Number A-1 in Series, Assessing New Federalism: Issues and Options for States
The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders.
The welfare initiative President Clinton signed into law in August 1996
1 replaces Aid to Families with Dependent Children (AFDC) with Temporary Assistance to Needy Families (TANF). As a result, the federal responsibility to match state expenditures on cash assistance to low-income families with children has become a fixed block grant to states with certain requirements focusing on a new philosophy of work rather than dependency.
How states respond to the challenges and opportunities presented by TANF is likely to differ substantially, because the state AFDC programs that TANF is replacing vary more widely than is generally understood. This variation across states has increased in recent years, as particular states have chosen to use waivers from federal requirements to test the effects of a variety of deviations from federal AFDC rules.2
This brief depicts the diversity of state AFDC programs prior to passage of the new legislation, as a guide in assessing the potential implications of TANF for different states. We begin with a brief overview of two major dimensions along which pre-TANF differences will affect how states fare under the new legislation.
Implications of AFDC Funding Levels and New Program Requirements
TANF differs from the federal provisions of the AFDC program in two ways that are particularly important in assessing how states will fare as they change from AFDC to TANF rules.
Funding Levels. Under TANF, the federal government provides each state an annual block grant, which is fixed for the next six years on the basis of that state's AFDC spending prior to the new law.3 Since states have always had the power to set benefit levels and income thresholds for AFDC eligibility and to define certain coverage parameters within federal guidelines, AFDC program expenditures span a wide range.
Since the TANF block grants depend on a state's prior AFDC spending level, the higher the historic spending level the more generous the TANF benefit will be and the greater fiscal leeway the state will have.
Program Requirements. The new legislation restricts state use of federal TANF funds in ways designed to encourage work and discourage long-term welfare dependency:
- Families may not receive assistance for more than five years during their lifetime, with up to 20 percent of the caseload potentially exempt for reasons of hardship.
- The percentage of TANF families who are working must increase over time, and virtually all TANF families must work after receiving benefits for two years.
- Each state must maintain 80 percent of its own recent AFDC spending level (which can drop to
75 percent if the state meets the new federal work participation targets).
Although these provisions are new at the federal level, states have already implemented a variety of AFDC program changes designed to increase work effort. These include triggering work requirements after two or three years, implementing strict sanctions that reduce or eliminate benefits for parents who fail to comply with program rules, imposing time limits, and encouraging work by increasing the amount of other income families may keep without losing benefit eligibility.4
States that have used waivers to make progress in replacing welfare with work will have less trouble meeting TANF's time limits and work requirements in the short run, though not necessarily over the longer term.
Caseload Size, Program Costs, and Financing
State AFDC caseload size and costs spanned a wide range in 1995 (table 1).
Caseload Size
Monthly caseloads varied from lows of about 5,000 in states like Wyoming and the Dakotas to over 900,000 in California. Nationwide, about half of all persons in poor families with children received AFDC benefits, with the share varying from 21 percent at the low end (in states such as Alabama, Arkansas, and Idaho) to more than 80 percent at the high end (in states such as Alaska, Hawaii, and Rhode Island).
Program Costs
The federal government spent $12 billion for AFDC cash benefits in 1995, compared to the states' total share of $10 billion. Federal expenditures accounted for 54 percent of total AFDC costs, varying from a statutory minimum of 50 percent (14 states), to a maximum of 79 percent, with less affluent states having a higher federal match. Monthly AFDC benefits per family averaged $381 for the nation as a whole, varying from a low of $120 per month in Mississippi to about $550 per month in California and New York and even higher amounts in Alaska and Hawaii. The two states with the highest expenditures California and New York both had relatively large caseloads and relatively high average monthly benefits. They accounted for 38 percent of federal spending in 1995, even though they had the minimum 50 percent federal match rate.
How states financed the state/ local share of their AFDC costs also varied. Eleven states required their localities to share the cost of benefits (California, Colorado, Indiana, Minnesota, Montana, New Jersey, New York, North Carolina, North Dakota, Ohio, and Wisconsin). The local share among these states (not shown) ranged from a low of 4 percent (Ohio) to a high of 50 percent (New York and North Carolina). Local governments in 18 states shared the cost of benefit administration in 1995.
Financing Implications
Since caseloads and costs have been declining, current expectations for 1997 indicate that the federal block grants will be more than enough to finance the federal share of cash assistance at pre-1996 levels in nearly all states representing a "bonus" or windfall gain to states at least in the first year of implementation.5 Since TANF legislation freezes current differences across states in federal spending on low-income families with children at least for the next six years,6 some states will continue to get more federal resources per recipient than others, with the affluent states that have traditionally paid higher benefits receiving larger shares.
States will be free to change benefit levels, however, and could distribute the federal dollars across beneficiaries in new ways under TANF. In the longer run, available resources per TANF family could be strongly affected by changes in caseloads. The amount of federal resources per family under TANF obviously will increase in states that reduce caseloads relative to current levels and decrease in states where caseloads increase.
Caseload Characteristics
The type of family receiving AFDC and how often members of the family combine earnings with welfare and other income are two other factors that will affect states as they respond to TANF.
Family Composition
Historically, AFDC has covered three types of needy families: a single parent with one or more children, two parents with children (when the primary wage earner was either disabled or unemployed), and children living with no eligible adult.
Overall, 75 percent of AFDC families were headed by a single adult in 1995; only 7 percent had two adults; and 18 percent of the families had no eligible adult (table 2). The share of families headed by a single adult ranged from a low of 65 percent in Alabama to a high of 86 percent in Rhode Island. Many states had extremely low proportions of two-parent families (around 1 percent of the caseload in Alabama, Delaware, the District of Columbia, Georgia, Mississippi, Oklahoma, and South Dakota) compared with states at the high end of the two-parent range with over 10 percent (Alaska, California, Hawaii, Iowa, Montana, New York, Vermont, Washington, and West Virginia).
The share of AFDC families with no eligible adult in 1995 also spanned a wide range. In seven states (Alabama, Arizona, Arkansas, Louisiana, Nebraska, Nevada, and South Carolina) it exceeded 25 percent of the caseload. In the seven states at the other end of the range, it was 10 percent or below (Alaska, Hawaii, Maine, Minnesota, Montana, Rhode Island, and Vermont).
Both the share of child-only units (units with no eligible adults) and the share of two-parent units will affect a state's ability to avoid financial penalties by meeting TANF work requirements.7
Child-only units are important because states must demonstrate that 25 percent of TANF families are working in 1997, and 50 percent by 2002. Since child-only units are exempt from these work requirements, they do not count in the states' caseload base used to calculate work participation rates. Thus, states with equal caseload sizes but relatively more child-only cases will be required to move fewer adults into employment activities in order to meet federal targets. Whether the states with the highest proportions of child-only units in 1995 will be those that have the largest child-only share under TANF depends, of course, on whether and how eligibility rules change.
Four types of families account for the vast majority (85 percent) of child-only units: families without parents in which the adult caretaker is not eligible for benefits, 39 percent; families in which the parent is on SSI, 21 percent; families in which the parent is an illegal immigrant but the children are citizens, 15 percent; and families in which the parent has been removed from the unit as a sanction for failure to meet program requirements, 10 percent (chart 1).
Two-parent families are important because states must reach higher targets for two-parent families, with 75 percent working in 1997 and 90 percent by 2002. This implies that states with more two-parent families will have to either meet tougher work requirements for a larger share of their caseloads or restrict their two-parent eligibility policies.
Family Earnings
The family earnings profile of a state's AFDC caseload provides further insight into how states may fare in meeting the TANF work requirements. States with relatively large percentages of their caseloads already working may have an easier time than other states in the short run. In the longer run, however, as participants with more education and recent work experience move increasingly from TANF to work, the opposite may be true, as the share of those remaining in the caseload who are "harder to employ" families increases.
Nationally, only 10 percent of AFDC families had some earnings in 1995 (table 2). However, states ranged from highs of more than 20 percent with earnings (Iowa and Michigan) to lows of less than 4 percent (Alabama, the District of Columbia, Louisiana, Nevada, New Jersey, and West Virginia).
Variations in the percent of the caseload with earnings are due to substantial differences in states' AFDC policies. Prior to the early 1990s, federal rules defined the amount of earnings that could be omitted from the benefit calculation in most states.8 But 15 states had "fill-the-gap" budgeting policies that allowed families to keep a larger share of earnings without a reduction in their AFDC payment, to make up for maximum benefit levels that were below what the state regarded as necessary to cover minimal needs. In addition, expansions of the amount of earnings recipients were allowed to keep were popular features of states' waiver policies. This type of waiver was in effect in 17 states during 1995, either statewide or in selected counties.
States with relatively high shares of AFDC beneficiaries with earnings had either fill-the-gap budgeting policies (for example, Maine and Wyoming) or waivers from federal rules defining earnings disregards (Iowa and Michigan). (Fill-the-gap and waiver states as of 1995 are shown in table 3.)9
Other Income
Other income is primarily child support. Under AFDC federal rules families could retain up to $50 a month of this income, with the remainder going to the state to offset AFDC costs. Under waivers, several states allowed families to retain a larger share of child support payments without losing eligibility. In 1995 the proportions of state caseloads with other income ranged from highs of over 20 percent (Kentucky, Maine, Maryland, and Ohio) to lows of under 4 percent (Arizona, District of Columbia, Illinois, and Nevada).
Program Generosity
AFDC program generosity significantly affects AFDC families' incomes. States vary considerably in their maximum payments and in the maximum income allowed before losing eligibility.
Average Family Incomes
Overall, families on AFDC in 1995 averaged cash incomes equal to 41 percent of the poverty line (table 3).10 At the high end, in 4 states AFDC families' average incomes reached 60 percent of poverty or above (Alaska, California, Connecticut, and Hawaii). At the low end, in 15 states, average incomes for AFDC families were only 30 percent of the poverty threshold or below (Alabama, Arkansas, Florida, Indiana, Kentucky, Louisiana, Mississippi, North Carolina, Missouri, Oklahoma, South Carolina, Tennessee, Texas, Virginia, and West Virginia).
Program Rules
Reflecting the relatively low percentages of AFDC families with earnings, families' average incomes tended to mirror their states' maximum AFDC benefit levels. For the nation, maximum benefit levels averaged 42 percent of the poverty line. Families in states that had policies allowing them to retain a larger share of other income were somewhat more likely to have average incomes above maximum benefit levels.
The maximum income allowed before families became ineligible for any AFDC benefit averaged 67 percent of the poverty line for the nation. But there was considerable variation in maximum income levels, even for states whose maximum benefit levels were fairly similar. For example, Kentucky's maximum benefit level was 22.5 percent of the poverty threshold, but families could have incomes as high as 59 percent of poverty before losing eligibility. In contrast, although Arkansas's maximum benefit was very similar, at 20 percent of the poverty line, families could reach income levels of only 28 percent of the poverty threshold before losing eligibility. This explains why AFDC family income in Kentucky averaged 30 percent of the poverty threshold, compared with only 20 percent of poverty for families in Arkansas, even though AFDC benefit levels in the two states were similar.
Implications
What does this variation suggest for TANF implementation? All states will be required to move larger percentages of their caseloads into work. This could, in itself, increase incomes for families on TANF.
On the other hand, the lifetime time limit may discourage families from participating in TANF whenever they can conceivably get along on other income from working or from child support, for example. Each month on TANF will count toward the lifetime time limit, even when TANF benefits are simply providing a supplement to other income. Thus, families may choose to save their TANF eligibility for periods when they have absolutely no other income opportunities. To the extent that they follow this pattern, the average income levels of families on TANF may fall, even if the maximum benefit levels and income eligibility limits remain unchanged.
Major Messages
Perhaps the most important message from this review is fiscal. Since the new block grant world bases state funding allocations on prior federal AFDC spending in a state, states with low current benefit levels will have the hardest time responding to the new reform legislation. They will have fewer federal resources per family to prepare recipients for moving into jobs than other states. In addition, even if those states are able to reduce caseloads, the bonus or windfall gain arising from the fixed six-year block grant amount will be relatively small.
The second message is that states with relatively larger shares of child-only cases will have an easier time meeting work participation requirements, other things equal, because these cases are excluded from the work target calculation.
For states that have moved the furthest toward the new philosophy of replacing welfare with work, the future is mixed. In the initial implementation of TANF, these states will find it easier to reach the work targets because they tend to have larger shares of their caseload already working. Later, however, their early success will leave them with larger shares of hard-to-employ adults, making TANF's increasingly strict work participation requirements harder to reach.
Notes
The authors would like to thank L. Jerome Gallagher for outstanding research assistance in preparing this brief.
1. The Personal Responsibility and Work Opportunity Reconciliation Act of 1996.
2. TANF permits states to continue their waiver programs until they are scheduled to expire, even if they are at odds with TANF requirements.
3. Each state gets the federal share of its AFDC spending either averaged over 1992 1994, or in 1994, or in 1995, whichever is highest.
4. States also vary considerably in the employment and training opportunities provided for parents in the unit, a topic discussed in later briefs.
5. New Mexico is the one exception, where poverty rates and caseloads have continued to increase, implying higher total assistance spending if benefits are retained at pre-TANF levels.
6. Exceptions could occur, since the federal legislation provides a small contingency fund for higher than expected spending needs, to be allocated among states on a discretionary basis. This fund is capped at $2 billion for all states over a five-year period, about 3 percent of expected federal spending on TANF over that period. Supplemental funds allow a 2.5 percent per state adjustment (up or down) in federal dollars to adjust for differential growth rates.
7. Single parents with children under age 6 meet the requirement by working 20 hours per week. Others can meet the requirement with 20 hours per week of activity beginning in 1996 increasing to 30 hours per week by 2000. Two-parent families must work 35 hours per week. Activities may include subsidized or unsubsidized employment, on-the-job training, community services, and up to 12 months of vocational training. Future policy briefs will provide further discussion of work requirements.
8. Families with earnings could disregard up to $90 per month for work-related expenses and up to $200 per month per child for child care expenses depending on the child's age. Also, the first $30 of earnings per month was disregarded for 12 months, and one-third of the remainder was disregarded for the first 4 months of any employment.
9. Most of those states that had statewide waiver policies in effect in 1995 included stronger work incentives than were allowed under the federal AFDC rules. A future policy brief will summarize the status of all waiver provisions in effect prior to the federal welfare reform legislation.
10. Nearly all (90 percent) of AFDC families also received food stamps in 1995 and 22 percent had some type of housing assistance. These noncash federal benefits raised the effective level of income available to AFDC families above what is shown in table 3.
Tables & Charts
Source: Urban Institute 1997, from Administration for Children and Families - 241 AFDC 5 Year Line by Line Report and 1995 AFDC Administrative Data (Administration for Children and Families, Department of Health and Human Services).
a. Unless otherwise noted, data in this table are for the federal fiscal year.
b. Percent is calculated as number of persons receiving AFDC benefits in the average month of the calendar year, relative to total number of poor persons in families with children. The number of poor persons in families with children in each state uses an average poverty rate calculated from annual incomes in the March 1994, 1995, and 1996 Current Population Surveys, applied to the 1995 population.
c. Excludes administrative costs of the program.
d. States that require local areas to fund part of the benefits. Many other local areas share the cost of benefit administration.
Source: Urban Institute 1997, from 1995 AFDC Quality Control Survey.
a. This group includes a very small number of pregnant women with no children, eligible at state option. (They comprise about 1% of the national caseload.)
Source: Urban Institute 1997, from 1995 AFDC Quality Control Survey, HHS documents, CRS benefits survey, and other sources.
a. For a family of three. Benefit levels relative to poverty vary across family size.
b. For a family of three. The maximum income a family can receive and remain eligible for AFDC after one year of working.
c. "Fill-the-Gap" states are those with a maximum payment that is less than the payment standard.
d. Waiver implementation status by end of 1995. 0=no waiver, 1=less than statewide, 2=statewide (though some provisions may be less than statewide.)
e. States whose waiver included an expansion of earnings disregards for all units.
About the Authors
Sheila Zedlewski is director of the Urban Institute's Income and Benefits Policy Center. Her areas of special interest include income security, health benefits, retirement, aging, and taxes.
Linda Giannarelli is a senior research associate at the Urban Institute. Her special interests include microsimulation, in-kind benefits, and the distributional effects of taxes and transfers.
PDF