Introduction

The decline in the number of families and individuals receiving family cash assistance is a distinctive characteristic of the period after enactment of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (P.L. 104-193, PRWORA, also referred to as the 1996 welfare reform law). PRWORA ended Aid to Families with Dependent Children (AFDC), a program that provided dedicated funding for cash assistance to needy families with children. The law folded funding for such assistance into a newly created, broad-purpose Temporary Assistance for Needy Families (TANF) block grant. The TANF block grant helps fund state-designed and state-run family cash assistance programs for needy families with children, but states may also use their funds for a wide range of benefits and services to address both the effects and root causes of childhood economic disadvantage.¹

Family cash assistance, sometimes called welfare, is generally a benefit of last resort and provided only to families with very low or no cash income. Historically, family cash assistance was most often provided to families headed by a single mother, though in the post-PRWORA period an increased share of families receiving cash assistance represent those with nonparent relatives (e.g., grandparents, aunts, uncles).

In 1994, before PRWORA's enactment, the number of cash assistance recipients reached its historical peak. An estimated 18.0 million people (including children, parents, and nonparent caretakers) received assistance at some point in that year. In 2018, an estimated 3.8 million received assistance during the year. Research has found that this reduction resulted from both faster exits from state assistance programs and fewer entries into them.²

Previous CRS reports have examined the decline in the cash assistance caseload, putting it in the context of the estimated number of people eligible for assistance both pre- and post-PRWORA.³ In 2018, the 3.8 million people who were estimated to receive benefits at some point in that year represented 26% of the estimated 14.3 million who were eligible for family cash assistance in that year. This report extends previous CRS analyses in two main ways. It examines

• the number of people eligible for and receiving assistance in each year from 1994 to 2018, highlighting the variation in the trends that are the result of cycles of economic growth and recession during this period; and
• the variation by state in the rate at which eligible individuals received benefits, and how this has changed from the mid-1990s.

This analysis is based on estimates from the Census Bureau's Annual Social and Economic Supplements (ASEC) to the Current Population Survey (CPS) and the Transfer Income Model version 3 (TRIM3).⁴ TRIM3 is funded primarily by the U.S. Department of Health and Human Services (HHS), and maintained at the Urban Institute. For more technical details, see the Appendix. The analysis shows eligibility and receipt through 2018. Comparable data are not available for the recent period that included the COVID-19 pandemic and its economic fallout.

Differing Rates of Receipt Among Need-Tested Programs

All major need-tested programs have receipt rates of less than 100%. Benefit receipt among eligible individuals is not automatic. Typically, an individual must apply for benefits and provide a government agency with information about their finances, family structure, and sometimes other circumstances. That information is taken in by the administering agency and then initial eligibility is determined and benefits authorized. This process usually must be repeated periodically as eligibility and benefits are recertified. Additionally, some programs might condition eligibility on certain behaviors, such as work or participation in job preparation activities. Recipients may be required to document such activity.

Economists have described an eligible individual's decision to receive benefits as one that weighs the costs versus the gains of that benefit receipt.⁵ That is, does the benefit provided by a program outweigh the administrative burden (i.e., costs) of completing the application, certification, and recertification processes and remaining in and proving compliance with program rules? Additionally, lack of information, stigmas related to benefit receipt, and other factors may be important for influencing program participation.

Different programs have different receipt rates.⁶ Medicaid and the State Children's Health Insurance Program (CHIP), as well as the Supplemental Nutrition Assistance Program (SNAP), tend to have comparatively high receipt rates: an estimated 75% of those eligible for Medicaid/CHIP and 63% of those eligible for SNAP in 2018. Medicaid and SNAP are entitlement programs, with those eligible for benefits being entitled to them, and have no explicit caps in federal funding of them.

Other programs, such as child care subsidies from the Child Care and Development Fund (CCDF) and housing subsidies, tend to have comparably low receipt rates among those eligible: 15% and 22%, respectively, in 2018. In addition to eligible individuals' choices in taking up benefits, the nature of federal funding also affects receipt rates, as some programs (including CCDF and housing subsidies) receive capped federal funding and sometimes have waiting lists for those eligible to receive benefits.

The 26% receipt rate among those eligible for family cash assistance under TANF tends to be on the low end of the range of rates for need-tested programs. It has also declined from the receipt rate for family cash assistance from AFDC before the enactment of PRWORA. In 1994, the receipt rate among eligible individuals for AFDC was 79%.

PRWORA Ends AFDC and Creates TANF

The modern form of family cash assistance has its origins in the early 1900s, with state and locally run mother's pensions programs that provided aid to single mothers, usually widows. The Social Security Act of 1935—enacted in the midst of the Great Depression—provided federal funding for state family cash assistance by creating the Aid to Dependent Children (ADC) program, later renamed AFDC. AFDC's stated purpose was to provide aid to mothers of children whose family had lost the support of the father because of death, disability, or absence, so that the mother could stay at home and rear her children.⁷ Family cash assistance was seen at the time as a way to keep children in their homes, and an alternative to placing children in institutional foster care.

AFDC provided dedicated federal funding for cash assistance for needy families and was eventually considered an entitlement to individuals who were eligible for benefits under state-determined rules. The entitlement meant that those eligible had a legal right to AFDC benefits, and that procedural due process was required to terminate them.⁸ Funding was provided on an open-ended, unlimited basis. That is, the federal government reimbursed states for a share of their expenditures on AFDC family cash assistance without limit.⁹

Support for the original view that single mothers should be permitted to stay home and out of the workforce eroded by the 1960s.¹⁰ AFDC raised concerns that providing assistance without work reduced work effort and led to more children raised outside of married couple families. Research generally found that the pre-PRWORA system of need-tested aid did reduce work, but its impact on family composition was less clear.¹¹ Research also found that while many AFDC families received benefits for a short period of time (e.g., less than one year), some who received assistance did so for long periods of time.¹²

Daniel Patrick Moynihan—a Department of Labor official during the Johnson Administration, an advisor to President Nixon, and later a U.S. Senator—argued that family cash assistance was about dependence, as distinguished from poverty.¹³ Throughout the four-decade debate, the issue of dependence on assistance was raised and the term was used often—though it was rarely precisely defined. Historically, the number of people and families receiving assistance was viewed as the preeminent indicator of dependence, though mere receipt of benefits might not capture all that is meant or implied by the term.¹⁴

Proposals to end AFDC and replace it with a new system of family support were debated over four decades from the 1960s to the mid-1990s. A first set of proposals was presented in the 1960s during the first large increase in the number of families receiving AFDC assistance in the late 1960s and into the early 1970s.¹⁵ These proposals were not enacted.

PWRORA was enacted in 1996 following then-candidate Bill Clinton's promise during the 1992 presidential campaign to "end welfare as we know it." That campaign occurred as the number of families receiving AFDC was increasing again, eventually reaching its all-time high in 1994.¹⁶ It was the record number of families receiving AFDC in 1994 that helped set the stage for the 1995 and 1996 debates that led to the enactment of PRWORA.

In creating TANF, PWRORA ended dedicated and open-ended funding for family cash assistance, ended the entitlement of individuals to it, and time-limited family cash assistance from federal funds to 60 months for families with a head of household receiving benefits. The law also revised the rules relating to work and participation in job preparation activities, altering requirements on states with respect to work participation, including allowing states to either partially or fully meet their work participation requirement through reducing the number of families receiving assistance (a caseload reduction credit against the state's work standard).¹⁷ PRWORA also altered work participation requirements that apply to individuals, including allowing states to end benefits entirely to a family (rather than reducing benefits) if an individual in it refused to comply with the requirements.

A statutory goal of TANF is to "end the dependence of needy parents on government benefits by promoting job preparation, work, and marriage."¹⁸ Another of TANF's statutory goals, though, harkened back to the original goal of family cash assistance: "provide assistance to needy families so that children may be cared for in their own homes or in the homes of relatives."¹⁹

Declines in the Receipt Rate Among Individuals Eligible for Family Cash Assistance

The estimates of the number of people eligible for family cash assistance are based on those in families

• that meet family composition requirements (i.e., having a child); and
• have their incomes fall below state-specified income eligibility thresholds, and if applicable in their state, have estimated countable assets below a state-specified threshold.²⁰

Figure 1 shows the estimated numbers of people eligible for and receiving assistance (left panel), as well as the receipt rate among those eligible (right panel), for each year from 1994 to 2018. The span from 1994 to 2018 covers three periods of economic growth and two recessions. As noted previously, the number of people receiving family cash assistance was at its all-time high in 1994.

The figure shows a general pattern of fluctuation in the number of people eligible for family cash assistance, associated with the differences in the state of the national economy over the years. It also shows an almost continuous decline in the number of people actually receiving benefits. Thus, the receipt rate among those eligible generally declined since 1994.

Particularly notable is the period of recession associated with the 2008 global financial crisis (the Great Recession). The number of people eligible for family cash assistance increased, almost reaching pre-PRWORA levels. The number of people receiving family cash assistance also increased, but that increase was a fraction (about 30%) of the increase in the number of eligible people in this period. In 2009, the number of people eligible for assistance but not receiving it was at its highest level for the 1994-2018 period. In that year, 14.7 million individuals were in families estimated to be eligible for assistance but did not receive it.

During the long period of economic growth that began in 2009, the numbers of people eligible for and receiving assistance declined. The receipt rate also declined during this period, albeit at a slower rate than during the previous economic expansions. By 2018, the receipt rate had fallen to 26%—and the number of people eligible for assistance at some point in the year but not receiving it totaled 10.5 million.²¹

Figure 1. Estimated Numbers of People Eligible For and Receiving Family Cash Assistance, and Receipt Rate: 1994-2018

Source: Congressional Research Service (CRS), based on data from the U.S. Census Bureau's Annual Social and Economic Supplement (ASEC) to the Current Population Survey (CPS) and the TRIM3 microsimulation model.

Comparable information is not available for the recent period that included the economic effects of the COVID-19 pandemic. However, the number of people receiving cash assistance did not increase during 2020 or 2021.

State Variations in Assistance Receipt Rates Among Eligible Individuals

The only federal eligibility rules for TANF-funded family cash assistance are that there be a child present in the family and that the state apply a test of financial need to the family to qualify it for benefits. Other rules, including the specific test of financial need, are determined by the states. States also set additional conditions for receipt of benefits, such as the work requirements that apply to individuals and the sanctions for failure to comply with them.

Figure 2 shows the receipt rate among those eligible for family cash assistance for the 50 states and District of Columbia. The figure shows states ranked from the highest to the lowest percentage of receipt among eligible individuals. It shows this percentage for the three-year average from 2016 to 2018.

A three-year average is used to provide more reliable estimates of the TANF receipt rate at the state level than is possible with a single year's data. The sampling error associated with these estimates is larger in small population jurisdictions than in states with larger populations even with three years of data. Additionally, some of the simplifying assumptions needed to model eligibility and receipt might affect some states more than others.

The figure shows that the estimated rate of receipt among eligible individuals was over 50% in four jurisdictions: California, the District of Columbia, New York, and Oregon. The estimated rate was below 10% in seven states: Georgia, Idaho, Indiana, Louisiana, North Carolina, South Carolina, and Wyoming.

Figure 2. Estimated TANF Receipt Rates Among Eligible Individuals by State: Three-Year Average, 2016-2018

Source: Congressional Research Service (CRS), based on data from the U.S. Census Bureau's Annual Social and Economic Supplement (ASEC) to the Current Population Survey (CPS) and the TRIM3 microsimulation model. Notes: Estimates for larger population states tends to be more reliable than those for the smaller states.

States' Benefit Amounts Influence Assistance Receipt Rates

In addition to determining eligibility thresholds for family cash assistance, states also determine the amount of benefits a family is eligible to receive. The benefit amounts vary by state.

Figure 3 shows the relationship between the estimated receipt rate among people eligible for family cash assistance and the TANF maximum benefit for a family of two. It is based on the three-year average from 2016 to 2018. The dashed line in the figure summarizes the statistical relationship between the receipt rate and maximum benefit. It shows, with a few exceptions, that states with lower maximum benefits tend to have lower receipt rates, and those with higher maximum benefits tend to have higher receipt rates. This suggests that in lower benefit states, families are more likely to consider that the costs of applying for and continuing to receive assistance outweighs the economic benefit of doing so. Conversely, in higher benefit states the higher benefits counter-balance the costs of applying for and receiving benefits. However, at higher benefit amounts there are larger differences among the states in receipt rates. The variation in benefit amounts is associated with some – but not all – of the variation in the rate at which eligible individuals receive benefits.²²

Figure 3. Relationship Between States' Maximum Benefit Levels and the Rate at Which Eligible Individuals Receive Assistance Receipt Rates and Maximum Benefits Are Averages for 2016 to 2018

Source: CRS analysis of the U.S. Census Bureau's Annual Social and Economic Supplement (ASEC) to the Current Population Survey (CPS) and the TRIM3 microsimulation model.

Differences by State in Declining Receipt Rates Among Eligible Individuals

The intent of the 1935 legislation to permit single mothers to stay out of the workforce to raise their children clashed with the reality that many single mothers did work, even at that time. This was particularly true for African-American women.²³ That goal was met with resistance in some states. Multiple states, particularly in the South, adopted policies barring aid to certain families: those with children born out of wedlock and households where a man was present. Some states also adopted so-called farm policies, requiring able-bodied mothers and children over a certain (state-defined) age to work in seasonal agricultural employment.²⁴

A series of federal administrative, legislative, and U.S. Supreme Court decisions ended some of the state practices that removed families from cash assistance.²⁵ On the eve of PRWORA's enactment, family cash assistance was generally considered an entitlement to individuals—those persons eligible based on state rules had legal protections of an entitlement to a benefit.

Figure 4 compares the family cash assistance receipt rates among eligible recipients pre- and post-PRWORA by comparing three-year averages for 1994 to 1996 and 2016 to 2018. The receipt rate declined in all jurisdictions. ²⁶

Before PRWORA, there was variation in the AFDC receipt rates among the states. However, the range of these rates was narrower under AFDC than under TANF. In the three years from 1994 to 1996, estimated receipt rates varied by 33 percentage points between the state with the highest rate and the state with the lowest. This compares with the 2016-2018 range of 58 percentage points between the state with the highest rate and the state with the lowest.

Some states that had relatively high receipt rates among eligible individuals in 1994-1996 experienced some of the largest declines in rates since then. Michigan's receipt rate in 2016-2018 was 77 percentage points lower than it was in 1994-1996. In 1994-1996, it was ranked at the top, with an estimated receipt rate among eligible individuals at 93%. In 2016, it had a receipt rate of 16%, dropping it to the middle of the receipt rate distribution among the states (see Figure 2).

Figure 4. Receipt Rates Among Eligible Individuals by State: Three-Year Average for 1994-1996 Compared with 2016-2018 (states ranked in the order of their 1994-1996 receipt rates)

Source: Congressional Research Service (CRS), based on data from the U.S. Census Bureau's Annual Social and Economic Supplement (ASEC) to the Current Population Survey (CPS) and the TRIM3 microsimulation model.

Policy Implications

TANF policy currently gives states incentives to reduce the number of people receiving assistance, whether or not they improve the financial circumstances of individuals who meet the eligibility criteria for family cash assistance. TANF also affords states the flexibility to act on these incentives.

TANF's incentives to reduce the number of families receiving assistance include the following:

• The ability to use savings from fewer families receiving assistance to fund a wide range of other activities. In FY2021, basic assistance accounted for 22.6% of all TANF and MOE funds used; this was a decline from 70.7% of AFDC and related spending used for basic assistance in FY1995. Much of the remainder of TANF and MOE funding was spent on a wide range of activities, such as child care, pre-Kindergarten services, and services for families where children are at risk of removal from their homes because of neglect or abuse. These activities also served a broader range of families than did family cash assistance.
• Permitting states to meet their work requirement through fewer families receiving assistance. For FY2020, 26 states achieved their main work requirement solely through reducing the number of families receiving assistance.²⁷ They did not need to engage a single work-eligible member of a family receiving assistance in either work or job preparation to meet their standard.

Is TANF Meeting Its Goals?

TANF's statute explicitly defines its overall purpose as "increasing state flexibility" to meet four goals. One of the goals is ending dependence on assistance; another is providing assistance so that children may remain in their homes.²⁸ The tension between these two goals has played out differently in different states, though in all states the number of families receiving assistance and the receipt rate among individuals eligible for TANF was lower in 2018 than in 1994.

The creators of TANF envisioned a reduced role for family cash assistance. As noted above, one of TANF's statutory goals is to "end dependence on government benefits." Additionally, a committee report that is part of TANF's legislative history, supporting the idea that a state could meet its work requirements through fewer people receiving family cash assistance, said:

This provision, in effect, provides States with the ability, and in fact the incentive, to do away with the concept of measuring participation rates, (which is by and large a "process" measurement), and move toward having their performance based on a true outcome—a reduction in welfare dependency, a goal in which no one can argue.²⁹

The decline in the number of families receiving assistance after PRWORA was used repeatedly as an indicator of the success of that law.³⁰ However, over time, there was an increase in receipt of need-tested aid from programs other than family cash assistance. CRS estimates that in 2017, 57% of children were in families that received some form of need-tested aid (commonly, medical care from Medicaid or the State Children's Health Insurance Program, nutrition assistance from SNAP, or refundable tax credits).³¹ Of children receiving benefits, 38 million out of 42 million were in families with an adult worker. The high rate of receipt of need-tested benefits, though not family cash assistance, by families with children raises questions about whether or not the goal of ending dependence on government benefits is being achieved. It also points to the vague nature of the term dependence, as much of the help received supplements earnings.

Another statutory goal of TANF is to provide assistance to needy families. The legislative history of TANF includes the following statement:

Destroying the narcotic effect of welfare while preserving its function as a safety net for families experiencing temporary financial problems is the major intent of the Committee bill. Based on the fact that it is precisely the permanent guarantee of benefits that induces dependency, the Committee fundamentally alters the nature of the AFDC program by making its benefits temporary and provisional.³²

However, many people eligible for assistance—in many states, most eligible people—do not receive assistance at all, even on a temporary basis.

The Broader Context Related to Child Poverty

There is a broader context for the diminished role of family cash assistance. In 1994, the child poverty rate, when measured inclusive of taxes and noncash benefits and anchored to a fixed standard of living, was 25%.³³ In 2009, when the gap between the number of people eligible for family cash assistance and those receiving it was at its greatest, the child poverty rate was 16.5%—still below the 1994 rate. In 2018 (the last year for which data are shown in this report), the child poverty rate inclusive of noncash benefits and taxes was 13%—half the 1994 rate. The decline in absolute child poverty may address one potential concern regarding the decline in family cash assistance, though other concerns have also been raised.

The Shift Away from Ongoing Cash Assistance

Both before and after enactment of PRWORA, a series of legislative changes transformed need-tested aid for families with children. These changes expanded benefits mostly for families with children and working adults. Need-tested benefits have a comparatively large role in reducing child poverty relative to other demographic groups, with much of that reduction occurring among families with earnings.³⁴ The biggest impacts come from noncash benefits (particularly nutrition assistance) and the refundable tax credits. The refundable tax credits—the Earned Income Tax Credit and the Child Tax Credit—are explicitly tied to earnings. Though they are cash benefits, the tax credits are not paid on an ongoing basis during the year; instead, they are received the following year in a lump sum when federal income tax returns are filed.

The 2018 child poverty rate of 13% is based on an income concept that counts—in a manner equivalent to ongoing cash—the refundable tax credits (that would actually be paid in 2019) as well as noncash benefits. These two forms of benefits play different roles in family budgets than ongoing cash benefits. Families have less flexibility in the use of these benefits than they do in the use of ongoing cash income. Because the refundable tax credits are paid in a lump sum once a year, they may not be timed to maintain incomes when needs occur. Noncash benefits such as SNAP might be available to address those needs, but only partially because they can only be used to purchase certain items.

Responding to Recessions

Typically during economic recessions, income from work falls as unemployment increases. Some programs usually respond to aggregate changes in earnings, as more people become eligible for benefits and receive them. Examples of such programs include unemployment insurance and SNAP.³⁵ Trends in receipt of family cash assistance have not always tracked economic cycles, even under AFDC, but the caseload rise that occurred from 1989 through 1994 was in part attributable to the 1990-1991 recession.

As noted previously, during the Great Recession there was an increase in the number of people receiving benefits, but that was only a fraction of the increase in the number of people who were eligible for family cash assistance. This led to the peak of 14.7 million people eligible but not receiving benefits in 2009, compared with 6.1 million people receiving family cash assistance at some time during that year. Yet in 2009, the absolute child poverty rate (inclusive of noncash benefits and taxes) was 16.5%—still below the 1994 rate.

Congress and the President responded to the Great Recession with temporary legislative enhancements to existing programs, including changes targeted to the low-income population. Provisions were added to unemployment insurance to give states incentives to expand eligibility for benefits for workers who otherwise would not have been eligible for them.³⁶ SNAP benefits were increased by 15%.³⁷ This approach offset, at least partially, the reduced role of TANF assistance in maintaining incomes for families with children. TANF itself was provided extra funding to respond to the recession. A temporary $5 billion fund was established that helped finance the increase in family cash assistance, but it was also used to pay TANF emergency short-term assistance as well as subsidized employment. The substantial ad-hoc legislative response to the economic downturn raises the question of how families with children would fare, given the lower post-PRWORA receipt rates of family cash assistance, should Congress's response to a future economic downturn not include similar temporary legislative changes.

The Federal Interest in Child Economic Well-Being

Family cash assistance originated with state and local programs (although these programs did not serve all families or all single-parent families). Even when federal funding was provided, states retained a great deal of discretion in the design of these programs. The brief history of family cash assistance provided in this report notes how federal actions—administrative actions, legislation, and U.S. Supreme Court decisions—ultimately created a federal entitlement to benefits for eligible families. This was reversed by PRWORA, which ended the individual entitlement to benefits and had an explicit purpose of increasing state flexibility.

PRWORA ushered in an era of greater state differences in family cash assistance than what prevailed in the later years of AFDC. In addition to historical differences in eligibility levels and benefit amounts, and new flexibility in terms of services provided, there was an increase in differences among states in accessing family cash assistance. In the meantime, as discussed above, federal policy had much to do with an overall decline in child poverty, attributable in part to 100% federally financed SNAP benefits and 100% federally financed refundable tax credits. For families with children, a large share of these benefits redistributed income to parents with low earnings. However, the role of family cash assistance—dealing with earnings loss associated with childbirth, as well as misfortunes associated with loss of earnings to a family through unemployment, the onset of an illness or disability, an earner leaving the family, or an earner never joining the family—is under the control of the states, and how they deal with it varies considerably among them.

Appendix. Methods

The estimates of family cash assistance eligibility in this report are from the TRIM3 microsimulation model based on data from the Census Bureau's ASEC survey. Estimates of the receipt of family cash assistance represent a combination of receipt reported by respondents of the ASEC and estimates from TRIM3. Benefit receipt is under-reported on household surveys, and the TRIM3 estimates impute benefit receipt for additional persons to address that under-reporting. The TRIM3 estimates bring, as closely as possible, the total number of families and individuals receiving family cash assistance up to totals as reported by administrative data. TRIM3 also matches, as closely as possible, the characteristics of those in its estimates to those reported in the administration data.

Estimates in recent years on TANF cash assistance receipt include those from federally funded TANF, state dollars expended to meet the TANF MOE requirement, and solely state funded programs (SSFs). SSFs are state-funded programs with expenditures that are not counted toward the TANF MOE, but that serve TANF-related populations. States primarily use SSFs to fund assistance for two-parent families so they do not have to meet the two-parent TANF work participation standard. The estimates on family cash assistance receipt are for assistance under traditional assistance programs and exclude those receiving TANF benefits that are technically considered assistance but paid outside the traditional program only to those with earnings.

The estimates are subject to error. Two main categories of errors are modeling error (TRIM3 assigning eligibility or benefit receipt to the wrong people) and sampling error. Sampling error is quantifiable.

The ASEC provides a sufficient sample to make reliable national-level estimates. However, it does not have a big enough sample in all states to make reliable state-level inferences using a single year of data. The Census Bureau recommends pooling three years of data to make state-level estimates; thus, the state-level information uses such pooled three-year data. However, this still produces estimates in smaller population states that have a comparatively larger sampling margin of error than estimates in larger population states (larger population states have larger samples than do smaller population states). Table A-1 shows the estimated average family assistance receipt rate among eligible individuals for 2016-2018 and the margin of error (at the 90% confidence level) by state. While margins of error based on a comparable methodology are not possible for the 1994-1996 estimates, the pattern (more precise estimates for larger states with larger samples) is likely repeated for them as well.