describe these shifts as changes in the way certain
programs are financed- not supplantation- and note that state spending on
social services has increased by more than this amount since federal welfare
reform was enacted. For example, they claim that while $160 million in state
funds was freed up as a result of shifts in TANF funds in the last biennium,
this ?refinancing? was part of a $400 million increase in state funding for
health and human services programs for the needy in the last 2 years. The
development of the 2000- 2001 budget in Texas illustrates the complex

relationship among federal policies regarding TANF funds and state decisions
on the use of state funds. Historically, cash assistance and job training
programs were funded with a combination of state and federal dollars. In
1995, federal funds represented just over 63 cents and state funds
represented almost 37 cents of every dollar spent on cash assistance in
Texas. In 1997, when Texas first deliberated how it would use the new TANF
funds in its welfare program, state officials believed that cash

assistance and training would be a good place to direct state resources to
count toward the state?s MOE requirement. However, as caseloads dropped more
quickly and steeply than anticipated, it became clear that the state funds
appropriated for these purposes would not be sufficient to meet the state?s
MOE requirement. A decision was made to fund job training primarily with
federal TANF dollars and to direct state funds to other priorities that had
projected outlay rates that would provide a more stable base to use as their
MOE.

Child welfare programs were seen as a priority because of growing foster
care caseloads and a state court ruling that called for heightened
investigation and prevention of child abuse and neglect. Because spending

on these programs was growing, these programs? expenditures were viewed as a
better source of MOE financing than CHOICES. The legislature increased
appropriations for an expanded child welfare program and directed the state
agency to count the state?s share of the increase towards it MOE. After
reviewing final TANF regulations, however, state agency

officials found that certain foster care expenditures included in the child
welfare expansion could not be counted toward the MOE requirement. The state
funds already appropriated for child welfare were not withdrawn, but state
budget officials had to look elsewhere for allowable MOE expenditures. In
the end, they identified as MOE a portion of state aid used

to support preschool programs. Since this state aid is allocated based on
the number of needy children in each district the state believes that a
portion of it qualifies as an MOE expenditure. The state spent nearly $320
million on this effort in 2000.

Texas state officials maintain that the changes that they have made in
program financing made it possible for the state to draw down more TANF
funds. In so doing, the state reduced the levels of unspent balances
remaining at the U. S. Treasury, but more importantly proponents of welfare
program expansions gained political leverage to get new programs and program
expansions enacted. A state official told us that the political environment
in Texas makes it very difficult to fund program increases or new programs
with state funds unless there are trade- offs in other

programs to offset some of the new spending. The official said that the
ability to refinance some existing programs with TANF funds in turn freedup
some state resources, which helped ensure enactment of new state spending in
other areas that also serve the low- income population.

While the advocate community agrees with state officials that the state has
expanded its investment in programs that serve the social services needs of
low- income families and individuals, they also note that once state funds
are freed up, it is impossible to tell with certainty how those funds are

ultimately used. Their concern is that funds designated for needy families
eligible under TANF may have been used for other purposes, thus diverting
state funds to other priorities.

Spending Trends From 1995 to 2000

Figure 12: Texas? TANF- Related Expenditures for State Fiscal Years 1995 and
2000 1995 dollars in millions 2, 500 2, 000 1, 500 1, 000

500 0

95 00 95 00 95 00 95 00 Basic welfare Support for Support for

TANF- related services

working poor working poor +

social services health care

Federal State

To assess how state budget decisions might have affected spending, we
collected data on how Texas used state and federal funds to finance
TANFrelated social services expenditures. The results are grouped into four
categories, as shown in figure 12, and described in more detail in appendix
I. Each category is identified on the graph by two bars- one for spending
that occurred in state fiscal year 1995 and one for spending that occurred
in state fiscal year 2000. From left to right, the categories are
cumulative. Expenditures for cash assistance and employment training are
included in expenditures for support of the working poor. These expenditures
in turn

are included in expenditures for support of the working poor plus health
care. Finally, all expenditures are included in social services
expenditures.

Texas is the only state in our study that saw nominal growth in total
expenditures for basic welfare services- cash assistance and job training
programs- between state fiscal years 1995 and 2000. Spending remained

constant, in real terms, which is significant because Texas also experienced
a 54 percent decline in the number of families receiving cash assistance
over this same period. Spending on the training component of the CHOICES
program drove this trend as the state invested heavily in these

programs. The new investment was, in large part, financed with federal
dollars. Our survey data show that while total spending on these programs
grew from $710 million to nearly $780 million over the period, the state?s
contributions declined from $254 million to $216 million (see table 30). As
a result, while the state financed about 36 percent of these programs in
1995, in 2000 it financed only about 28 percent.

Table 30: Texas? TANF- Related Expenditures

Dollars in millions

Real Nominal (expressed in 1995 dollars) 1995 2000 1995 2000

Basic welfare services Total funds $710 $778 $710 $710 State funds $254 $216
$254 $197 Support for the working poor Total funds $1,476 $2, 066 $1,476 $1,
885 State funds $638 $774 $638 $707 Source: GAO survey and analysis of state
expenditure data.

As in other states we surveyed, Texas officials claim that most of the state
funds freed- up with TANF funds were reinvested in programs that support the
broad goals of federal welfare reforms. Total spending rose significantly-
28 percent in inflation adjusted dollars- among a group of programs that
include those same basic welfare services as well as programs that provide
transitional support for families moving from

welfare to work (e. g. child care and child development, transportation
subsidies, and other post- employment services). Although it is impossible
to tell with any certainty what programs a freed up state dollar funded, the
state did substantially increase its own investment in programs that broadly
support the working poor- state spending rose by about

11 percent in real terms from 1995 to 2000. (See table 30.) In line with the
state?s budget strategy to maximize the use of federal funds, in 2000
federal funds represent a much larger share of these programs? finances than

before welfare reform. In 1995 the state financed about 43 percent of these
programs, in 2000 the state share receded to 38 percent.

State spending on health care for the working poor rose faster than federal
spending between 1995 and 2000 chiefly because the Medicaid match rate
changed requiring the state to finance a greater share of that program than

it did in 1995. When these state expenditures are combined with state
spending on other programs supporting the working poor, the rate of spending
outpaced inflation. Total federal and state spending rose- in real terms- by
12 percent, while state spending rose by 11 percent. (See table 31.)

Table 31: Texas? TANF- Related Expenditures

Dollars in millions

Real Nominal (expressed in 1995 dollars) 1995 2000 1995 2000

Support for the working poor plus health care Total funds $4,648 $5, 722
$4,648 $5, 221 State funds $1, 800 $2,181 $1, 800 $1, 990 TANF- related
social services Total funds $8, 333 $10, 908 $8, 333 $9, 953 State funds $4,
361 $5,574 $4, 361 $5, 086 Source: GAO survey and analysis of state
expenditure data.

As noted previously, Texas has focused recently on expanding its child
welfare programs and other social services. TANF funds can be used in all
these programmatic areas provided they are used on TANF- eligible

families. For example, Texas devoted more than $173 million of its TANF
funds to child welfare activities in 2000 (39 percent of TANF funds spent in
2000 compared to 14 percent of Title IV- A funds spent in 1995). In
addition,

these programs received a large share of state funding as well. For all
social services, total federal and state spending rose 19 percent in real
terms; real state spending kept nearly the same pace, rising 17 percent. As
a result, the fiscal balance across the broadest category of social services
spending has remained stable since 1995; the state financed 52 percent in
1995 and in 2000 it financed about 51 percent.

Additional federal safety net programs provide food assistance and
supplemental security income for needy families and children in Texas. For
example, the federal government provided $2. 3 billion in food stamp
benefits to low income individuals in Texas in 1995 and $1.2 billion in
2000. In addition, from 1995 to 2000, federal supplemental security income
payments to Texans increased from $1.4 billion to $1.6 billion, about 73

percent of which went to individuals under the age of 65. Rainy Day Funds
The state has a statutorily established contingency fund for TANF

expenditures. However, the state has not funded this account based on
projections of future program needs. Rather an appropriation equal to the
balance of all federal TANF funds not already appropriated for other
purposes is made for this purpose every 2 years. The contingency
appropriation requires that the expenditure of funds from this account be
made only upon approval of the Governor and the Legislative Budget Board. 2
Although state general revenues are another potential source of

rainy day funds, their use for cash assistance is limited by the state
constitution to 1 percent of the total biennial budget.

After other state priorities were funded for state fiscal years 2000- 2001,
budget analysts estimated that $107 million would be available for the
contingency fund. Federal regulations require that states report all rainy
day funds as unobligated balances and therefore these funds are commingled
with all other funds the state reports as unobligated. Soon

after the beginning of the second year of its biennium (September 1, 2000),
Texas reported $141.2 million in unobligated balances. (See table 32.)

2 The 10- member Legislative Budget Board was created by statute in 1949 for
the primary purpose of developing recommended legislative appropriations for
all agencies of the state government. Membership of the LBB is provided by
law. The Lieutenant Governor is the Chairman, and the Speaker of the House
of Representatives is Vice Chairman. Chairs of the House Appropriations
Committee, House Committee on Ways and Means, Senate Finance

Committee, and Senate State Affairs Committee are automatically members. The
Lieutenant Governor appoints two additional members of the Senate, and the
Speaker appoints two additional members of the House to complete the 10-
member Board.

Table 32: Unspent TANF Funds in Texas as of September 30, 2000

Dollars in millions

Unspent funds as a Unobligated

Unliquidated obligations Tot al unspent

percent of annual TANF funds of TANF funds funds

TANF grant

$141. 2 $41. 6 $182. 8 37. 6% Source: HHS? Administration for Children and
Families.

An analysis of the balances Texas has left unobligated at the U. S. Treasury
for future needs is complicated because these reported balances also
represent funds the state has budgeted but has not yet spent. TANF funds for
Texas are shown on federal reports as obligated to the extent that they are
contracted out to state workforce boards for the CHOICES program,

but are shown as unobligated where they are awaiting disbursement by state
agencies that provide cash assistance and other services. About onequarter
of TANF funds are contracted through the workforce boards and the remainder
are split among several state agencies.

Generally, CHOICES is run by local workforce boards that offer job training
and education programs, child care subsidies, and other supportive services
designed to help welfare recipients transition from welfare to work. Each
year the state executes performance- based contracts with these boards,
providing a variety of federal and state funds. When the state contracts
with the boards for TANF- funded services, the state appropriately records
this in reports to HHS as an obligation of federal

funds. However, many other services are funded with TANF through state
social service, child welfare, and substance abuse treatment agencies.
Because these are state agencies, the TANF funds needed to provide services
must be recorded as unobligated balances until the agencies have spent the
funds or have contracted for provision of the services.

Appendi x XI

Wisconsin Background

State fiscal year: July 1- June 30 Budget Cycle: Biennial TANF Grant: a $318
million 75% MOE: $169 million Cash Assistance caseload: b 74, 000 families
in 1995

16, 000 families in 2000 78 percent decline a Wisconsin received about $16
million in federal TANF funds in 2001 as a bonus for successful program
outcomes. b Caseloads as reported to HHS for January 1995 and June 2000 and
rounded to the nearest thousand.

Wisconsin was one of the first states to reform its welfare program,
starting in 1987, and went statewide with its current program, Wisconsin
Works (W- 2), in September 1997. As a result, the decline in the state?s
cash

assistance caseload began at an earlier date and proceeded at a faster rate
than in many other states, but by 2000 the number of families receiving cash
assistance was relatively stable. In accordance with the state?s workfirst
philosophy, the program serves those who are willing to work to their full
capability. Eligible families receive cash assistance payments that are
equivalent to hourly wages for time spent in work and other assigned
activities, such as

education or drug treatment. All families with children and incomes under
115 percent of the federal poverty level are eligible. Once enrolled, the
maximum monthly assistance a family can receive is $673, regardless of
family size. This amount is paid when a family member performs community
service work for 30 hours a week and participates in 10 hours of education
and training. The amount is a 30- percent increase over the

amount Wisconsin paid in 1995 for a family of three, making it one of the
highest cash assistance payments in the nation. However, the payment is
reduced for each hour that the family member fails to attend the assigned

activity. In addition to cash assistance, parents may receive any child
support that the state collects on their behalf. 1 Families participating in
1 The state and federal governments may retain collections as reimbursement
for cash

assistance payments made to these families. Wisconsin has chosen instead to
give its full state share to W- 2 families.

W- 2 who have received monthly cash assistance payments for the equivalent
of 5 years are ineligible for further receipt of cash assistance, in line
with TANF law. For recent legal immigrants, the state has set aside its own
state funds to pay cash assistance because federal law prohibits use of TANF
funds for

this purpose. Children receiving cash assistance whose parents are not
eligible for TANF, such as those receiving supplemental security income, are
served in a separate program and are not enrolled in W- 2. Services and
benefits available to W- 2 participants include case management, job
training, transportation assistance, and matching grants of up to $500 for
vocational training. In addition, subsidized healthcare, called Badger Care,
and child care are available to W- 2 participants and

other families with children and incomes up to 200 percent of the poverty
level. Other support services were made available to families with incomes
below 200 percent of poverty starting in mid- 1999 under the state?s
community reinvestment program. TANF funds are also used to support an
earned income tax credit (EITC) as well as child welfare, juvenile justice,

mental health programs, services for the developmentally disabled, substance
abuse treatment activities, and incentives for contractors who meet
performance goals.

Most of the federal and state funds for the W- 2 program are obligated by a
single state agency in contracts with local agencies that determine
eligibility and provide cash assistance and services to program
participants. W- 2 is administered by over 70 local agencies, primarily
county social services agencies, under 2- year, performance- based contracts

with the state. Supplantation in the

The dramatic changes in Wisconsin?s cash assistance caseload as well as
State Budget

the breadth of services the state had decided to fund prompted budget
officials to begin looking for creative ways to finance state programs for
the working poor. In Wisconsin?s 1999- 2001 biennial budget, the state used
about $164 million of its TANF grant to replace state general funds. The

state plans to continue using about $52 million of TANF funds per year to
change the source of financing for these programs. 2 The largest single 2
About $48 million of the total $164 million was used retroactively to adjust
the appropriations level in the final year of the previous budget.

program affected by the refinancing was the state?s tax credit program, the
EITC. Each year Wisconsin plans to use about $52 million in federal TANF
funds to finance the state tax credit that previously had been financed with
state funds. Under current projections for use of TANF funds, the state will
use about 16 percent of its annual block grant in this manner. The
circumstances surrounding the state?s decisions to use its federal funds in
this way are illustrative of the nature of budget negotiations and the role
that federal funds can play in influencing trade- offs among a variety of
financing options. According to legislative members and staff we

interviewed, when TANF was enacted in 1996, Wisconsin was already spending
less than the MOE required them to spend because they had already instituted
substantial reforms and had already achieved significant savings as a result
of their reforms. So when MOE levels were pegged to state spending levels in
1994, it created a fiscal burden for Wisconsin. The state had to raise its
spending in order to meet the 75 percent MOE level.

During the 1999- 2001 budget deliberations, state budget analysts projected
that the state would not meet its MOE requirement because the number of W- 2
families receiving cash assistance had fallen so dramatically. Even

with agreement on several new spending proposals, budget analysts calculated
that the state would run $40 million short over the biennium. A shortfall of
this size presented the state with a challenge. State budget officials
predicted that it would be difficult to gain bipartisan approval in the
legislature for new social service programs that required an investment of
new state funds. The legislature had already enacted expansions to the
state?s subsidized child care program to serve more lowincome families and
had approved a new health care program for lowincome families- BadgerCare.
Budget officials looked to existing programs already in the budget that
might be used to meet the MOE.

In the midst of the state?s budget negotiations, the final TANF regulations
were issued that allowed use of TANF funds to establish or expand state EITC
programs. State officials realized that the regulations permitted them

to replace about $53 million a year in state funds that they were spending
on their EITC with federal TANF funds. However, doing so still left them
short of meeting their MOE requirement. The state had a shortfall to begin
with, and the shift in funds only served to increase the shortfall. However,
the legislative committee leadership realized that freeing up state general
revenues provided funds it could use as leverage to ensure bipartisan
support for a package of spending and revenue options. Some of the freed-

up state general funds could be used for a property tax cut, and the rest to
expand certain programs for the state?s low- income population. In addition,
the expanded programs serving the low- income population could be used to
meet the state?s MOE requirement so there would be no shortfall.

According to the committee leadership, the compromise also resulted in
increased investment of the federal TANF funds in a variety of programs. The
package presented during the negotiations afforded the committee the
opportunity to bring TANF spending on new programs up even farther thereby
increasing the state?s investment on low- income families. The

leadership stressed that these new programs could not have been funded
during the 1999- 2001 biennium in the absence of the options arising from
the freed- up state general funds.

Spending Trends From 1995 to 2000

Figure 13: Wisconsin?s TANF- Related Expenditures for State Fiscal Years
1995 and 2000

1995 dollars in millions 3, 500

3, 000 2, 500 2, 000 1, 500 1, 000

500 0

95 00 95 00 95 00 95 00 Basic welfare Support for Support for TANF- related
services

working poor working poor +

social services health care

Federal State

To assess how state budget decisions might have affected spending, we
collected data on how Wisconsin used state and federal funds to finance
TANF- related social services. The results are grouped into four categories,
as shown in figure 13 and described further in appendix I. Each category is
identified on the graph by two bars-- one for spending that occurred in
state

fiscal year 1995 and one for spending that occurred in state fiscal year
2000. From left to right, the categories are cumulative. In other words,
expenditures for cash assistance and employment training are included in
expenditures for support of the working poor. These expenditures in turn

are included in expenditures for support of the working poor plus health
care. Finally, all expenditures are included in social services
expenditures.

Given the dramatic decline in the cash assistance caseload, the decline in
total spending on these basic activities would be expected. Since fewer
families are receiving cash assistance, spending on cash assistance has
plummeted despite the higher monthly assistance payment level. The decline
in total real spending was 46 percent. (See table 33.) The decline in
spending on cash assistance was partially offset by spending on job

training, which more than tripled. In 2000, for every dollar spent on cash
assistance, three dollars were spent on job training. Notably, however, and
in line with the state?s struggles to meet its MOE requirement, state
spending declined only marginally- about 4 percent- in real terms. As a
result, the state now finances a greater share of these costs. Whereas the

state financed 41 percent of the costs of these services in 1995, in 2000 it
financed about 72 percent. Table 33: Wisconsin?s TANF- Related Expenditures

Dollars in millions

Real Nominal (expressed in 1995 dollars) 1995 2000 1995 2000

Basic welfare services Total funds $412 $244 $412 $223 State funds $167 $176
$167 $161 Support for the working poor Total funds $898 $881 $898 $804 State
funds $497 $485 $497 $443 Source: GAO survey and analysis of state
expenditure data.

Total federal and state spending on programs that more broadly support the
working poor also declined over the same period. However, when viewed
cumulatively, the steep decline in spending on basic welfare services was
moderated by large investments in other programs. For example, spending on
child care alone has increased by about $123 million.

However, state spending on all components of support for the working poor
declined by 11 percent, roughly the same decline seen in total spending.
This indicates that the state has maintained its own investment in these
programs. Wisconsin continues to finance about 55 percent of the costs of
these programs with state funds.

In 1999, Wisconsin began enrolling certain low- income families without
health insurance in a new statewide program called Badger Care. The program
is supported by state funds and by two federal funding sources- Medicaid and
the State Children?s Health Insurance Program (SCHIP). 3 Budget forecasts
projected that spending would rise by about $100 million

a year- a substantial new investment- once the new program was fully
implemented. About a third of the increase was projected to have been funded
with state funds. Given the nature of the matching grants, it is not
surprising that federal spending would rise faster than state spending. (See
table 34.) Overall, combined spending on support for the working poor plus
health care grew by about 2 percent, while state spending declined about 3
percent in real terms. The state?s share of these total expenditures fell
from 45 percent in 1995 to 43 percent in 2000.

3 The two federal programs are matching grants: the federal government
finances about 71 percent of SCHIP program in Wisconsin and about 59 percent
of the Medicaid program. In January 2001, the state received approval for
its request to waive provisions of federal

law that prohibit the use of SCHIP funds for services provided to adults.
The state can now use the more favorable match rate (71 percent) to finance
a larger share of its program.

Table 34: Wisconsin?s TANF- Related Expenditures

Dollars in millions

Real Nominal (expressed in 1995 dollars) 1995 2000 1995 2000

Support for the working poor plus health care Total funds $2,591 $2,908 $2,
591 $2, 653 State funds $1, 163 $1, 238 $1, 163 $1, 129 TANF- related social
services Total funds $3,006 $3,518 $3, 006 $3, 210 State funds $1, 455 $1,
578 $1, 455 $1, 440 Source: GAO survey and analysis of state expenditure
data.

The decline in the state?s share of funding for programs supporting
lowincome families that was first shown in our analysis of programs that
support the working poor is reflected in the state?s spending on social
services. Total spending on social services grew by about 7 percent; but

state spending decreased by about 1 percent. Wisconsin?s share of spending
on all social services that support the goals of TANF declined from about 48
percent in 1995 to about 45 percent in 2000.

Additional federal safety- net programs provide food assistance and
supplemental security income (SSI) for needy families with children in
Wisconsin, but are not shown in figure 13. The federal government provided
$227 million in food stamp benefits to low- income individuals in

Wisconsin in 1995, and $115 million in 2000. In addition, from 1995 to 2000,
federal supplemental security income payments increased from $362 million to
$374 million in nominal dollars- about 88 percent of these benefits went to
individuals under 65 years of age.

Rainy Day Funds In 1997 Wisconsin created a $50 million reserve account for
(1) its W- 2 program, (2) additional benefit payments for Milwaukee, and (3)
benefits to children of SSI recipients. The state drew down most of the
reserves it had

set aside for the benefit payment for Milwaukee and children of SSI
recipients by the end of the first biennium. The state approved a $102
million deposit to the Contingency Fund for the 1999- 2001 biennium for the
W- 2 program. These funds are available to W- 2 agencies in the event of an
economic downturn. 4 However, disbursement from this fund must be

approved by the Joint Finance Committee. In addition to the Contingency
Fund- which is specifically set aside for a rainy day- the state left about
$16 million in TANF funds unbudgeted over the biennium.

State funds that are set aside for a rainy day must, under HHS regulations,
be reported as unobligated balances. The state?s most recent submission for
the end of federal fiscal year 2000 is difficult to reconcile with the
state?s own budget outline. (See table 35.) This difficulty highlights the
inadequacy of HHS? financial reports to offer any insight into the states
plans for their unspent balances.

Table 35: Unspent TANF Funds in Wisconsin as of September 30, 2000 Unspent
funds as a Unobligated

Unliquidated Total unspent

percent of annual balance obligations funds

TANF grant

$40.5 $252.2 $292. 4 92% Source: HHS? Administration for Children and
Families.

The data in table 35 is prepared at the end of the federal fiscal year and
as such represents a snap shot in time rather than a flow of funds. A state
budget official said that to avoid listing too much as ?unobligated? it
treats its Contingency Fund more like a revolving fund. For example, the
balance in the Fund is always $102 million, but by the end of the federal
fiscal year the state would obligate the balance to pay for ongoing program
needs and then use the next fiscal year?s grant for the Contingency Fund.
This state official said that, in part, the motivation to record obligations
in this manner stems from regulations that limit the use of any unobligated
TANF

4 Unspent federal funds held in reserve for a rainy day must remain on
deposit at the U. S. Treasury until the state needs them to make a payment.

balances. He also expressed concern that if they reported significant
unobligated balances, it might send a signal to Congress that those funds
were not needed.

The vast majority of Wisconsin?s unspent balances are claimed as
unliquidated obligations. As noted previously, Wisconsin operates W- 2
through local contractors who provide the services available under the

program. At the time the state signs the contracts, it reports the full
amount of the federal funds involved as obligated, in accordance with
federal regulations. The reported obligations are shown as unliquidated,
however, until the local contractors provide or purchase services for W- 2
families. These contracts extend for 2 calendar years. In September 1999,
the state signed 75 contracts that covered the period from January 1, 2000,
to December 31, 2001. The total value of the contracts for 2 years of
services and benefits is about $370 million, only $60 million more than the

state?s annual block grant. But the state has other TANF- related
expenditures that cannot be recorded as obligations because of the nature of
the expenditures. For example, child care subsidies are paid as the services
are rendered. While the state has an underlying commitment to pay child care
subsidies to eligible families, the amount of the subsidy and the
recipient?s eligibility for the subsidy depend on the recipient?s income and
therefore a legal obligation,

as required under the regulations, does not exist until the payment is made.
So even though the state might project considerable outlays in its child
care program, these future expenditures do not represent an obligation as
defined by federal regulations until those expenditures are due.

No funds have been drawn from the Contingency Fund, and the Governor
proposed to eliminate it for the 2001- 2003 biennium. The $102 million on
deposit will be transferred out of the state?s Contingency Fund and is
projected to be spent on existing programs over the 2001- 2003 biennium.

The Governor?s proposed 2001- 2003 budget projects that about $66 million
will be left unbudgeted in the biennium?s first year, but that by the end of
the second year of the biennium, projected spending will rise leaving only
about $600,000 unbudgeted.

Appendi x XII

Survey Instrument Attachment 1

          

When completing the survey, please keep the following in mind: 1.  Identify
all state programs serving social service needs, particularly programs

targeted towards reducing dependence on public assistance. 2.  Distribute a
copy of the survey to all agencies that oversee these programs. Please

explain to these agencies what MOE means and what funds should be shown in
the MOE column. 3.  Include all federal, state and local expenditures that
are incorporated in the state

budget. For local expenditures, include local spending of locally raised
revenue that is incorporated in the state budget such as a local match.
Include expenditures or estimated expenditures only (not amounts budgeted or
authorized). 4.  Include TANF and MOE expenditures, as well as expenditures
from any other funding

source for each program. Include all TANF spending; if some expenditures do
not fit into one of the specific program categories, include them in one of
the lines labeled

?other.?

5.  Provide expenditures for the state fiscal years of 1994- 1995, 1998-
1999, and estimated expenditures for 1999- 2000 (not the federal fiscal
year). 6.  Do not include capital expenditures. 7.  Do not include
indirect administrative costs or management information systems

(MIS) expenditures, but do include direct administrative costs such as case
management expenditures in relevant program line items. If case management
was included in administrative expenditures in 1994- 1995, estimate and
include where relevant. 8.  Include the costs of fringe benefits for state
personnel. (A rough estimate of fringe

benefit costs is all that is necessary.) 9.  Identify funding streams
included in the columns labeled ?Other? on a separate

worksheet. 10.  For columns labeled ?SSBG? (Social Services Block Grant):
If state officials cannot

isolate spending on individual programs, obtain either estimates for these
amounts, or totals and document the general areas in which SSBG funds are
spent. For TANF funds transferred from TANF to SSBG document as SSBG
expenditure with note on level attributable to the transfer. 11.  Compile
and provide copies of all supporting documentation for the data entered in

the survey: e. g. expenditure reports, annual financial statements, or
enacted budgets (for fiscal year 2000 estimates). 12.  When possible,
identify caseload and eligibility criteria for each program in

supporting documentation.

Attachment 1 Suggested format for data submission: For example, Line 2b,
Training, could include expenditures from several programs across two or
more agencies. Each of these agencies would complete the survey as well as
provide the supporting documentation.

Agency 1

Dept of Social

I II III IV

Services

Work Preparation and Education b) training Program w 0 0 18,000 6,000
Program x 250,000 0 2, 000 0 Total 250,000 0 20,000 6,000

Agency 2

Dept. of Family

I II III IV

Employment

Work Preparation and Education b) training Program y 650,000 40,000 0 4, 000
Program z 100,000 10,000 0 0 Total 750,000 50,000 0 4, 000

The various agencies? contributions could be compiled centrally and
summarized

State Summary

I II III IV

Totals

Work Preparation and Education b) training 1,000,000 50,000 20,000 10,000

Glossary of column headings: IV- A: Former Aid to Families with Dependent
Children (AFDC) program IV- F: Former Job Opportunities and Basic Skills
(JOBS) program SSBG: Social Services Block Grant, title XX of the Social
Security Act TANF: Temporary Assistance for Needy Families TANF- MOE: TANF
Maintenance of Effort, see your state TANF director or http:// www. acf.
dhhs. gov/ programs/ ofa/ funds2.htm MOE- SSP: Maintenance of Effort -
Separate State Programs

Attachment 2

         

Line 1: Poverty Relief

Line 1a: cash assistance. Include expenditures on cash payments or vouchers
provided to families to meet ongoing, basic needs, net of child support
collections. Any cash assistance program that was formerly funded by IV- A
should be included here. (Note: This definition is adapted from the
definition of basic assistance in line 5a of the ACF- 196 Financial Report.)

Line 1b: child support payments. Include all child support collections from
non- custodial parents that are passed on to custodial parents who are
receiving cash assistance through TANF, in excess of $50 per monthly
payment.

Line 1c: emergency assistance. Include all expenditures for emergency
assistance (the former IV- A program), including prevention of eviction,
utility cut- off, etc. Document, to the extent possible, how emergency
assistance funds are allocated.

Line 1d: food assistance. Include expenditures on programs designed to
provide food or nutritional assistance to low- income people. Include state
administrative expenses for Food Stamps. If known, also provide federal
expenditures on Food Stamps in your state fiscal year.

Line 1e: housing assistance. Include expenditures on programs designed to
provide housing assistance to low- income people, such as vouchers, state
low- income housing tax credits, or any other state support for low- income
housing efforts.

Line 1f: SSI supplements. Include expenditures on state supplementation of
the federal Supplemental Security Income program. Do not include federal
expenditures.

Line 1g: other. Include expenditures on any other programs related to
poverty relief that are not included above. Describe such programs on an
attached sheet.

Line 2: Work Preparation and Education

Line 2a: education. Include expenditures on educational activities that
prepare the recipients for work. For example, include secondary education
(including alternative programs); adult education, GED, and ESL classes;
education directly related to employment; education provided as vocational
educational training; and post- secondary education. Do not include
education

Attachment 2 programs for children below high- school age. In this instance,
limit spending to TANFeligible people.

(Note: This definition is adapted from the definition of education in line
6a2 of the ACF196 Financial Report.)

Line 2b: training. Include expenditures on programs to prepare people who
are not yet working with skills to make them employable. Examples include
skills development programs, community service placements, etc. Do not
include expenditures on people who are in the paid workforce.

Line 2c: other. Include expenditures on any other programs related to work
preparation and education that are not included above. Document programs
descriptions on an attached sheet.

Line 3: Employment Support

Cells are shaded. Include expenditures in this category on lines 3a- 3e
below.

Line 3a: post- employment services. Include expenditures on programs
designed to keep people employed after they have found employment. Examples
include coaching to ensure that individuals arrive at work on time,
counseling to address problems that may arise in the workplace, and any
other case management services for this working population. If known,
include spending for on- the- job training.

Line 3b, c: state EITC. Include expenditures on state earned income tax
credits paid to families. Include state and local tax credits that are
designed to defray the costs of employment for low- income families.

Line 3d: transportation. Include the value of transportation benefits (such
as allowances, bus tokens, car payments, auto insurance reimbursement, and
van services) provided to employed families (related either to their work or
related job retention and advancement activities) and provided as a
nonrecurring, short- term benefit to non- working families (e. g. during
applicant job search). (Note: this definition is adapted from the definition
of transportation in line 6c of the ACF- 196 Financial Report.)

Line 3e: wage subsidies. Include payments to employers or third parties to
help cover the costs of employee wages, benefits, supervision, or training.
Also include any wage- related tax credits that benefit employers. (Note:
This definition is adapted from the definition of work subsidies in line 6a1
of the ACF- 196 Financial Report.)

Attachment 2

Line 3f: other. Include expenditures on any other programs related to
employment support that are not included above. Describe such programs on an
attached sheet.

Line 4: Poverty Prevention

Cells are shaded. Include expenditures in this category on lines 4a- 4d
below.

Line 4a: diversion payments. Include expenditures on nonrecurrent, short-
term benefits to families in the form of cash payments, vouchers, or similar
form of payment to deal with a specific crisis situation or episode of need.
An example is an emergency rent payment to prevent eviction. (Note: this
definition is adapted from the definition of diversion payments in line 6g
of the ACF- 196 Financial Report.)

Line 4b: family formation and pregnancy prevention. Include expenditures on
programs aimed to keep families together, prevent teen pregnancy and prevent
single parenthood. Examples of program expenditures are: responsible
fatherhood initiatives that will improve the capacity of needy fathers to
provide financial and emotional support for their children; premarital and
marriage counseling, and mediation services; counseling services or classes
that focus on teen pregnancy prevention; media campaigns to encourage young
people to delay parenting or to encourage fathers to play a responsible role
in their children's lives; and incentives for single parents to marry or for
two- parent families to stay together. (Note: this definition is identical
to the section in ACF?s ?Helping Families Achieve SelfSufficiency? guide
entitled, ?Appropriate Uses of Funds/ Family Formation and Pregnancy
Prevention? at http:// www. acf. dhhs. gov/ programs/ ofa/ funds2.htm )

Line 4c: other. Include expenditures on any other programs related to
poverty prevention that are not included above. Describe such programs on an
attached sheet.

Line 5: Child Protection/ Juvenile Justice

Cells are shaded. Include expenditures in this category on lines 5a- 5c
below.

Line 5a: child welfare. Include expenditures on adoption assistance, foster
care, and independent living programs; on any program intended to prevent
out- of- home placements, promote reunification of families, or provide a
safe environment for children; and on programs that focus on child abuse
prevention and neglect prevention. Examples of expenditures include using
funds for family counseling; parent support programs; appropriate supportive
services (e. g., referral services, child care, transportation, and respite
care) to caregiver relatives who can provide a safe place for a needy child
to live and avoid his or her placement in foster care; and screening
families for risk of child abuse or neglect and providing case management.
(Note: Any cash assistance program that was formerly funded by IV- A, such
as cash assistance to needy caretaker relatives, should be included in the
?cash assistance? category.)

Attachment 2

Line 5b: juvenile justice programs. Include expenditures on social services
programs for youth who have violated the state juvenile code. Do not include
institutional spending.

Line 5c: other. Include expenditures on any other programs related to child
protection/ juvenile justice that are not included above. Describe such
programs on an attached sheet.

Line 6: Other

Cells are shaded. Include expenditures in this category on lines 6a- 6c
below.

Line 6a: provider profits. Include all bonuses, incentive payments or
profits to contractors for provision of services to low- income people.

Line 6b: substance abuse prevention and treatment. Include expenditures on
programs aimed to prevent alcohol, drug and tobacco abuse and to provide
intervention services to individuals with alcohol, drug and/ or tobacco
dependency in their families. Examples of prevention programs are media
campaigns, educational programs and community- based planning programs.
Examples of expenditures on treatment include counseling, treatment
facilities, and outpatient medical care.

Line 6c: developmental disabilities. Include expenditures on programs that
provide services to individuals with developmental disabilities and their
families, including outpatient care and public education, but excluding
institutional facilities.

Line 6d: mental health services. Include expenditures on programs that
provide prevention and/ or intervention services to the mentally ill and
their families, including community- based treatment facilities, outpatient
care and public education. Exclude all expenditures provided at/ through
mental health institutions.

Line 6e: other Include expenditures on any other programs that are not
included above. Describe such programs on an attached sheet.

Child Care/ Child Development Spending Survey

Instructions for Child Care/ Child Development: Include expenditures on any
child care or child development program, either custodial or educational,
in- home or out- of- home, aimed at either working or non- working people,
including pre- K programs, after- school programs, vouchers for child care,
state expenditures on Head Start, and subsidies to child care centers, and
child care tax credits (if available). Include programs for both TANF-
eligible and non- TANF- eligible people. Please identify each child care/
child development program in the spaces below and identify their funding
streams. Please identify eligibility criteria for these programs, as well as
caseloads (numbers of children, not families, if possible), on an attached
sheet.

1994 - 95 Child Care/ Child Development Spending Federal Expenditures State
Expenditures

TOTALS

Program Names CCDBG Other

TOTALS 0 0000

1999 - 00 Child Care/ Child Development Spending Federal Expenditures State
Expenditures

TOTALS

Program Names CCDF Other CCDF- MOE CCDF- Match Other

TOTALS 0 00000 CCDBG - Child Care and Development Block Grant CCDF - Child
Care Development Fund

Healthcare Coverage Spending Survey

Instructions for Healthcare Coverage Spending Survey: Include expenditures
on any healthcare program, in- home or out- of- home, aimed at either
working or non- working people and their children, excluding longterm care.
Include programs for both the TANF- eligible and non- TANF- eligible non-
elderly population. Identify each program in the spaces below and their
funding streams. Identify eligibility criteria for these programs, as well
as caseloads on an attached sheet. For Medicaid- funded programs, identify
target populations (e. g. "transitional assistance," "expansion population")
where possible. State expenditures should capture local spending if it flows
through the state budget (e. g. a local match).

1994 - 95 Healthcare Coverage Spending Federal Expenditures State
Expenditures TOTALS

Program Names Title XIX Title XXI Other Title XIX Title XXI Other

TOTALS 0 0 0 0 0 0 0

1999 - 00 Healthcare Coverage Spending Federal Expenditures State
Expenditures TOTALS

Program Names Title XIX Title XXI Other Title XIX Title XXI Other

TOTALS 0 0 0 0 0 0 0

State FY 1994 - 95 Social Services Spending Survey Federal Expenditures
State Expenditures Local Expenditures TOTALS

I II III IV V VI VII VIII IX X IV- A IV- F SSBG Other IV- A IV- F Other IV-
A Other 1 Poverty Relief

a) cash assistance b) child support payments c) emergency assistance d) food
assistance e) housing assistance f) SSI supplements g) other (please
specify) h) total 0 0 0 0 0 0 0 0 0 0 2 Work Preparation and Education

a) education b) training c) other (please specify) d) total 0 0 0 0 0 0 0 0
0 0 3 Employment Support

a) post- employment services b) state EITC- non- refundable c) state EITC-
refundable d) transportation e) wage subsidies f) other (please specify) g)
total 0 0 0 0 0 0 0 0 0 0 4 Poverty Prevention

a) diverson payments b) family formation/ preg. prev. c) other (please
specify) d) total 0 0 0 0 0 0 0 0 0 0 5 Child Protection/ Juvenile justice

a) child welfare b) juvenile justice programs c) other (please specify) d)
total 0 0 0 0 0 0 0 0 0 0 6 Other

a) provider profits b) substance abuse prev. & treat. c) developmental
disabilities d) mental health services e) other (please specify) f) total 0
0 0 0 0 0 0 0 0 0 7 TOTALS 0 0 0 0 0 0 0 0 0 0

State FY 1999- 2000 Social Services Spending Survey (estimated) Federal
Expenditures State Expenditures Local Expenditures TOTALS

I II III IV V VI VII VIII IX TANF SSBG Other TANF- MOE MOE- SSP Other TANF-
MOE Other 1 Poverty Relief

a) cash assistance b) child support payments c) emergency assistance d) food
assistance e) housing assistance f) SSI supplements g) other (please
specify) h) total 0 0 0 0 0 0 0 00 2 Work Preparation and Education

a) education b) training c) other (please specify) d) total 0 0 0 0 0 0 0 00
3 Employment Support

a) post- employment services b) state EITC- non- refundable c) state EITC-
refundable d) transportation e) wage subsidies f) other (please specify) g)
total 0 0 0 0 0 0 0 00 4 Poverty Prevention

a) diverson payments b) family formation/ preg. prev. c) other (please
specify) d) total 0 0 0 0 0 0 0 00 5 Child Protection/ Juvenile justice

a) child welfare b) juvenile justice programs c) other (please specify) d)
total 0 0 0 0 0 0 0 0 0 6 Other

a) provider profits b) substance abuse prev. & treat. c) developmental
disabilities d) mental health services e) other (please specify) f) total 0
0 0 0 0 0 0 0 0 7 TOTALS 0 0 0 0 0 0 0 00

Comments From the Department of Health Appendi x XI II and Human Services

See comment 1. See comment 2. See comment 3. See comment 4.

See comment 5. See comment 6.

See comment 7.

GAO Comments 1. As we reported, many aspects of the MOE requirement play a
significant role in the maintenance of the federal- state fiscal balance.

We agree with HHS that many aspects of the MOE could be taken up during the
forthcoming TANF- reauthorization debate. We believe that footnote 2
provides adequate information for the reader to understand the specifics of
the states? financing requirements. We also refer readers to our previous
report for additional information on this important provision in the law.

2. We added footnote 8 on page 9 to clarify that HHS regulations restrict
the use of TANF carry- forward funds and clarified our discussion of the

?new spending test? on pages 23- 24 of the report. 3. Now on page 7 of the
report. 4. See footnote 8. HHS suggests that the restriction on the use of
state

carry- forward funds may be a factor in explaining why states are using
federal funds to pay an increased share of the costs of basic cash
assistance. Clearly, regulations play an important role in state
decisionmaking. However, in many of the states we reviewed this restriction
was not the primary factor that influenced their budget decisions.
Furthermore, many factors- regulations, the economy, and other state
budgetary constraints- influence state decision- making; the fiscal balance
is not static and will continually evolve. This is the primary

reason we recommend that the department work with the states to gather data
periodically to assess the federal- state fiscal balance relevant to
achieving the broader goals of the TANF program. Such assessments will
provide important information on how the myriad factors that influence state
decision- making affect the fiscal balance.

5. We added footnote 20 to distinguish the two components of ?unspent
funds?- unliquidated obligations and unobligated balances. The body of
report discusses our view that such distinctions have significant
limitations because states do not report this information in a consistent
manner.

6. We clarified our discussion of the ?new spending test? on pages 23- 24 of
the report. 7. In our discussion on pages 25- 26 of this report we address
the lack of nonduplicative case counts and the implications it has on our
analysis.

Specifically, it becomes problematic when one tries to compare the level of
effort under TANF with the level of effort under AFDC on a per case basis.
Absent better data on caseloads, our focus on the fiscal balance was
designed to address the extent to which states were using their TANF funds
to supplant their own funds controlling for some changes to the program?s
workload.

GAO United States General Accounting Office

Page 1 GAO- 01- 828 Welfare Reform

Contents

Contents Page 2 GAO- 01- 828 Welfare Reform

Contents Page 3 GAO- 01- 828 Welfare Reform

Page 4 GAO- 01- 828 Welfare Reform

Page 5 GAO- 01- 828 Welfare Reform United States General Accounting Office

Washington, D. C. 20548 Page 5 GAO- 01- 828 Welfare Reform

Page 6 GAO- 01- 828 Welfare Reform

Page 7 GAO- 01- 828 Welfare Reform

Page 8 GAO- 01- 828 Welfare Reform

Page 9 GAO- 01- 828 Welfare Reform

Page 10 GAO- 01- 828 Welfare Reform

Page 11 GAO- 01- 828 Welfare Reform

Page 12 GAO- 01- 828 Welfare Reform

Page 13 GAO- 01- 828 Welfare Reform

Page 14 GAO- 01- 828 Welfare Reform

Page 15 GAO- 01- 828 Welfare Reform

Page 16 GAO- 01- 828 Welfare Reform

Page 17 GAO- 01- 828 Welfare Reform

Page 18 GAO- 01- 828 Welfare Reform

Page 19 GAO- 01- 828 Welfare Reform

Page 20 GAO- 01- 828 Welfare Reform

Page 21 GAO- 01- 828 Welfare Reform

Page 22 GAO- 01- 828 Welfare Reform

Page 23 GAO- 01- 828 Welfare Reform

Page 24 GAO- 01- 828 Welfare Reform

Page 25 GAO- 01- 828 Welfare Reform

Page 26 GAO- 01- 828 Welfare Reform

Page 27 GAO- 01- 828 Welfare Reform

Page 28 GAO- 01- 828 Welfare Reform

Page 29 GAO- 01- 828 Welfare Reform

Page 30 GAO- 01- 828 Welfare Reform

Page 31 GAO- 01- 828 Welfare Reform

Page 32 GAO- 01- 828 Welfare Reform

Page 33 GAO- 01- 828 Welfare Reform

Page 34 GAO- 01- 828 Welfare Reform

Page 35 GAO- 01- 828 Welfare Reform

Page 36 GAO- 01- 828 Welfare Reform

Page 37 GAO- 01- 828 Welfare Reform

Page 38 GAO- 01- 828 Welfare Reform

Page 39 GAO- 01- 828 Welfare Reform

Page 40 GAO- 01- 828 Welfare Reform

Page 41 GAO- 01- 828 Welfare Reform

Page 42 GAO- 01- 828 Welfare Reform

Page 43 GAO- 01- 828 Welfare Reform

Page 44 GAO- 01- 828 Welfare Reform

Page 45 GAO- 01- 828 Welfare Reform

Page 46 GAO- 01- 828 Welfare Reform

Page 47 GAO- 01- 828 Welfare Reform

Page 48 GAO- 01- 828 Welfare Reform

Page 49 GAO- 01- 828 Welfare Reform

Page 50 GAO- 01- 828 Welfare Reform

Page 51 GAO- 01- 828 Welfare Reform

Page 52 GAO- 01- 828 Welfare Reform

Page 53 GAO- 01- 828 Welfare Reform

Page 54 GAO- 01- 828 Welfare Reform

Appendix I

Appendix I Scope and Methodology

Page 55 GAO- 01- 828 Welfare Reform

Appendix I Scope and Methodology

Page 56 GAO- 01- 828 Welfare Reform

Appendix I Scope and Methodology

Page 57 GAO- 01- 828 Welfare Reform

Appendix I Scope and Methodology

Page 58 GAO- 01- 828 Welfare Reform

Appendix I Scope and Methodology

Page 59 GAO- 01- 828 Welfare Reform

Appendix I Scope and Methodology

Page 60 GAO- 01- 828 Welfare Reform

Page 61 GAO- 01- 828 Welfare Reform

Appendix II

Appendix II California

Page 62 GAO- 01- 828 Welfare Reform

Appendix II California

Page 63 GAO- 01- 828 Welfare Reform

Appendix II California

Page 64 GAO- 01- 828 Welfare Reform

Appendix II California

Page 65 GAO- 01- 828 Welfare Reform

Appendix II California

Page 66 GAO- 01- 828 Welfare Reform

Appendix II California

Page 67 GAO- 01- 828 Welfare Reform

Appendix II California

Page 68 GAO- 01- 828 Welfare Reform

Appendix II California

Page 69 GAO- 01- 828 Welfare Reform

Page 70 GAO- 01- 828 Welfare Reform

Appendix III

Appendix III Colorado

Page 71 GAO- 01- 828 Welfare Reform

Appendix III Colorado

Page 72 GAO- 01- 828 Welfare Reform

Appendix III Colorado

Page 73 GAO- 01- 828 Welfare Reform

Appendix III Colorado

Page 74 GAO- 01- 828 Welfare Reform

Appendix III Colorado

Page 75 GAO- 01- 828 Welfare Reform

Appendix III Colorado

Page 76 GAO- 01- 828 Welfare Reform

Appendix III Colorado

Page 77 GAO- 01- 828 Welfare Reform

Page 78 GAO- 01- 828 Welfare Reform

Appendix IV

Appendix IV Connecticut

Page 79 GAO- 01- 828 Welfare Reform

Appendix IV Connecticut

Page 80 GAO- 01- 828 Welfare Reform

Appendix IV Connecticut

Page 81 GAO- 01- 828 Welfare Reform

Appendix IV Connecticut

Page 82 GAO- 01- 828 Welfare Reform

Appendix IV Connecticut

Page 83 GAO- 01- 828 Welfare Reform

Appendix IV Connecticut

Page 84 GAO- 01- 828 Welfare Reform

Page 85 GAO- 01- 828 Welfare Reform

Appendix V

Appendix V Louisiana

Page 86 GAO- 01- 828 Welfare Reform

Appendix V Louisiana

Page 87 GAO- 01- 828 Welfare Reform

Appendix V Louisiana

Page 88 GAO- 01- 828 Welfare Reform

Appendix V Louisiana

Page 89 GAO- 01- 828 Welfare Reform

Appendix V Louisiana

Page 90 GAO- 01- 828 Welfare Reform

Appendix V Louisiana

Page 91 GAO- 01- 828 Welfare Reform

Page 92 GAO- 01- 828 Welfare Reform

Appendix VI

Appendix VI Maryland

Page 93 GAO- 01- 828 Welfare Reform

Appendix VI Maryland

Page 94 GAO- 01- 828 Welfare Reform

Appendix VI Maryland

Page 95 GAO- 01- 828 Welfare Reform

Appendix VI Maryland

Page 96 GAO- 01- 828 Welfare Reform

Appendix VI Maryland

Page 97 GAO- 01- 828 Welfare Reform

Appendix VI Maryland

Page 98 GAO- 01- 828 Welfare Reform

Page 99 GAO- 01- 828 Welfare Reform

Appendix VII

Appendix VII Michigan

Page 100 GAO- 01- 828 Welfare Reform

Appendix VII Michigan

Page 101 GAO- 01- 828 Welfare Reform

Appendix VII Michigan

Page 102 GAO- 01- 828 Welfare Reform

Appendix VII Michigan

Page 103 GAO- 01- 828 Welfare Reform

Appendix VII Michigan

Page 104 GAO- 01- 828 Welfare Reform

Appendix VII Michigan

Page 105 GAO- 01- 828 Welfare Reform

Appendix VII Michigan

Page 106 GAO- 01- 828 Welfare Reform

Appendix VII Michigan

Page 107 GAO- 01- 828 Welfare Reform

Page 108 GAO- 01- 828 Welfare Reform

Appendix VIII

Appendix VIII New York

Page 109 GAO- 01- 828 Welfare Reform

Appendix VIII New York

Page 110 GAO- 01- 828 Welfare Reform

Appendix VIII New York

Page 111 GAO- 01- 828 Welfare Reform

Appendix VIII New York

Page 112 GAO- 01- 828 Welfare Reform

Appendix VIII New York

Page 113 GAO- 01- 828 Welfare Reform

Appendix VIII New York

Page 114 GAO- 01- 828 Welfare Reform

Appendix VIII New York

Page 115 GAO- 01- 828 Welfare Reform

Appendix VIII New York

Page 116 GAO- 01- 828 Welfare Reform

Page 117 GAO- 01- 828 Welfare Reform

Appendix IX

Appendix IX Oregon

Page 118 GAO- 01- 828 Welfare Reform

Appendix IX Oregon

Page 119 GAO- 01- 828 Welfare Reform

Appendix IX Oregon

Page 120 GAO- 01- 828 Welfare Reform

Appendix IX Oregon

Page 121 GAO- 01- 828 Welfare Reform

Appendix IX Oregon

Page 122 GAO- 01- 828 Welfare Reform

Appendix IX Oregon

Page 123 GAO- 01- 828 Welfare Reform

Appendix IX Oregon

Page 124 GAO- 01- 828 Welfare Reform

Appendix IX Oregon

Page 125 GAO- 01- 828 Welfare Reform

Page 126 GAO- 01- 828 Welfare Reform

Appendix X

Appendix X Texas

Page 127 GAO- 01- 828 Welfare Reform

Appendix X Texas

Page 128 GAO- 01- 828 Welfare Reform

Appendix X Texas

Page 129 GAO- 01- 828 Welfare Reform

Appendix X Texas

Page 130 GAO- 01- 828 Welfare Reform

Appendix X Texas

Page 131 GAO- 01- 828 Welfare Reform

Appendix X Texas

Page 132 GAO- 01- 828 Welfare Reform

Appendix X Texas

Page 133 GAO- 01- 828 Welfare Reform

Appendix X Texas

Page 134 GAO- 01- 828 Welfare Reform

Page 135 GAO- 01- 828 Welfare Reform

Appendix XI

Appendix XI Wisconsin

Page 136 GAO- 01- 828 Welfare Reform

Appendix XI Wisconsin

Page 137 GAO- 01- 828 Welfare Reform

Appendix XI Wisconsin

Page 138 GAO- 01- 828 Welfare Reform

Appendix XI Wisconsin

Page 139 GAO- 01- 828 Welfare Reform

Appendix XI Wisconsin

Page 140 GAO- 01- 828 Welfare Reform

Appendix XI Wisconsin

Page 141 GAO- 01- 828 Welfare Reform

Appendix XI Wisconsin

Page 142 GAO- 01- 828 Welfare Reform

Appendix XI Wisconsin

Page 143 GAO- 01- 828 Welfare Reform

Page 144 GAO- 01- 828 Welfare Reform

Appendix XII

Appendix XII Survey Instrument

Page 145 GAO- 01- 828 Welfare Reform

Appendix XII Survey Instrument

Page 146 GAO- 01- 828 Welfare Reform

Appendix XII Survey Instrument

Page 147 GAO- 01- 828 Welfare Reform

Appendix XII Survey Instrument

Page 148 GAO- 01- 828 Welfare Reform

Appendix XII Survey Instrument

Page 149 GAO- 01- 828 Welfare Reform

Appendix XII Survey Instrument

Page 150 GAO- 01- 828 Welfare Reform

Appendix XII Survey Instrument

Page 151 GAO- 01- 828 Welfare Reform

Appendix XII Survey Instrument

Page 152 GAO- 01- 828 Welfare Reform

Appendix XII Survey Instrument

Page 153 GAO- 01- 828 Welfare Reform

Page 154 GAO- 01- 828 Welfare Reform

Appendix XIII

Appendix XIII Comments From the Department of Health and Human Services

Page 155 GAO- 01- 828 Welfare Reform

Appendix XIII Comments From the Department of Health and Human Services

Page 156 GAO- 01- 828 Welfare Reform

Appendix XIII Comments From the Department of Health and Human Services

Page 157 GAO- 01- 828 Welfare Reform

Appendix XIII Comments From the Department of Health and Human Services

Page 158 GAO- 01- 828 Welfare Reform

Appendix XIII Comments From the Department of Health and Human Services

Page 159 GAO- 01- 828 Welfare Reform

Ordering Information The first copy of each GAO report is free. Additional
copies of reports are $2 each. A check or money order should be made out to

the Superintendent of Documents. VISA and MasterCard credit cards are
accepted, also. Orders for 100 or more copies to be mailed to a single
address are discounted 25 percent.

Orders by mail: U. S. General Accounting Office P. O. Box 37050 Washington,
DC 20013

Orders by visiting: Room 1100 700 4th St. NW (corner of 4th and G Sts. NW)
U. S. General Accounting Office Washington, DC

Orders by phone: (202) 512- 6000 fax: (202) 512- 6061 TDD (202) 512- 2537

Each day, GAO issues a list of newly available reports and testimony. To
receive facsimile copies of the daily list or any list from the past 30
days, please call (202) 512- 6000 using a touchtone

phone. A recorded menu will provide information on how to obtain these
lists.

Orders by Internet: For information on how to access GAO reports on the
Internet, send an e- mail message with ?info? in the body to:

info@ www. gao. gov or visit GAO?s World Wide Web home page at: http:// www.
gao. gov

To Report Fraud, Waste, or Abuse in Federal Programs

Contact one:  Web site: http:// www. gao. gov/ fraudnet/ fraudnet. htm  e-
mail: fraudnet@ gao. gov  1- 800- 424- 5454 (automated answering system)

United States General Accounting Office Washington, D. C. 20548- 0001

Official Business Penalty for Private Use $300

Address Correction Requested Presorted Standard

Postage & Fees Paid GAO Permit No. GI00
*** End of document. ***