Contents

• Introduction
• Data and Caveats
• Estimates from the Transfer Income Model Version 3 (TRIM3)
• The "Value" of Benefits to Individuals and Families
• The Special Case of Medicaid
• Programs Examined in this Report
• Program Characteristics
• How Many People Are Eligible for Need-Tested Benefits, and How Many Actually Receive Them?
• Eligibility for Selected Need-Tested Benefits
• Receipt of Selected Need-Tested Benefits
• Benefit Receipt among Those Eligible
• What Family Characteristics Are Associated with Receipt of Need-Tested Benefits?
• Pre-welfare Income
• Family Characteristics
• How Much Do Families Typically Receive in Benefits?
• Do Families Typically Receive Benefits From One Program or From Multiple Programs?
• What Family Characteristics Are Associated with Receipt of Relatively Large Amounts of Need-Tested Aid?
• Pre-welfare Income-to-Poverty Ratios
• Family Characteristics
• Families with an Individual with Disabilities
• Families with Children
• Benefit Receipt by Program
• Conclusion

Summary

Need-tested benefits have received increased attention from policymakers in recent years, as spending levels for these programs remain elevated well into the economic expansion that followed the 2007-2009 recession. While information is available on the number of people who receive benefits from individual programs, it is more challenging to examine how these programs interact and the cumulative benefits families receive from them. Case studies based on hypothetical families often show how much in benefits a family may potentially receive from multiple programs under federal and state policies. However, these case studies assume families receive all the benefits they are eligible for and receive them all year. This is often not true.

This report examines estimated benefit receipt by families from nine major need-tested benefit programs in 2012. The nine programs are the Supplemental Nutrition Assistance Program (SNAP); the Earned Income Tax Credit (EITC); Supplemental Security Income (SSI); subsidized housing assistance; the Additional Child Tax Credit (ACTC); the special supplemental nutrition program for Women, Infants, and Children (WIC); Temporary Assistance for Needy Families (TANF) cash assistance; the Child Care and Development Fund (CCDF); and the Low-Income Home Energy Assistance Program (LIHEAP). The estimates are derived from a combination of information from a Census Bureau household survey and a model that estimates program eligibility and participation based on information from that survey.

An estimated 135 million persons, 4 in 10 persons in the noninstitutionalized population, were eligible for benefits from at least one of these programs in 2012. However, not all persons eligible for need-tested benefits actually received them. Among the programs examined in this report, an estimated 70% of eligible families actually received SNAP and 65% of eligible families received WIC in 2012. However, the estimated rate of benefit receipt among eligible persons was 28% for TANF cash assistance, 22% for LIHEAP, 18% for subsidized housing, and 17% for CCDF (based on eligible children).

An estimated 106 million persons (1 in 3 persons in the population) actually received benefits from one of these programs in 2012. Benefits were concentrated among people in families with children and families with an individual with disabilities with those two groups accounting for an estimated 78% of total benefit dollars from the selected programs. Many families that received need-tested benefits had characteristics not typically associated with economic disadvantage; a substantial portion of families that received aid had pre-welfare incomes above the poverty line in 2012. Among families with children in 2012, an estimated 45% of those who had a worker and 38% with at least one adult working full-time all year received at least one need-tested benefit.

The estimated median annual benefit amount from the nine programs in 2012 was $3,300 (i.e., half the families that received benefits received less than $3,300 and half received more). About 40% of families that received need-tested aid did so from only one of the nine selected programs.

Some families received relatively large amounts of need-tested aid. In 2012, an estimated 25% of families that received benefits from one or more of the selected programs received a total of $9,027 or more. These families accounted for two-thirds of all spending for these programs in 2012. Families with children who received $9,027 or more had characteristics indicative of a more disadvantaged population: working less than full-time all year, lacking a high school diploma, being in a family headed by a single woman, being of a racial/ethnic minority (other than Asian-American), and being in a large family.

---

Need-Tested Benefits: Estimated Eligibility and Benefit Receipt by Families and Individuals

Introduction

Benefits and services for low-income families have received increased attention from policymakers in recent years. This is due in part to elevated federal spending in the wake of the deep recession of 2007-2009.

The federal government, sometimes in partnership with state governments and local entities, provides an array of economic assistance to persons and families with low income. There are various ways to identify these "need-tested" programs and the individuals who are eligible for and served by them.¹ However, it is more challenging to identify how these programs interact with each other and how they cumulatively provide benefits to families and individuals. There is no single data source that identifies all need-tested benefits for which individuals and families are eligible or those they actually receive.

One way to analyze the interaction of need-tested benefits is to look at case studies, which create hypothetical families and present a suite of benefits and services for which those families would be eligible under program rules. For example, a Congressional Research Service (CRS) analysis estimated that in 2014, a single mother of two children, working all year, full-time at the minimum wage could receive almost $9,000 in combined benefits from the Earned Income Tax Credit (EITC), the Additional Child Tax Credit (ACTC), and the Supplemental Nutrition Assistance Program (SNAP).² However, such case study analyses are limited because their hypothetical families do not reflect the wide variation in characteristics and circumstances of actual families. They typically assume that families receive all the benefits for which they are eligible, which is not the case in reality. Further, they sometimes assume that families receive benefits for a full year, though many families are financially needy for only part of a year and receive benefits for only some months as their circumstances change.

This report takes a different approach to exploring the interaction of need-tested benefits. It uses data collected by the U.S. Census Bureau, along with microsimulation modeling, to explore individuals' and families' eligibility and benefit receipt. The report uses the same data source as, and methods similar to, a Government Accountability Office (GAO) report released in 2015 (U.S. Government Accountability Office, Federal Low-Income Programs, Multiple Programs Target Diverse Populations and Needs, GAO 15-516, July 2015). The GAO report describes federal programs for people with low incomes, examines selected household characteristics of people in poverty, examines the poverty status and household characteristics of selected programs' recipients, and discusses the research on how selected programs may affect incentives to work. This report complements the information in the GAO report, providing further analysis of benefit receipt from specific need-tested programs.

Specifically, this report provides information on eligibility and benefit receipt among individuals and households in 2012³ from nine major need-tested programs for which adequate data were available, listed in the order of the amount of their FY2012 federal obligations:

• Supplemental Nutrition Assistance Program (SNAP);
• Earned Income Tax Credit (EITC);
• Supplemental Security Income (SSI);
• housing assistance provided through the Section 8 Housing Choice Voucher program, the public housing program, and the project-based rental assistance program (collectively referred to as "housing assistance");
• Additional Child Tax Credit (ACTC);
• special supplemental nutrition program for Women, Infants, and Children (WIC);
• cash assistance from the Temporary Assistance for Needy Families (TANF) block grant;
• Child Care and Development Fund (CCDF); and
• Low-Income Home Energy Assistance Program (LIHEAP).

This report is organized around and addresses six questions that may help inform policy discussions about the future of selected need-tested benefits. The six questions and key findings related to them are presented below in Table 1.

Data and Caveats

While the federal government supports a number of need-tested benefit programs, uniform data on eligibility, participation, and benefits across programs are generally available only from household surveys. However, respondents from those surveys tend to under-report receipt of need-tested assistance. ⁴ Thus, this report combines information from a Census household survey with estimates from a microsimulation model, the third version of the Transfer Income Model (TRIM3). ⁵ The need-tested benefit programs examined in this report are limited to nine cash or in-kind transfer programs and tax provisions for which eligibility, benefit receipt, and benefit amounts are estimated by TRIM3.⁶

Estimates from the Transfer Income Model Version 3 (TRIM3)

TRIM3 is a microsimulation model for government benefit receipt that is primarily funded by the U.S. Department of Health and Human Services (HHS) and maintained at the Urban Institute. TRIM3 combines administrative data on program rules with survey data from the U.S. Census Bureau's Annual Social and Economic Supplement (ASEC) to the Current Population Survey. Simulations conducted with TRIM3 attempt to correct for some limitations of the underlying survey data, such as under-reporting of certain benefits in the ASEC.

TRIM3 uses data from the ASEC to estimate the number of people eligible for benefits from selected need-tested programs. For most programs, TRIM3 estimates actual benefit receipt based on the number of people federal agencies report as receiving benefits and the probabilities that an eligible person receives benefits. An exception to this general method is the refundable tax credits. The information on the ASEC identifies fewer tax filers eligible for the EITC and the ACTC than claimed these credits on their federal income tax returns. Therefore, the TRIM3 estimates of people receiving these tax credits are set equal to the number that are identified as eligible for them on the ASEC. Thus, in these estimates there are fewer tax filers and there is a shortfall in aggregate credits for the EITC and the ACTC compared to what is claimed on federal tax returns.

TRIM3, like all models, is subject to its own limitations. Appendix A provides detailed information on the methodology used to develop the estimates in this report, including data limitations and key assumptions. The estimates presented in this report were derived using ASEC and TRIM3 data for calendar year 2012. Readers should be aware that estimates in this report may not match administrative data for several reasons, including the following:

• The estimates in this report reflect annual measures of income, program participation, and benefit amounts. Thus, the estimates may not match administrative data, which commonly report average monthly, rather than annual, participation and benefit amounts.
• The estimates measure poverty using the Supplemental Poverty Measure (SPM), rather than the official poverty measures.⁷ The SPM differs from the official poverty measures in several ways, including its use of a broader definition of the family unit. It also measures net income available to meet a family's non-medical needs. From gross income (including the value of government "in-kind" benefits), the SPM subtracts taxes paid, work expenses, child support payments for children outside the household, and out-of-pocket medical expenses.
• The estimates present benefit amounts using a common family unit. TRIM3 estimates benefits based on the filing unit used in each program and, generally, assigns a per-person benefit amount to members of that filing unit or a benefit amount to the head of the filing unit. For this analysis, persons and their benefits are grouped into a family unit based on that used for the SPM. This single definition of "family" allows analysis of eligibility and benefit receipt across all programs in a comparable manner, which would otherwise not be possible.

The "Value" of Benefits to Individuals and Families

Readers should also be aware that the "benefit amounts" in this report reflect the estimated dollar value of aid received by individuals and families. However, a large share of need-tested benefits is paid in forms other than cash (e.g., medical, food, housing, and child care cost assistance). In theory, a dollar in benefits received as cash is worth a dollar. A recipient who receives cash assistance can choose what goods and services to purchase with that dollar or whether to save all or part of it.

On the other hand, a recipient may value a dollar received in noncash form (medical, food, or housing assistance) differently from a dollar in cash. This is because noncash benefits are not "fungible"—they must be used for a particular type of good or service. Policy analysts have developed several different methods for determining the value of noncash benefits such as medical, food, and housing assistance.⁸ However, it is beyond the scope of this report to analyze the relative value of such benefits to any given family or to assess whether receipt of such benefits measurably improves family well-being.

The Special Case of Medicaid

Medicaid is the largest need-tested program in terms of federal spending. The $270.9 billion spent on Medicaid in FY2012 exceeded the total amount of federal spending for all of the nonmedical need-tested programs examined in this report. Medicaid is an entitlement program that finances the delivery of primary and acute care health services, as well as long-term services and supports.⁹

Medicaid is not included in the central analysis of this report. Rather, it is discussed separately (in Appendix B) for several reasons:

• TRIM3 did not include estimates of Medicaid enrollment and the value of benefits for 2012.
• This report focuses on receipt of need-tested benefits in 2012. However, the Affordable Care Act (ACA) made substantial changes to Medicaid eligibility, which became effective in 2014. Thus, the picture of Medicaid today and in future years might be significantly different from that of 2012.
• While there are conceptual and technical issues in estimating the "value" to families and individuals of all noncash benefits (including food and housing), the valuation of health care benefits is particularly challenging.

Because of these issues, topics discussed in this report, such as eligible population, enrollees, and percentage of eligible individuals actually enrolled in a program, are addressed separately for Medicaid in Appendix B.

Programs Examined in this Report

This report examines benefit receipt from nine major need-tested benefit programs; they are listed in the report's introduction and discussed in more detail in Table 2, below.

In some cases, the report considers only a portion of the benefits provided under a program (e.g., only the refundable portion of the child tax credit (known as the ACTC), or only the cash assistance portion of TANF). In other cases, several programs are aggregated and treated as one program (e.g., housing assistance).

The nine programs are not necessarily the largest need-tested benefit programs, either in terms of spending or individuals served. As discussed above, the largest need-tested benefit program, Medicaid, is not included in this analysis but is examined separately in Appendix B. Nor do these programs represent the full breadth of assistance potentially available to low-income individuals and families. (For a more comprehensive list, see CRS Report R43863, Federal Benefits and Services for People with Low Income: Programs and Spending, FY2008-FY2013, by [author name scrubbed] and [author name scrubbed].) Rather, as discussed previously, they represent a subset of need-tested benefit programs—those for which sufficient data are available in TRIM3.

Program Characteristics

Though all nine programs examined here provide benefits based on individual or family financial need, they differ considerably in terms of who is eligible for assistance, the type of assistance provided, the conditions placed on receipt of benefits, and how benefits are funded. These differences in program characteristics are important to consider, as they help to explain some of the findings explored later in the report.

Table 2 presents basic characteristics of each of the programs examined in this report, including the following:

• Income eligibility. Basing eligibility on income and other financial need criteria, such as low levels of assets, is the defining characteristic of a need-tested program. All of the programs examined in this report use explicit income eligibility criteria that individuals, families, or households must meet, but the specific levels and measures of income vary. Some measures are uniform throughout the country (e.g., those in the tax code); others vary by geography and/or are based on relative measures (e.g., state or local median income). Further, some programs use alternative criteria that allow specified groups or categories of people to qualify automatically without having to meet an individual income test. Thus, eligibility for these programs is not necessarily tied to being "poor" by official measures, and many programs consider families with income above the poverty level to be eligible. For example, in the case of WIC the income eligibility threshold is 185% of the poverty line, and for housing assistance it is as high as 331% of poverty in Washington, DC. TANF income eligibility thresholds are set by states, although families typically must have incomes below poverty (in fact, incomes must be less than half of the poverty level in many states) to receive TANF cash assistance.¹⁰
• Populations eligible. For many programs, there are additional requirements beyond income eligibility rules, so that an applicant must be both income-eligible and a member of the program's target population. Of the nine programs considered in this report, four (the ACTC, TANF, CCDF, and WIC) would only be available for families with children (for WIC, and sometimes TANF, this includes families expecting a child). That is, families without a child are ineligible for benefits under these programs regardless of their income. SSI is restricted to aged, blind, or disabled individuals (including children).
• Other requirements. Some benefits are tied specifically to workers. For example, the EITC and the ACTC are available only to workers with earnings and their families. To receive child care subsidies from the CCDF, parents must be working or in training, in addition to meeting income eligibility criteria. Some programs also impose behavioral requirements as a condition of eligibility. For example, recipients of TANF cash assistance must comply with work and training requirements and able-bodied adults without dependent children must comply with work and training requirements to receive SNAP benefits for more than a limited time.
• Form of assistance. Some of the programs provide cash to families, such as the refundable tax credits, TANF, and SSI. Others provide in-kind benefits to help families meet their basic needs, such as housing and food. Child care assistance is often considered a work support, providing full or partial reimbursement for an expense associated with employment.
• Funding category. Another important way in which the programs vary is how they are funded. Some receive mandatory funding, some receive discretionary funding, and one (CCDF) receives both. Mandatory funding may be structured as open-ended or capped. In an open-ended program, no pre-determined ceiling is imposed on federal expenditures; instead, federal payments are made to all eligible beneficiaries for eligible expenditures as defined in law. (SSI is an example of an open-ended mandatory program.) In a capped program, the authorizing law limits the total amount of federal spending that can occur. (TANF is an example of a capped mandatory program.) The amount of federal funding for discretionary programs, on the other hand, is determined by Congress through the annual appropriations process. It is important to note that capped mandatory and discretionary programs may not have sufficient funding to serve all eligible individuals. Refundable tax credits, though administered through the tax code, effectively are financed like open-ended mandatory programs.
• Size. The programs also vary significantly by size, which may be due in part to a program's eligibility rules or funding category. In terms of federal obligations, SNAP is the largest of the programs considered in this report (accounting for 30% of the combined $261.5 billion in total obligations across all nine programs in FY2012) and LIHEAP is the smallest (accounting for just over 1% of the total obligations). However, the programs rank differently when measuring size by the number of individual recipients, with the largest program being the EITC and the smallest being CCDF. (Note, however, that child care recipients in this report reflect only the children served, not the parents.) Table 2 orders programs based on their size in terms of federal dollars spent on them.

How Many People Are Eligible for Need-Tested Benefits, and How Many Actually Receive Them?

The selected need-tested programs examined in this report target different populations and have different eligibility criteria. Some entitle all eligible individuals to benefits; others have limited funding and can only serve a limited number of eligible individuals and families. This section discusses how these rules translated to the population in 2012, quantifying how many individuals were eligible for these programs and how many actually received benefits.

Eligibility for Selected Need-Tested Benefits

In 2012, the total number of people who were estimated to be eligible for at least one of the need-tested programs examined in this report was 135 million, or more than 4 in 10 persons among the nation's non-institutionalized population.¹¹ These 135 million people resided in 58 million families. Figure 1 shows the number eligible for any of the nine programs, as well as the number eligible for each of the nine programs. The programs are ranked by the size of their eligible populations.

The SNAP program has the largest number of people eligible for benefits. In 2012, an estimated total of 83.3 million were eligible, representing 27% of the total non-institutionalized population. Like several other programs (e.g., LIHEAP, housing assistance), SNAP was available in 2012 to people in eligible households of all family types (aged, disabled, families with children, and childless adults). ¹² By contrast, TANF and child care are limited to low-income families with children. WIC is limited to low-income families with a pregnant woman or a young child. SSI is restricted to aged, blind, or disabled individuals.

The two tax credits (EITC and ACTC) are administered through a universal system (the federal income tax system) and primarily benefit families with children and earnings; the number of people shown as eligible for the EITC includes childless workers who receive a relatively small benefit, and the ACTC is restricted to families with children.

Figure 1. Estimated Number of People Eligible for Selected Need-Tested Benefits, 2012 Millions of People

Source: Congressional Research Service (CRS) estimates using data from the Census Bureau's 2013 Annual Social and Economic Supplement (ASEC) to the Current Population Survey, supplemented with estimates of program eligibility, receipt, and benefits from the TRIM3 microsimulation model. Notes: Represents people eligible for benefits at any time during the year. For the refundable tax credits, represents an estimate of tax credits based on 2012 earnings, claimed on the 2012 return, and filed and paid in 2013. Number of people eligible for tax credits represents the tax filer, spouse, and dependents of the tax filer. For subsidized child care, the number of people eligible represents children eligible for subsidized care.

Receipt of Selected Need-Tested Benefits

Not all those eligible for a need-tested benefit actually receive one. In 2012, the total number of people who were estimated to have received at least one of the need-tested benefits examined in this report totaled 106 million, out of 135 million estimated to be eligible. The 106 million recipients resided in 42 million families.

Figure 2 shows estimates of the number of people who actually received benefits in 2012, both overall and for each of the nine programs. Programs are ranked by the number of recipients. As discussed in "Estimates from the Transfer Income Model Version 3 (TRIM3)," because of limitations of the data from the ASEC, TRIM3 estimates the number of people receiving benefits from the two refundable tax credits as being equal to the number of people estimated to be eligible for these benefits.¹³ For all the other programs, TRIM3 estimates the number of people receiving benefits as a proportion of those who were estimated to be eligible (discussed in the next section). Thus, programs are ranked differently than in Figure 1, programs varied in the percentage of those eligible who actually received benefits.

The EITC was the most widely received benefit in 2012, with an estimated 62.9 million persons (representing 20% of the total U.S. population) benefiting from the refundable tax credit. SNAP was the second largest benefit in terms of recipient population, with an estimated 58 million persons (19% of the total U.S. population) receiving the benefit at some time during 2012. The other refundable tax credit, the ACTC, was the third most widely received benefit in 2012. The number of recipients for each of the remaining programs was well below that of these top three programs.

Figure 2. Estimated Number of People Receiving Selected Need-Tested Benefits, 2012 Millions of People

Source: Congressional Research Service (CRS) estimates using data from the Census Bureau's 2013 Annual Social and Economic Supplement (ASEC) to the Current Population Survey, supplemented with estimates of program eligibility, receipt, and benefits from the TRIM3 microsimulation model. Notes: Represents people receiving benefits at any time during the year. For the refundable tax credits, represents tax credits earned during the year. Recipient counts for tax credits represent the tax filer, spouse, and dependents of the tax filer. For subsidized child care, the number of people represents children receiving subsidized care at any time during the year.

Benefit Receipt among Those Eligible

Some programs provide benefits to a large share of their eligible populations, while others serve a relatively small portion of those eligible. Figure 3 shows the estimated percentage of the eligible population served in 2012 by seven of the programs discussed in this report. The two refundable tax credits are not shown as their estimated recipient populations are assumed to be identical to their estimated eligible populations. Programs are ranked by their percentage of eligible populations receiving benefits.

Three of the seven programs had relatively high rates of receipt among their eligible populations: the two nutrition programs (SNAP and WIC) and SSI. It should be noted that estimating the eligible population for SSI is somewhat difficult. Estimates of the SSI-eligible population are based, in part, on self-reported information about impairments, functional limitations, and health status. Such information is subjective in nature, because the data reflect the respondent's (or household head's) concept of "disability" at the time the survey was administered.¹⁴ In contrast, in order to actually receive SSI, an adult claimant must be certified by an independent examiner that he or she is unable to perform substantial work due to a severe physical or mental impairment that is expected to last for at least one year or result in death. Although needy individuals who report having a work-limiting disability are potentially eligible for SSI, not all of them would ultimately qualify for benefits if they applied, because their impairment may not meet the program's statutory standards.¹⁵

SNAP was estimated to provide benefits to about 7 in 10 eligible persons in 2012.¹⁶ WIC and SSI were estimated to provide benefits to about two-thirds of eligible persons in that year. The rate of receipt for the remaining four programs was much lower. For example, TANF served about 3 in 10 eligible persons in 2012, and the rate of receipt for housing, LIHEAP, and child care subsidies was even lower. A common feature among these four programs is that funding for each is capped, and states (and localities, in the case of housing) must sometimes ration aid, using mechanisms such as waiting lists for housing and child care benefits. Federal law also explicitly states that TANF is not to be considered an entitlement to individuals. In addition, TANF has work requirements and time limits that might deter some individuals from applying for aid. However, lack of entitlement status does not necessarily mean that a program serves a small number of those eligible. WIC is technically not an entitlement, but it received enough federal funding to serve all those that sought its benefits.

Figure 3. Estimated Percentage of Eligible Persons Receiving Selected Need-Tested Benefits, by Program, 2012

Source: Congressional Research Service (CRS) estimates using data from the Census Bureau's 2013 Annual Social and Economic Supplement (ASEC) to the Current Population Survey, supplemented with estimates of program eligibility, receipt, and benefits from the TRIM3 microsimulation model.

What Family Characteristics Are Associated with Receipt of Need-Tested Benefits?

A substantial minority of all individuals and families benefited from at least one of the selected need-tested programs in 2012. However, as discussed earlier in "Programs Examined in this Report," these programs have different income eligibility rules and provisions that target benefits toward certain populations, such as families with children or the aged and disabled. Thus, some types of families may be more likely to receive need-tested aid than others.

Pre-welfare Income

As would be expected, families with lower incomes before receipt of need-tested benefits had a greater likelihood of receiving them than families with higher incomes in 2012. However, not all people in poor families receive a need-tested benefit, and conversely, not all need-tested benefits go to people who are poor.

This section examines benefit receipt for families based on the ratio of their pre-welfare incomes to poverty thresholds. (See Table 3, below, for an explanation of "pre-welfare" income.)

Figure 4 shows the estimated percentages of families that received any need-tested assistance by their pre-welfare income-to-poverty ratio for 2012. The figure shows that families with pre-welfare incomes below their poverty thresholds (pre-welfare income-to-poverty ratios of 0% to 49% and 50% to 99%) were highly likely to receive a need-tested benefit, with about 8 out of 10 such families estimated as receiving aid. However, families with pre-welfare incomes above the poverty thresholds also received benefits. A little more than half (53.7%) of all families with pre-welfare incomes of between 100% and 150% of their poverty threshold received need-tested benefits, and about 3 in 10 families with pre-welfare incomes between 150% and 200% of their poverty thresholds also were estimated to receive benefits.

Figure 4. Estimated Percentage of Families Receiving Need-Tested Benefits from At Least One Program, by Pre-welfare Income-to-Poverty Ratio, 2012 (Pre-welfare income-to-poverty ratio is expressed as a percentage)

Source: Congressional Research Service (CRS) estimates using data from the Census Bureau's 2013 Annual Social and Economic Supplement (ASEC) to the Current Population Survey, supplemented with estimates of program eligibility, receipt, and benefits from the TRIM3 microsimulation model. Notes: Single individuals are considered a family of one in this tabulation.

Table 4 shows the estimated percentage of families, by their pre-welfare income-to-poverty ratio, who received benefits from each of the nine programs examined in this report in 2012. The three most widely received benefits were the two refundable tax credits (ACTC and EITC) and SNAP. SNAP was more widely received than any other benefit for families with pre-welfare incomes below the poverty thresholds. For families with incomes above the poverty thresholds, the most common benefits were the EITC and/or the ACTC.

Family Characteristics

Some of the programs examined in this report base eligibility on low income alone and do not restrict benefits by family type. Other programs restrict their benefits to families or individuals of a certain type, such as families with children or with aged persons or individuals with disabilities. These restrictions generally are based on societal expectations regarding work participation. According to economic theory, provision of government benefits without a tie to work reduces incentives to work. Empirical studies have generally supported that government benefits that are not tied to work have an effect on reducing work, though studies often differ about whether that effect was large or small.¹⁷

Historically, the aged and individuals with disabilities have not been expected to work. Thus, concerns about providing individuals in these two groups with basic needs have tended to outweigh concerns about the work disincentive inherent in government benefits without work requirements. Social Security retirement and disability benefits are also available to the aged and disabled, based on sufficient past work in covered employment¹⁸ and meeting age or disability criterion. Income from Social Security may be sufficient to raise family incomes above the eligibility thresholds used by need-tested programs.

Expectations about work are often at the center of debates about aid to families with able-bodied, non-aged adults. As will be discussed in this section, most aid to such families goes to those with children. The Social Security Act of 1935 established Aid to Dependent Children, later called Aid to Families with Dependent Children (AFDC), with the explicit goal of providing single mothers with assistance so they did not have to work. But given the changing role of women in the workforce, particularly since the 1960s, more recent policies have sought both to require and support work for single mothers. These policies eventually resulted in large expansions of refundable tax credits for families with children, paid only to families with earnings; subsidized child care for working parents; and the replacement of AFDC with TANF in the 1990s.

In order to examine benefit receipt by the presence of aged or disabled individuals, children, and earners in the family, for this report CRS classified all families as being in one of six groups: (1) families with an aged member (aged 65 or older); (2) families with an individual with disabilities (3) families with children and no earners; (4) families with children and earners; (5) other families, without a member who was aged, disabled, or a child, with no earners; and (6) other families, without a member who was aged, disabled, or a child, with earners.

The classification of families is sequential; for example, if a family has both an aged member and an individual with disabilities it is assigned to the category of a family with an aged member. Only those families without an aged member are considered when the next assignment is made, which would be if the family had an individual with disabilities. Families can be assigned only to one category for the purpose of this classification. However, some families meet the criteria for more than one category. (For a discussion of families that could be classified in more than one category, see Appendix C.)

In 2012, there were an estimated 128.8 million "families" identified under the Census Bureau's SPM definition of family. Single persons were classified as a family of one for this purpose. Table 6 shows the number and percentage of families in each category. It shows, for example, that 32.5 million families (25.2% of all families) had an aged member in 2012. Families without an aged member that included an individual with disabilities totaled an estimated 18.6 million (14.5% of all families).

Families without an aged person or individual with disabilities were further divided into four groups based on whether they contained at least one child and whether an adult in the family had worked at all during the year. Most of these families had at least one adult worker in 2012.¹⁹ Only an estimated 0.9 million families (0.7% of all families) with children had no adult worker, and 3.5 million families (2.7% of all families) of non-aged, non-disabled adults without children had no adult worker.

The table also shows the relative pre-welfare poverty status for families in each category. The poorest category was families with a child and no adult worker. In 2012, 92.3% of all such families were pre-welfare poor. The least poor group was non-disabled childless adults (without an aged member or individual with disabilities) who had at least one worker.

The pre-welfare poverty rates for families with an aged member and those without an aged member or individual with disabilities that had children and an adult worker were fairly similar (19.4% for the former, 19.9% for the latter). Families without an aged member that had an individual with disabilities had a relatively high overall pre-welfare poverty rate (42.2%), with 25.1% of these families having pre-welfare incomes of less than 50% of the poverty line.

Figure 5 shows the estimated rate of receipt for need-tested benefits by family category in 2012. These rates apply to families by category at all income levels. The highest rates of benefit receipt occurred among families with children with no workers (91.6% received aid) and among families with a disabled member (60.4% received aid). Of families with children with workers, 44.6% received need-tested aid in 2012. It should be noted that families "with workers" represent those with any adult in the family working at any time during the year. About 1 in 4 aged persons received a need-tested benefit in 2012.

Families that included a worker and had no members who were aged, disabled, or children had the lowest rate of need-tested benefit receipt among all the family categories shown in Figure 5, 16.1%. Among families with no workers and no members who were aged, disabled, or children, 37.5% received a need-tested benefit. The differences in benefit receipt among family categories reflect both different economic circumstances and differences in the programs for which each category of family is eligible. Moreover, there are differences in the "take-up" rate of benefits (the rate at which those who are eligible for benefits actually receive them) among eligible categories of families. For example, previous research has indicated that the aged tend to take up benefits at a lower rate than other population groups.²⁰

Figure 5. Estimated Percentage of Families Receiving Selected Need-Tested Assistance, by Family Category, 2012

Source: Congressional Research Service estimates using data from the Census Bureau's 2013 Annual Social and Economic Supplement (ASEC) to the Current Population Survey, supplemented with estimates of program eligibility, receipt, and benefits from the TRIM3 microsimulation model. Notes: Single individuals are considered a family of one in this tabulation. Families with children are those with a child under the age of 18 who do not have an aged person or individual with disabilities.

As discussed earlier, some categories of families are ineligible for certain benefits, as reflected in the different rate of receipt by families in different categories of each of the nine programs examined in this report. Table 7 shows the estimated percentage of families that received benefits in each of the nine programs by family category in 2012. It shows that SNAP was the most widely received benefit for all family categories with the exception of families with children with workers. A greater percentage of families in that category received benefits from the two refundable tax credits (EITC and ACTC) than received SNAP. The table also shows the effect of policies that restrict receipt of the ACTC, TANF, WIC, and child care subsidies to families with children. (WIC and, at the option of a state TANF, can also be received by pregnant women, which explains the small number recipients from these programs in the childless adult categories.) As discussed above, families with an aged or disabled member may also have children, so some of these families received benefits from the programs restricted to families with children.

Figure 6 shows estimated total spending for the selected need-tested benefits by family category in 2012. The two family categories accounting for the largest share of spending were families with children and families with a member with disabilities. Most families with children had workers, and families with children with workers accounted for an estimated $91.6 billion, or 38% of all spending for the selected need-tested assistance. Even though almost 9 in 10 families with children without workers also received need-tested benefits, there were relatively few such families. Need-tested spending for families with children with no worker was $9.4 billion, or about 4% of all spending for the selected need-tested assistance.

Families with individuals with disabilities accounted for an estimated $89.5 billion, or 37% of the selected need-tested spending in 2012. On the other hand, families with an aged member accounted for $38.5 billion, or 15.7% of the selected need-tested spending. Families without an aged, disabled, or child member accounted for a small share of need-tested spending. Such families with workers accounted for $9.1 billion; such families without workers accounted for $3.3 billion.

The figure also shows spending by category of benefit. A large share of spending for families with children and earners is on refundable tax credits. In 2012, refundable tax credits accounted for 50% of the selected need-tested spending for families with children and earners. These tax credits were also prominent in aid to families with a disabled member. Such families often have earners—sometimes the disabled person and sometimes other adults. These families also sometimes have children.

As shown in the figure, tax credits are relatively small for families with earnings that do not have an aged, disabled, or child member. Although 10% of these families receive the EITC, the EITC benefits for "childless" workers are relatively small. The bulk of EITC benefits go to families with children.

Figure 6. Estimated Selected Benefit Spending, by Family Category, 2012 Dollars in Billions

Source: Congressional Research Service estimates using data from the Census Bureau's 2013 Annual Social and Economic Supplement (ASEC) to the Current Population Survey, supplemented with estimates of program eligibility, receipt, and benefits from the TRIM3 microsimulation model. Notes: Single individuals are considered a family of one in this tabulation. Families with children are those with a child under the age of 18 who do not have an aged person or individual with disabilities. The cash assistance category represents benefits from TANF cash assistance and SSI. Tax credits represent refundable tax credits from the ACTC and the EITC. Food assistance represents SNAP and WIC. Housing assistance represents Section 8 rental assistance, rent subsidies in public housing, and LIHEAP.

In addition to pre-welfare income and family type—two factors explicitly taken into account in determining eligibility for many need-tested programs—the rate of need-tested benefit receipt varies by characteristics commonly associated with economic disadvantage, specifically job attachment, educational attainment, and family structure. These factors are discussed in a later section of this report, "What Family Characteristics Are Associated with Receipt of Relatively Large Amounts of Need-Tested Aid?"

How Much Do Families Typically Receive in Benefits?

As discussed earlier, the selected programs examined in this report (which exclude the large Medicaid program) provide benefits to more than 1 in 3 persons in the population residing in 42 million families. How does this translate in terms of dollars received²¹?

In 2012, the annual median benefit to families who received at least one need-tested benefit was estimated to be $3,300 (i.e., half the families who received at least $1 in aid received an amount less than or equal to $3,300, the other half received more than $3,300). However, the median does not necessarily reflect the circumstances of most families receiving need-tested aid.

Figure 7 shows the estimated number of families receiving selected need-tested benefits by the total dollar amount of annual benefits received from all programs in 2012. The figure shows that many families received a relatively small benefit during the year—more than 10 million families received less than $1,000 in benefits. This represents about 1 in 4 families that received need-tested benefits. As noted above, half of all families received less than $3,300 during the year. Relatively few families received large benefits. The figure shows that as the annual benefit amount per family increased, the number of families receiving benefits decreased.

Figure 7. Estimated Number of Families Receiving Need-Tested Benefits, by Dollar Value of Annual Benefits from All Selected Programs, 2012

Source: Congressional Research Service estimates using data from the Census Bureau's 2013 Annual Social and Economic Supplement (ASEC) to the Current Population Survey, supplemented with estimates of program eligibility, receipt, and benefits from the TRIM3 microsimulation model. Notes: Single individuals are considered a family of one in this tabulation.

The distribution of annual benefits among families in 2012 is what is known as a "skewed" distribution. It is asymmetrical, with many families bunched at one end of the distribution—in this case at the lower dollar amounts. However, a few families tend to receive high benefit amounts.

A characteristic of this skewed distribution is that the relatively few families that receive high annual need-tested benefit amounts account for a disproportionately large share of all spending. For the purposes of this report, the benefit total received by the 25% of families that received the highest amounts—$9,027 or more in 2012—is used as the high benefit amount concentrated among relatively few families. Figure 8 examines total spending for the selected need-tested programs by the amount of annual benefits received by families. It shows that 11% of all spending on the selected need-tested benefits was for families with annual benefits of less than $3,300 (the median). This means that the 50% of all families with annual benefits less than $3,300 accounted for 11% of all spending. On the other hand, families with annual benefits of $9,027 or more accounted for 64% of all spending.

Figure 8. Estimated Distribution of Families Receiving Selected Need-Tested Benefits and Benefit Dollars, 2012

Source: Congressional Research Service estimates using data from the Census Bureau's 2013 Annual Social and Economic Supplement (ASEC) to the Current Population Survey, supplemented with estimates of program eligibility, receipt, and benefits from the TRIM3 microsimulation model. Notes: Single individuals are considered a family of one in this tabulation.

Do Families Typically Receive Benefits From One Program or From Multiple Programs?

An earlier section of this report ("How Many People Are Eligible for Need-Tested Benefits, and How Many Actually Receive Them?") discussed the extent to which individuals received benefits from any and each of the need-tested programs examined in this report. Of additional interest is the extent to which benefits were received from one program only or from more than one.

Table 8 shows families that received at least one need-tested benefit in 2012, by the number of programs from which they were estimated to receive benefits and certain benefit combinations. The table shows that an estimated 4 in 10 families that received at least one need-tested benefit received benefits from only one program. The remainder, or a majority of families that received at least one benefit, did so from more than one program.

In 2012, an estimated 14% of those who received any benefit received only SNAP and 13% received only the EITC. Other program benefits were received alone less frequently. Among families that received benefits from two programs, the most common combination was simultaneous receipt of the two refundable tax credits, the EITC and the ACTC. The most common combination of three programs was the two tax credits together with SNAP.

As would be expected, families that received benefits from more programs tended to receive higher total annual benefits. Figure 9 shows the estimated median total annual benefits for families by the number of programs from which they received benefits. The median annual benefit was $800 for families that received aid from only one program, $3,595 for families that received aid from two programs, and the amounts rise for each additional program. For families that received benefits from five or more programs, the median benefit was $17,180 for the year.

Figure 9. Estimated Number of Selected Need-Tested Programs Families Received Benefits From and Median Total Annual Benefits, 2012

Source: Congressional Research Service estimates using data from the Census Bureau's 2013 Annual Social and Economic Supplement (ASEC) to the Current Population Survey, supplemented with estimates of program eligibility, receipt, and benefits from the TRIM3 microsimulation model. Notes: Single individuals are considered a family of one in this tabulation.

Table 9 shows the estimated median benefits received by families and the estimated percentage of families receiving benefits from multiple programs in 2012, by their pre-welfare income-to-poverty ratios. The table shows that median benefits were higher for families with lower pre-welfare income relative to need. The median benefit for the poorest families (pre-welfare incomes less than 50% of the poverty threshold) was $9,262, and fell to $1,000 for the highest income group (pre-welfare incomes of 300% of poverty or more). Additionally, families with lower incomes were more likely to receive benefits from more than one program, while a majority of families in the higher income groups (pre-welfare incomes of 200% of poverty or higher) received benefits from only one program.

What Family Characteristics Are Associated with Receipt of Relatively Large Amounts of Need-Tested Aid?

About one-third of all persons were estimated to receive at least some benefit from the selected need-tested programs in 2012, and the rate of benefit receipt among some groups—such as those in families with an individual with disabilities or a family with a child—was even higher. The median annual benefit for those who received any benefit in 2012 was $3,300, although some families received substantially larger amounts. As discussed earlier, 25% of all families that received any aid received benefits of $9,027 or more. (The $9,027 amount was the 75^th percentile—75% of families received less than that and 25% of families received that or more.) These families, while accounting for 25% of all families receiving benefits, accounted for two-thirds of the selected benefit spending. This section examines the question of how these families differ from families that received less in benefits.

Pre-welfare Income-to-Poverty Ratios

The rate of receipt of any need-tested aid was related to both a family's pre-welfare income and its family type. Likewise, the rate of receipt of relatively large amounts of need-tested aid ($9,027 or more) was also related to these characteristics.

Figure 10 shows the estimated rate of receipt of any need-tested aid and receipt of aid totaling $9,027 (75^th percentile) or more, by pre-welfare income in 2012. It shows that lower income groups were more likely than higher income groups to receive either any benefits or benefits at or above $9,027 for the year. However, 53.7% of families with pre-welfare incomes just above the poverty line (pre-welfare incomes of 100% to 149% of poverty) received some benefit, while only 5.8% of families in that income category received $9,027 or more.

Figure 10. Estimated Receipt of Selected Need-Tested Benefits and Benefits Equal to or Exceeding $9,027, by Pre-welfare Income, 2012 (Pre-welfare income-to-poverty ratio is expressed as a percentage)

Source: Congressional Research Service estimates using data from the Census Bureau's 2013 Annual Social and Economic Supplement (ASEC) to the Current Population Survey, supplemented with estimates of program eligibility, receipt, and benefits from the TRIM3 microsimulation model. Notes: Single individuals are considered a family of one in this tabulation.

Families who received $9,027 or more in annual benefits tended to be concentrated among the lowest income groups in 2012. Likewise, total spending on benefits also tended to be concentrated among the lowest income groups. Table 10 shows the estimated composition of all families, families receiving any of the selected need-tested benefits, and families that received $9,027 or more in annual benefits (i.e., those at or above the 75^th percentile), by their pre-welfare income-to-poverty ratios. It also shows total spending on benefits by pre-welfare income-to-poverty ratios. Families with pre-welfare incomes below the poverty line accounted for 55% of all families who received any of the selected need-tested benefits. Thus, 45% of all families who received a need-tested benefit had pre-welfare incomes above the poverty line. However, examining just those families who received $9,027 or more, 85.8% had pre-welfare incomes below the poverty line and 51.3% had pre-welfare incomes that were less than half of the poverty line.

Total spending was also concentrated among lower-income families. While 55% of families who received any benefits had pre-welfare incomes below the poverty line, these same families accounted for 76% of all benefit spending.

Family Characteristics

Figure 11 shows the estimated rate of receipt of any selected need-tested benefit and benefit amounts at or above the 75^th percentile, by family type in 2012. The three family types with the highest rates of benefit receipt—families with children with workers, families with children without workers, and families with an individual with disabilities—also were the most likely to receive benefit amounts at or above the 75^th percentile. An estimated 44.4% of families with children without workers received annual benefits equal to or in excess of $9,027, while 13% of families with children with workers received that level of benefits. Almost 1 in 4 families (23.3%) with an individual with disabilities had annual benefits at or above $9,027.

Figure 11. Estimated Receipt of Selected Need-Tested Benefits and Benefits Equal to or Exceeding $9,027, By Family Type, 2012

Source: Congressional Research Service estimates using data from the Census Bureau's 2013 Annual Social and Economic Supplement (ASEC) to the Current Population Survey, supplemented with estimates of program eligibility, receipt, and benefits from the TRIM3 microsimulation model. Notes: Single individuals are considered a family of one in this tabulation. Families with children are those with a child under the age of 18 who do not have an aged person or individual with disabilities.

Table 11 shows the estimated composition of all families, families receiving any need-tested benefit, and families that received $9,027 or more in benefits during 2012, by family type. It also includes the percentage of total spending accounted for by each family type. The table shows that families receiving benefits at or above the 75^th percentile are more concentrated in the categories of families with children and families containing an individual with disabilities. Families with children and those containing a member with disabilities accounted for more than 8 in 10 (83.2%) of the families that received benefits at or above $9,027 during the year and almost 8 in 10 (78.1%) of all selected need-tested dollars. Families with an aged individual accounted for 14.8% of all families with benefits at or above the 75^th percentile. Families without an aged person, an individual with disabilities, or a child accounted for few (2.1%) of all families receiving benefits above the 75^th percentile and for about 6% of all selected need-tested spending.

Benefit receipt rates—including those at or above the 75^th percentile—can also be examined relative to characteristics often associated with economic well-being, such as job attachment, educational attainment, and family structure. Because families who have an individual with disabilities and families with children collectively account for the largest share of all families that receive relatively large benefits—and also account for a large share of total spending—they are the focus in the next sections of this report of detailed examination of characteristics associated with economic well-being.

Families with an Individual with Disabilities

Historically, individuals with disabilities have been viewed as a population in need of assistance because their impairments often limit them from working enough to be economically self-sufficient.²² Need-tested benefits provide an income supplement for working-age individuals with disabilities who are unable to meet their basic needs for food, clothing, and shelter. For families with a severely disabled child, need-tested benefits are used to defray costs associated with taking care of someone with a severe disability. Disability-related costs can include (1) out-of-pocket medical expenses not covered by Medicaid or private health insurance and (2) lost wages of family members who reduce their attachment to the workforce in order to provide caregiving services for individuals with disabilities.²³

In 2012, an estimated 6 in 10 families who had an individual with disabilities received at least one of the selected need-tested benefits.²⁴ Families with an individual with disabilities represented 41.8% of all families who received benefits at or above $9,027 a year and accounted for 36.7% of all spending from the selected need-tested programs.

Among families with an individual with disabilities that received a selected need-tested benefit, the median benefit amount was $6,236 in 2012, higher than the median benefit for all families that received aid. Thus, these families represent a relatively expensive group, even without taking into account medical costs.²⁵

Table 12 shows selected characteristics of families with individuals with disabilities and their estimated rate of receipt of any selected need-tested benefit and of benefits at or above the 75^th percentile. Family characteristics are typically used in analyses of need-based benefits because most programs consider family rather than individual circumstances. Thus, the indicator used in the table (e.g., work attachment) might not reflect the circumstances of the individual with disabilities, but might reflect those of another family member.

Overall, 60% of families with an individual with disabilities received some need-tested benefits and 23.3% received benefits of $9,027 or more in 2012. These are relatively high rates of receipt, and occur even for those families that do not have characteristics typically associated with economic disadvantage. Only the families with the highest pre-welfare income relative to poverty had low rates of receipt. A substantial share of this population also received benefits at or above the 75^th percentile.

The table shows one characteristic specific to the individual with disabilities: their age.²⁶ It demonstrates that rates of receipt of need-tested benefits decline somewhat with increases in the age of a disabled person.

One possible reason for the higher incidence of need-tested benefit receipt among younger individuals with disabilities is that they often lack a sufficient work history to qualify for social insurance benefits such as Social Security.²⁷ As noted earlier, Social Security income may disqualify an individual or family from receiving a need-tested benefit. Another potential reason is that later onset of disability might allow a family or individual to build up assets and some savings, which also might disqualify a family or individual from need-tested aid.

Table 13 shows the estimated age composition of all individuals with disabilities, those who received benefits from the selected need-tested programs, and those who received Social Security Disability Insurance benefits (SSDI). SSDI coverage is earned through work in covered employment, and provides cash benefits to workers who become disabled after working a sufficient period of time. Younger individuals with disabilities are more likely to receive need-tested aid than older individuals with disabilities (Table 12); thus, younger individuals with disabilities disproportionately received need-tested benefits. In 2012, an estimated 14.2% of individuals with disabilities who received need-tested benefits were children (under age 18) and 6 out of 10 individuals with disabilities who received benefits were under age 50. In contrast, those who were reported on the ASEC as receiving SSDI payments (which are not need-tested) tended to be older. In 2012, an estimated 63% of SSDI recipients identified on the ASEC were aged 50 and older.

Some disabilities are present at birth, such as intellectual and developmental disabilities. On the other hand, some disabilities result from accident, injury, or a chronic disabling condition such as diabetes. Severe mental illness might also be a chronic disabling condition.

Information from the need-tested SSI program indicates that the nature of impairments tends to differ among younger versus older persons with disabilities. In 2012, almost 7 in 10 SSI recipients who were under the age of 50 had primary impairments such as intellectual disabilities, developmental disabilities, and mental impairments.²⁸ In contrast, 42% of those age 50 and older had such disabilities as their primary impairments. In 2012, 22% of those age 50 and older had primary impairments related to muscular-skeletal and connective tissue systems, compared to 6% for individuals under 50. Since younger individuals with disabilities are more likely than older individuals with disabilities to receive need-tested aid, the impairments of individuals with disabilities who receive need-tested aid are often intellectual and developmental disabilities or mental illness.

Families with Children

Families with children have been a focal point of policy for low-income individuals for decades. The Social Security Act of 1935 established grants to states to help them aid families with children who had been deprived of support because of the death, disability, or absence of one parent. Aid was provided so that the remaining parent (usually the mother) did not have to work and could care for the family's children. This policy became increasing controversial over time, leading to a series of changes that culminated with

• a major expansion of aid to low-income, working parents (including two-parent families) through the EITC in the 1980s and 1990s, and later the establishment and expansion of the ACTC; and
• the 1996 welfare reform law that changed the terms of cash assistance for needy families with children and provided additional support for state programs that subsidize child care for low-income families.

Thus, low-income assistance policy for families with children was transformed from providing support for a relatively narrow population (needy families headed by a non-working single mother) to supporting a much wider population, with a focus on requiring and supporting parental work. The broader reach of need-based aid to families with children is reflected in the 2012 data.

Table 14 shows the characteristics of families with children that were estimated to receive at least one of the selected need-tested benefits and were estimated to receive total benefits at or above the 75^th percentile in 2012. In general, receipt rates for any need-tested benefits were high for families with children, even for those with characteristics not often associated with economic disadvantage. For example, an estimated 38% of families with children that included a full-time, full-year worker received at least some benefit from a need-tested program. Among families where the highest level of educational attainment for the adults was a college associate's degree, almost half received a need-tested benefit. On the other hand, those families that had characteristics usually associated with economic disadvantage had a high rate of receipt of relatively large total benefit amounts.

Key takeaways from the table include the following:

• Families with children that lacked an adult worker almost always received a selected need-tested benefit. These families were also likely to receive a large benefit. In 2012, 44.4% of families with children that lacked an adult worker received benefits equal to or above $9,027. Additionally, 55.2% of families that had relatively weak job attachment (part-time work for part of the year) received a benefit equal to or above $9,027.
• While some families with relatively high levels of educational attainment (e.g., college credential) received selected need-tested benefits, receipt of relatively large benefit amounts (at or above the 75^th percentile) was concentrated among those with at most a high school diploma or those that lacked a high school diploma.
• Families with children headed by a single woman had higher rates of receipt of selected need-tested benefits and total benefits at or above the 75^th percentile than did families headed by married couples or men. However, close to 1 in 4 (38.7%) families headed by married couples or men received at least some selected need-tested aid, although such families were unlikely to receive benefits at or above the 75^th percentile. In 2012, 9.6% of families with children headed by a married couple or man received benefits at or above the 75^th percentile, compared to 33.8% of families with children headed by a woman.
• Families with children headed by racial and ethnic minorities except Asian-Americans were more likely than non-Hispanic whites to receive need-tested benefits and benefits at or above the 75^th percentile.
• Large families with children were much more likely than smaller families with children to receive selected need-tested benefits at or above the 75^th percentile of total benefit receipt.

Benefit Receipt by Program

Total need-tested benefit amounts are typically higher for families that combine benefits from multiple programs than they are for families that receive benefits from only one program. Moreover, families that receive benefits from certain programs are more likely to receive benefits from multiple programs. Thus, the percentage of families who are in the top 25% of need-tested benefit receipt varies by program.

Table 15 shows for families that received benefits from each of the nine programs, estimates of the median benefit total received from all programs and the median benefit from that specific program. The table also shows the percentage of families receiving benefits from each program that were in the top 25% of need-tested benefit receipt in 2012. Finally, the table shows the number of programs from which these families received benefits. Programs are ranked by the percentage of recipient families that received total benefits at or above the 75^th percentile ($9,027).

The top-ranked programs in the table—subsidized child care and TANF—are intended for families with children. For families receiving subsidized child care, the median total need-tested benefit (from all programs) was $14,810. Families receiving TANF had median total need-tested benefits of $13,937. Families participating in both programs were very likely to combine benefits from multiple programs, and a large share of the total benefits for those families came from programs other than child care or TANF. Families receiving subsidized child care were also very likely to receive benefits from the two refundable tax credits, SNAP, and sometimes TANF. The benefit packages for TANF families were more varied, though most TANF families also received SNAP.

The third-ranked program was subsidized housing. The median total need-tested benefit for families in subsidized housing was $11,349. However, the housing assistance benefit itself accounted for a large share of this total benefit. While more than 8 in 10 families in assisted housing received benefits from more than one program, only 14% received benefits from five or more programs. This compares with 4 in 10 TANF families receiving benefits from five or more programs and half of families receiving subsidized child care getting benefits from five or more programs.

The programs that ranked lowest in the table in terms of their percentage of families with total need-tested benefits in the top 25% (at or above $9,027) were families receiving the two refundable tax credits, LIHEAP, and SNAP.

Conclusion

Many people in the United States were eligible for need-tested benefits in 2012. Four in ten people were estimated to be eligible for benefits from at least one of the nine need-tested programs examined in this report. However, not all persons eligible for need-tested benefits actually received them. Among the programs examined in this report, an estimated 70% of eligible families actually received SNAP and 65% of eligible families received WIC in 2012. However, the estimated rate of benefit receipt among eligible persons was 28% for TANF cash assistance, 22% for LIHEAP, 18% for subsidized housing, and 17% for CCDF (based on eligible children).

In 2012, one in three people were estimated to have actually received benefits from at least one of these programs. Many of these families had characteristics not typically associated with economic disadvantage; a substantial portion of families that received aid had pre-welfare incomes above the poverty line in 2012. However, many of these families received relatively small benefits.

Benefit receipt differed considerably depending on family characteristics. Families with an aged member and those comprising non-aged, non-disabled childless adults represented a disproportionately small share of all families receiving assistance and total benefit dollars spent. For the aged, this partially reflects the role Social Security plays in providing income and Medicare plays in providing health care coverage. The aged also have a relatively low rate of take-up of the need-tested benefits for which they are eligible. This report focuses on nonmedical need-tested benefits: it does not reflect the protection that Medicaid offers in providing expensive long-term care services to the aged should they exhaust their income and assets to obtain them.

Families containing non-aged, non-disabled childless adults also received relatively small amounts of need-tested benefits. They are ineligible for benefits from the programs that are available only to families with children (ACTC, TANF, child care, and WIC). This group is also eligible for a relatively small EITC. In 2012, a single childless adult could receive a maximum EITC of $475 for the year, while a tax filer with three children was eligible for a credit of up to $5,891.

The wide reach of need-tested benefits is attributable to two major groups of families: those that contain an individual with disabilities and those with children. Families with a non-aged, disabled member totaled 18.6 million (14% of all families), and 60% of them received benefits from at least one need-tested program in 2012. Individuals are typically considered disabled if they have a physical or non-physical impairment that prevents work. Moreover, the presence of an individual with disabilities might also limit the ability of other family members to work. Families with a non-aged disabled member accounted for $89.5 billion, or 37% of total spending, for the nine need-tested benefit programs examined in this report.

Families with children accounted for the greatest amount and share of need-tested spending (41%). Many policy debates about low-income people focus on families with children and reflect the tension between trying to alleviate high levels of financial need while not undercutting expectations that non-disabled parents work. Both goals can be expensive. The large share of families with children that received need-tested aid is attributable to assistance provided to families with a working adult member. These families totaled 29.8 million (23.2% of all families), and 45% of them received benefits from at least one need-tested benefit program in 2012. Many of these families did not have characteristics that are often associated with economic disadvantage. Families with children who have working adults accounted for $91.6 billion, or 48%, of total spending for the nine need-tested benefit programs examined in this report.

Families with children without an aged or disabled member who also lacked an adult worker were a relatively small group in 2012 (fewer than 1 million families). However, they were the most likely group to be in the top 25% of families in terms of benefit amounts received, with 44% of such families having benefits of $9,027 or more.

Though only 25% of families receiving need-tested benefits had benefits of $9,027 or more, they accounted for 64% of total spending for the nine need-tested programs discussed in this report. Families that received large benefits tend to have characteristics that traditionally have been associated with economic disadvantage and discussed in past policy debates: being in poverty or even "deep poverty" (pre-welfare incomes under 50% of the poverty line), disabilities, weak or no job attachment, low levels of educational attainment, and headed by single mothers.

Data and Methods

Data Sources and Limitations

This report examines eligibility, participation, and benefit receipt in calendar year 2012. Estimates in this report were derived using data from the March 2013 Annual Social and Economic Supplement (ASEC) to the Current Population Survey (CPS), augmented by information from the Transfer Income Model, version 3 (TRIM3), funded by the U.S. Department of Health and Human Services (HHS) and maintained at the U.S. Census Bureau.

ASEC

The ASEC is a household survey of the non-institutionalized population conducted by the Census Bureau in March of each year. The non-institutionalized population excludes those persons residing in institutional group quarters such as adult correctional facilities, juvenile facilities, skilled-nursing facilities, and other institutional facilities such as mental (psychiatric) hospitals and in-patient hospice facilities. The non-institutionalized population includes members of the Armed Forces living in civilian housing units on a military base or in a household not on a military base.

The ASEC asks respondents to report on household members' demographic characteristics, work experience, and earnings in the prior year. In addition, the ASEC asks respondents about receipt of certain government benefits in the prior year, including six of the need-tested programs examined in this report: SSI, TANF, SNAP, WIC, housing assistance, and LIHEAP. While the ASEC does not directly ask questions about federal taxes, the Census Bureau provides its own estimates of the ACTC and the EITC (based on data provided by respondents about their household's economic and demographic circumstances) for inclusion in ASEC estimates of after-tax income. Data on CCDF benefits, however, are neither collected nor estimated by the Census Bureau in conjunction with the ASEC.

As with any household survey, ASEC data are subject to sampling error, as well as error from respondents misreporting household members' circumstances. Research has found that need-tested benefits are commonly under-reported on the ASEC. Respondents report fewer household members receiving need-tested benefits than are recorded by federal or state administering agencies. For example, in 2005 it was estimated that the ASEC captured 57% of SNAP recipients, 59% of TANF recipients, and 74% of SSI recipients.²⁹ In the case of housing assistance, it appears that there is under-reporting of benefits by families receiving federal housing assistance, and over-reporting of receipt of assistance by families who are not receiving federal housing assistance, perhaps because they are receiving some form of state or local assistance.³⁰ In addition, benefit amounts reported by ASEC respondents typically fall below the benefit amounts recorded in agency expenditure data.

TRIM3

TRIM3 is a microsimulation model for government benefit receipt that is primarily funded by the U.S. Department of Health and Human Services (HHS) and maintained at the Urban Institute.³¹ Microsimulation models apply a set of program rules to individual or family units to simulate various elements—in this case, eligibility for certain need-tested benefit programs, likelihood of receiving one or more benefits, and the amount of any benefits received. Simulations at the individual or family level are then aggregated to allow for comparisons across families and programs.

Estimates in this report were generally derived from the TRIM3 microsimulation model using administrative data on program rules and survey data on individuals and families from the March 2013 ASEC. This report used TRIM3 to adjust ASEC survey responses for under-reporting of applicable need-tested benefits. This was done by using family characteristics and likelihoods of benefit receipt to "assign" benefits to a portion of the families estimated to be eligible for benefits but not reported as receiving them on the ASEC.

In addition to using TRIM3 to adjust ASEC data for certain programs, this report also relied on TRIM3 estimates, rather than Census Bureau estimates, of federal tax benefits from the ACTC and the EITC. This was done so that the assumptions and methods underlying the estimates of the refundable tax credits were aligned with those of the government spending benefit programs.

Finally, this report used TRIM3 to estimate the receipt and amount of CCDF subsidies for families. The ASEC does not collect information on CCDF benefits, but the TRIM3 child care estimates were derived using data on income, work, and age of children, as reported on the CPS.

The TRIM3 microsimulation model generally brings recipient counts and total benefits (in dollars) in line with aggregate administrative totals for the need-tested benefit programs examined in this report. However, TRIM3 is a model and all models have limitations. For instance, TRIM3 makes a number of simplifying assumptions in order to estimate monthly income (often necessary for eligibility establishment under program rules) because the ASEC only asks respondents about annual income. To do this, TRIM3 "allocates" annual income amounts for individuals across the months of the year based on a variety of factors (e.g., number of weeks of employment). However, such allocations of income may not always be perfect, meaning that there is some margin of error in TRIM3 estimates of eligibility, participation, and benefits. Additionally, housing assistance estimates from TRIM3 rely on weighted statewide averages for fair market rents and income eligibility thresholds, rather than the local area data that are actually used for program administration purposes.

While TRIM3 corrects for under-reporting of benefit receipt on the ASEC, it generally does this on a program-by-program basis. Because of a lack of administrative data on multiple program participation, TRIM3 is only able to specifically target real-world overlap for a subset of program combinations when correcting for under-reporting.

Table A-1 summarizes some key estimation issues, by program, for the data used in this report.

Key Assumptions

Annual Estimates of Income and Benefit Receipt

The estimates of eligibility, receipt, and benefits in this report are based on program rules. Many of the need-tested programs examined here determine eligibility and benefits on a monthly basis. As a result, the number of people who received benefits at any point over the course of the year tends to be greater than the number of people receiving benefits in any one month.

The estimates in this report reflect receipt of one or more benefits at any point during 2012 (i.e., total recipients includes people who were estimated to receive one or more benefits for only part of the year). As such, this report's estimated number of recipients tends to be greater than the number of recipients reported by administrative data, which often show monthly or monthly average participation. In addition, benefit amounts in this report are estimated annual benefits received by a family over the course of the entire year.

The information on refundable tax credits from the EITC and the ACTC represents the amounts earned during the year, which are actually paid to families once a year when they receive their tax refunds, generally in the following year. For example, tax credits earned in 2012 would generally be received in early 2013 when taxpayers filed their 2012 federal income tax returns. This also comports to the way the refundable tax credits are considered in analyses of family income and poverty.

Definition of "Family"

This report presents estimates of benefits received at the family level. The family level is used because it best represents an economic unit—people who pool together financial resources, consume goods and services, and make economic decisions together. For purposes of this report, the family unit includes all those in a household who are related by blood, marriage, or adoption, as well as cohabiting couples, relatives of cohabiters, and unrelated children who are cared for by the family (e.g., foster children). This family construct is consistent with the family unit used for the Supplemental Poverty Measure (SPM), but differs from the family unit used for the official poverty measure, which excludes cohabitating couples, relatives of cohabiters, and unrelated children. The family unit used in this report may also differ from the unit used by programs for eligibility determination purposes (e.g., the EITC and the ACTC use tax filing units, which may be different than the family unit used for benefit estimates in this report). However, use of a common family unit allows this report to examine receipt of benefits across all programs in a comparable manner that would not otherwise be possible.

Supplemental Poverty Measure

The SPM is an alternative poverty measure developed by the U.S. Census Bureau and the Bureau of Labor Statistics that is based on a broader range of income sources and costs than the official poverty measure.³² The official poverty measure counts only earnings and cash benefits (e.g., Social Security and unemployment benefits). In addition to these, the SPM counts certain in-kind benefits (e.g., SNAP, WIC, housing assistance, LIHEAP) and tax credits (e.g., the EITC, the ACTC) and subtracts a number of necessary expenses (e.g., taxes, work-related child care, commuting costs, and medical out-of-pocket expenses) from a family's total resources to arrive at a measure of disposable income available to meet the family's basic needs. The SPM also sets separate income/resource thresholds for homeowners with a mortgage, homeowners without a mortgage, and renters; these thresholds are adjusted for variations in housing costs by geographic area (metropolitan and nonmetropolitan areas in a state).

CRS used the SPM for contextual purposes when considering the economic circumstances of families who are eligible for or receiving need-tested benefits. TRIM3 uses each program's family unit, income eligibility, and benefit computation rules for determining its estimates for eligibility and benefit receipt of need-tested programs.

Benefit Levels

Readers should be aware of additional considerations regarding the benefit levels presented in this report.

One consideration is that all in-kind benefits were monetized for the purpose of this analysis and several assumptions were made in that process. For example, in the case of housing assistance, the benefits provided by the Section 8 Housing Choice Voucher (HCV) program, the project-based Section 8 rental assistance programs, and the public housing program were all calculated the same way, using an approximation of the method used for calculating the maximum benefit a family could receive under the Section 8 HCV program. This approach is tied to the market cost of housing and is commonly used by researchers and policy analysts. However, the public housing and project-based Section 8 rental assistance programs provide affordable rental units to families rather than the vouchers provided by the Section 8 HCV program for use in the private market. While the dollar value of a voucher is fairly clear, the dollar value of an affordable rental unit is less clear; thus, this approach may over- or under-estimate the dollar value of the benefit received by a resident of public housing or project-based Section 8 rental assistance housing.

Estimates of the Refundable Tax Credits

In conducting the ASEC, the Census Bureau does not ask survey respondents about their federal tax liabilities or whether they benefit from tax credits. Therefore, estimates must be made in order to incorporate tax liabilities and estimates into analyses based on the ASEC.

TRIM3 applies the federal tax rules in effect during the year to the population for that year. This involves taking the ASEC's information on individuals and families and family structure and creating a "tax filing" unit (i.e., the tax filer and their spouse and dependents, if applicable). TRIM3 then applies the rules for counting income, deductions and exemptions, eligibility for tax credits, and tax rates to determine the unit's tax liabilities and/or tax credits.

It is well known in the research community that estimates of the total dollars and number of filers receiving refundable tax credits based on information from the ASEC is well below the true dollar amount and number of filers who actually claim these credits. For 2012, the TRIM3 estimate of the total dollars of the EITC is $46.4 billion. This is compared to Statistics of Income (SOI) data, based on tax returns that showed a total $64.1 billion. Note that this is not a function of using TRIM3 estimates; the Census Bureau's own estimates of the EITC using the ASEC were similar to those made by TRIM3.³³

Research has been done on why EITC estimates from the ASEC differ from those reported on the SOI, but the reasons for the different estimates have not been conclusively explained. Factors that have been raised as potentially explaining this discrepancy include EITC "compliance" issues, where taxpayers claim an EITC or higher EITC amount that they are not eligible for; misclassification of tax units from ASEC family structure information; or under-reporting of earnings on the ASEC that are used to estimate the EITC.

Similar shortfalls and results are also observed for the ACTC.

Medicaid

This report uses data from the TRIM3 microsimulation model for 2012 to provide information about who is eligible for and receiving benefits through need-tested programs and the value of these benefits. Medicaid is the largest need-tested program, but it was not included in most of the analysis provided throughout the report because (1) TRIM3 does not provide information on enrollment and benefit values of Medicaid; (2) using the 2012 information would miss the major expansion of Medicaid eligibility under the Patient Protection and Affordable Care Act (ACA, P.L. 111-148 as amended); and (3) there are issues with valuing Medicaid.

This appendix provides a brief overview of the Medicaid program followed by sections about TRIM3 and Medicaid, Medicaid participation rates, and the value of Medicaid coverage.

Medicaid Overview

Medicaid is a means-tested entitlement program that finances the delivery of primary and acute medical services as well as long-term services and supports (LTSS).³⁴ In FY2014, Medicaid is estimated to have provided health care services to 63 million individuals³⁵ at a total cost of $494 billion, with the federal government paying $299 billion (about 61%) of that total.³⁶

Medicaid coverage includes a wide variety of preventive, primary, and acute care services as well as LTSS.³⁷ Not everyone enrolled in Medicaid has access to the same set of services. Different eligibility classifications determine available benefits.

To be eligible for Medicaid individuals must meet both categorical (e.g., aged, individuals with disabilities, children, pregnant women, parents, certain non-aged childless adults) and financial (i.e., income and sometimes assets limits) criteria.³⁸ Historically, Medicaid eligibility had generally been limited to certain low-income children, pregnant women, parents of dependent children, the aged, and individuals with disabilities; however, as of January 1, 2014, states have the option to extend Medicaid coverage to most non-aged, low-income individuals through the Medicaid expansion enacted as part of the ACA.³⁹

Medicaid spending per full-year equivalent enrollee was $7,236 in FY2011.⁴⁰ However, Medicaid spending per enrollee varies significantly by population group. The following are per-enrollee spending amounts by major population group:

• Children = $2,854
• Adults = $4,368
• Aged = $16,236
• Disabled = $19,031.⁴¹

One reason the aged and disabled populations have higher per-enrollee expenditures is because these populations consume most of LTSS. The Medicaid per-enrollee spending amount for enrollees with no LTSS was $4,332 in FY2011, while per-enrollee spending for enrollees with LTSS was $44,719. Among the enrollees with LTSS, the Medicaid per-enrollee spending ranged from $25,837 for those with no institutional or home- and community-based service waiver⁴² services to $66,006 for those with LTSS consisting of only institutional services.⁴³

TRIM3 and Medicaid

For 2012, the TRIM3 microsimulation model only estimates which individuals are likely eligible for Medicaid coverage. In that year, TRIM3 estimated that 65 million persons (21% of the total U.S. population) were eligible for Medicaid. It ranked third only to SNAP and LIHEAP in terms of number of people eligible in the non-institutional population.

The Medicaid eligibility information in TRIM3 has a few limitations. First, TRIM3 only includes data for non-institutionalized individuals, but some Medicaid enrollees receive institutional services.⁴⁴ Second, TRIM3 simulates Medicaid eligibility for full Medicaid coverage, but some states provide limited benefit coverage (e.g., inpatient hospital-only coverage or preventative care-only coverage) to certain populations. Third, the eligibility information in TRIM3 is for 2012, and therefore doesn't take into account the ACA Medicaid expansion. For the states that have implemented the ACA Medicaid expansion (30 states and the District of Columbia), Medicaid eligibility looks significantly different in 2015 than it did in 2012, with substantially more parents, individuals with disabilities, and nonaged adults without dependent children being eligible.⁴⁵

Additionally, for 2012 TRIM3 did not include information about Medicaid enrollment (i.e., who is receiving Medicaid coverage) and the value of Medicaid coverage. Thus, it is not possible to treat Medicaid in the same manner as the other need-tested programs discussed in this report. Instead, the following sections summarize other research about Medicaid participation rates and the value of Medicaid coverage.

Participation Rate

The participation rate (or take-up rate) of Medicaid refers to the percentage of people eligible for the program that choose to enroll in Medicaid. Research indicates the Medicaid participation rates for adults tend to be lower than the participation rates for children.

The average estimates of Medicaid participation rates for adults from a number of studies from 1999 through 2010 range from 52% to 81%.⁴⁶ However, the participation rates vary significantly by state. For instance, a study of the Medicaid participation rate in 2009 estimated a national participation rate of 68%, with state Medicaid participation rate estimates ranging from 51% in Nevada to 94% in Massachusetts.⁴⁷

Research suggests that participation rates among children in Medicaid and the State Children's Health Insurance Program (CHIP)⁴⁸ tend to be greater than the Medicaid participation rate for adults, and the rate has grown significantly in recent years. The Medicaid and CHIP participation rate for children was 82% in 2008 and increased to 88% in 2012.⁴⁹ As with adults, these participation rates vary by state, ranging from 80% or lower in two states to 90% and higher in 21 states and the District of Columbia in 2012.⁵⁰

Research has also found that Medicaid participation rates are higher among non-whites,⁵¹ individuals in families with lower incomes,⁵² and people with health-related limitations (such as SSI recipients).⁵³ Also, Medicaid participation rates were found to be lower for older children (ages 13 to 18) relative to younger children,⁵⁴ and for nonaged, childless adults without functional limitations relative to all non-aged adults.⁵⁵

It is important to note that participation rates provide the number of people eligible for the program that choose to enroll. However, not all Medicaid enrollees access services.

Value of Medicaid

As noted previously in this report, estimating the dollar value of health benefits to families and individuals is particularly difficult. A seminal National Academy of Sciences report issued in 1995 examining potential changes in the measurement of family well-being for poverty analysis stated:

The issue of how to treat medical care needs and resources in the poverty measure has bedeviled analysts since the mid-1970s, when rapid growth in the Medicare and Medicaid programs (and in private health insurance) led to a concern that the official measure was overstating the extent of poverty among beneficiaries because it did not value their medical insurance benefits. Yet after two decades of experimentation, there is still no agreement on the best approach to use.⁵⁶

For most of the cash, food, and housing benefits discussed in this report, the amount a family receives in benefits can be estimated based on their circumstances as reported on the ASEC. That is, a dollar value of what is spent on benefits for the family can be estimated based on available data. This is not so for Medicaid.

Actual Medicaid spending for a family is determined by the amount and type of medical services it consumes. Moreover, the medical services consumed by a family are related to the health of family members. Medicaid is also paid to the provider of services. For these reasons, Medicaid benefits do not simply add to other family income to determine the family's well-being. That is, more dollars are spent on an individual when an individual is less healthy; those extra dollars spent do not mean that the individual or the individual's family is "wealthier."

However, Medicaid does have economic value to families that receive benefits. Historically, the valuation methods have relied on per-enrollee Medicaid spending for individuals in a given demographic group. There are two major approaches that have been used to value Medicaid: a method that uses its "average cost" to the government and another that attempts to measure its "fungible value."

Average Cost Method

The first method uses the average cost of Medicaid to the government. This is the valuation method used most recently by the Congressional Budget Office (CBO) in its analysis of the distribution of income.⁵⁷ It was also historically used as the "market value" of Medicaid by the U.S. Census Bureau. Usually, the average cost is estimated based on the person's state, age (aged or child), and disability status.

It is not clear that the average cost of coverage represents the value of that coverage to an individual. If the individual would prefer that the average cost of coverage be paid in cash instead, the average cost of coverage would overstate the value health insurance has to the individual. The 2008 Oregon Health Insurance Experiment examined enrollees' willingness to pay for coverage. It found that enrollees would not be willing to cover the cost to the government of Medicaid coverage, but they would be willing to pay $0.2 to $0.4 per dollar of government spending on their Medicaid coverage.⁵⁸

However, it is also possible that the bundle of health care services covered by Medicaid could not be purchased without coverage at the average cost of coverage. If this is the case, the value of coverage could be understated. Additionally, Medicaid coverage has spillover effects to other people, and thus its value to society may be greater than its value to individual recipients. The 2008 Oregon Health Insurance Experiment found that Medicaid coverage for low-income, uninsured adults increased health care use (i.e., outpatient care, preventive care, prescription drugs, hospital admissions, and emergency room visits); improved self-reported health; and reduced depression.⁵⁹ Medicaid coverage also has benefits to health care providers, who otherwise might be required to provide uncompensated services.⁶⁰

Fungible Value of Medicaid

The second method historically used to value Medicaid benefits attempts to measure its "fungible value." The fungible value is supposed to represent the amount of resources of Medicaid enrollees' families that are freed up for other uses by Medicaid coverage.⁶¹ That is, Medicaid coverage does not directly increase a household's income, but it can increase an individual's (or family's) ability to consume other goods and services because the individual (and family) does not have to use income to pay for health care services.⁶² Fungible values were based on the "average cost" of Medicaid per enrollee in a person's demographic group, adjusted for food and housing costs and family incomes. CBO has also used the fungible value of Medicaid in its estimates on the distribution of household income.⁶³

The nature of medical coverage means that Medicaid is not a relatively fungible benefit. Medicaid coverage for relatively low-cost services (e.g., a doctor's visit) might free up money for the purchase of other goods and services. However, Medicaid coverage for a very expensive benefit (e.g., a surgery) is unlikely to affect a family budget in a similar way. Such expenses would likely overwhelm a low-income family's current income and potentially their assets.

Medical Care Expenses and Economic Burden and Risk

As discussed above, in 1995 a National Academy of Sciences panel recommended revising poverty measurement. It recommended not placing a dollar value on either private health insurance or publically provided health care from programs such as Medicaid when computing poverty statistics. Rather, it recommended subtracting out-of-pocket medical expenses from family resources, so that the new poverty measure would consider a family's disposable income available to purchase non-medical goods and services. The SPM implemented by the U.S. Census Bureau followed the recommendation of not including the value of medical insurance as a family financial resource and deducting out-of-pocket medical care expenditures. Beginning with the release of the 2014 poverty data, the Census Bureau discontinued providing the average cost and the fungible value of benefits from Medicaid and Medicare.

In the fall of 2010, the Department of Health and Human Services (HHS) asked the National Academy to convene a panel to examine "the state of the science in the development and implementation of a new measure of medical care risk as a companion measure to the new Supplemental Poverty Measure." ⁶⁴ That panel recommended developing measures of medical care economic burden and medical care risk, separate from poverty, to capture effects on family finances and economic risk to families of having no or inadequate health insurance coverage. In terms of actual burden, the panel recommended that the Census Bureau provide information that would compare a family's or individual's out-of-pocket medical expenses with its resources available for medical care. These resources would include both current income and a portion of the family's or individual's liquid assets. Under the recommendation, the medical economic burden would be used to measure how medical out-of-pocket expenses affect a person's or family's poverty status—how much they reduce the resources available to the family to purchase non-medical goods and services relative to poverty.

Measuring medical care economic risk would entail measuring the likelihood that a family would incur a specified level of medical out-of-pocket expenses. The panel did not make a specific recommendation on the medical care economic risk, though it did recommend that the relevant federal agencies undertake research to develop that measure. The medical care burden and risk measure has yet to be implemented.

Family Categories in this Report

This report divides families into six categories: (1) families with an aged member (65 or older); (2) families with an individual with disabilities; (3) families with children and no earners; (4) families with children and earners; (5) other families, without a member who is aged, disabled, or a child, with no earners; and (6) other families, without a member who is aged, disabled, or a child, with earners. As discussed previously in this report, any individual family is placed into only one category even if it may meet the criteria of another category.

Families are assigned sequentially to a category based on the ordering of families listed above. For example, if a family had any aged member, it is placed in the first category (families with an aged member). This is the case even if the family also has an individual with disabilities or a child.

The method described above creates mutually exclusive categories of families that sum to the total number of families in the population. However, it fails to reflect the complexity of families and family structures present in the population.

Table C-1 shows families with an aged member by the presence of individuals with disabilities and children. The table shows that of the 32.5 million families with an aged member, 14.1% of them had either an individual with disabilities or a child. Moreover, these families also had higher poverty rates based on pre-welfare income and higher rates of need-tested benefit receipt than did families with an aged member who did not have an individual with disabilities or a child.

Table C-2 shows how many families without an aged member but with an individual with disabilities include a child. Of the total 18.6 million families with a non-aged disabled member, close to one-third (32.3%) also had a child. Almost 8 in 10 families with a non-aged disabled member and a child received a need-tested benefit compared with about half of families with a non-aged disabled member but no children.

Author Contact Information

Acknowledgments

This report is the product of a team effort that included individuals in addition to those listed as its authors. CRS staff that contributed to this report included [author name scrubbed], [author name scrubbed], Kirsten Colello, [author name scrubbed], Austin Frerick (formerly of CRS), Chris Lewis, [author name scrubbed], [author name scrubbed], and [author name scrubbed]. Amber Wilhelm produced the figures in this report. Additionally, the report benefited from the technical help from staff of the Urban Institute on the use of the TRIM3 microsimulation model, particularly Linda Giannarelli and Laura Wheaton.

Footnotes

1.	For example, see CRS Report R43863, Federal Benefits and Services for People with Low Income: Programs and Spending, FY2008-FY2013, by [author name scrubbed] and [author name scrubbed].
2.	CRS Report R43409, Federal Minimum Wage, Tax-Transfer Earnings Supplements, and Poverty, by [author name scrubbed], [author name scrubbed], and [author name scrubbed].
3.	This report examines benefit receipt in 2012 because it is the most recent year for which adequate data are available.
4.	For a discussion, see Bruce D. Meyer, Wallace K.C. Mok, and James X. Sullivan, "Household Surveys in Crisis," Journal of Economic Perspectives, vol. 29, no. 4 (Fall 2015), pp. 199-226.
5.	Documentation from TRIM3 can be found at http://trim3.urban.org/T3Technical.php. For a discussion of microsimulation and its use in policy analysis, see Gordon H. Lewis and Richard C. Michel, ed., Microsimulation Techniques for Tax and Transfer Analysis (Washington, DC: Urban Institute Press, 1990).
6.	The programs here focus on some of the large need-tested programs that provide economic assistance to families. The report does not include school meal programs, because information on their benefits is not available in TRIM3. The report does not examine receipt of student financial assistance or need-based veterans' benefits. The report also does not include data on programs receipt of employment and training or social services.
7.	For a discussion of the history and issues of measuring poverty, see CRS Report R41187, Poverty Measurement in the United States: History, Current Practice, and Proposed Changes, by [author name scrubbed].
8.	This has been done for computing alternative measures of poverty since the 1980s. For a discussion of different valuation techniques, see Timothy M. Smeeding, Alternative Methods for Valuing Selected In-Kind Transfer Benefits and Measuring Their Effect on Poverty, U.S. Department of Commerce, Bureau of the Census, Technical Paper 50, March 1982.
9.	For more information about the Medicaid program, see CRS Report R43357, Medicaid: An Overview, coordinated by [author name scrubbed].
10.	CRS Report R43634, Temporary Assistance for Needy Families (TANF): Eligibility and Benefit Amounts in State TANF Cash Assistance Programs, by [author name scrubbed].
11.	The non-institutionalized population excludes those persons residing in institutional group quarters such as adult correctional facilities, juvenile facilities, skilled-nursing facilities, and other institutional facilities such as mental (psychiatric) hospitals and in-patient hospice facilities. The non-institutionalized population includes members of the Armed Forces living in civilian housing units on a military base or in a household not on a military base.
12.	A rule that imposes a time limit (3 months in a 32-month period) on SNAP receipt for nondisabled adults aged 18 to 54 without children who lack employment or are not in training was suspended in most jurisdictions during the deep recession of 2007-2009 and remained suspended in 2012. This rule has subsequently been restored in an increasing number of jurisdictions.
13.	That is, the 62.9 million taxpayers eligible for the EITC are all assumed to receive the EITC. However, as discussed in "Estimates from the Transfer Income Model Version 3 (TRIM3)" estimates from TRIM3 and the ASEC tend to find fewer eligible tax filers and EITC dollars than what is actually reported on federal income tax returns. Therefore, estimates of the EITC from TRIM3 and the ASEC do not reduce the number of those eligible for EITC based on a probability that the tax filer actually receives the credit. Research has provided evidence that less than 100% of eligible tax filers actually received the credit. The Internal Revenue Service (IRS) has estimated that for 2012, 80% of those eligible to receive the EITC actually received it. See https://www.eitc.irs.gov/EITC-Central/Participation-Rate.
14.	See Paul S. Davies and T. Lynn Fisher, "Measurement Issues Associated with Using Survey Data Matched with Administrative Data from the Social Security Administration," Social Security Bulletin, vol. 69, no. 2 (July 2009), http://www.ssa.gov/policy/docs/ssb/v69n2/v69n2p1.html.
15.	Another problem in determining the SSI eligible population is that the CPS does not collect the data necessary for TRIM3 to determine potential eligibility for all children under age 15. The only children under age 15 who are identified as "SSI eligible" are children in receipt of SSI benefits.
16.	This rate of receipt is lower than the "participation rate" published by the U.S. Department of Agriculture (USDA). The TRIM3-based rate differs both conceptually and technically from the USDA participation rate. The conceptual difference is that the USDA rate accounts only for families with incomes below the federally established 130% of the poverty level. The USDA rate does not take into account families made eligible through expanded categorical eligibility, which provides the option to states to set an income limit as high as 200% of poverty. The TRIM3 rate examines the estimated rate of receipt with those estimated to be eligible for SNAP under state rules. The USDA report also uses a different estimating method. The USDA participation rate and the method used to estimate it is described in Karen E. Cunnynham, Reaching Those in Need: Estimates of State Supplemental Nutrition Assistance Program Participation Rates in 2012, U.S. Department of Agriculture, February 2015.
17.	For a review of the literature on the effect of need-tested benefits on work behavior prior to the changes in these programs that culminated with major expansions of the EITC and the creation of TANF in the 1990s, see Robert Moffitt, "Incentive Effects of the U.S. Welfare System: A Review," Journal of Economic Literature, vol. 30, no. 1 (March 1992), pp. 1-61. GAO, in their 2015 review of low-income assistance programs, concluded that the changes in programs since then have enhanced incentives for people to join the labor force. See U.S. Government Accountability Office, Federal Low-Income Programs, Multiple Programs Target Diverse Populations and Needs, 15-516, July 2015.
18.	Covered employment is all employment for which earnings are creditable for Social Security purposes. Most employment is covered by Social Security. Workers who are not covered by Social Security include some state and local government employees and federal civilian workers hired before 1984.
19.	For this analysis, having at least one adult worker means any adult (age 18 or older) who worked at any time, even for one week, in 2012. Thus, families without a worker represent families where no adult worked at all in 2012, a relatively rare event.
20.	See Norma B. Code and April Yanyuan Wu, What Impact Does Social Security Have on the Use of Public Assistance Programs Among the Elderly, Center for Retirement Research at Boston College, May 2014.
21.	In this report, dollars received represent the face value of benefits. This is not necessarily the "value" that a recipient family would place on the benefit. See "The "Value" of Benefits to Individuals and Families".
22.	Edward D. Berkowitz and Larry DeWitt, The Other Welfare: Supplemental Security Income and U.S. Social Policy (Cornell University Press, 2013), p. 17.
23.	Anne DeCesaro and Jeffrey Hemmeter, "Unmet Health Care Needs and Medical Out-of-Pocket Expenses of SSI Children," Journal of Vocational Rehabilitation, vol. 30, no. 3 (February 4, 2009), pp. 177-199. See also U.S. Government Accountability Office, SSI Children: Multiple Factors Affect Families' Cost for Disability-Related Services, HEHS-99-99, June 28, 1999, http://www.gao.gov/assets/160/156618.pdf.
24.	The total population of individuals with disabilities is based partially on survey respondents' answers to questions about work-related disabilities and partially based on receipt of disability-based benefits. Some issues related to this are discussed in the "Family Characteristics" section of this report.
25.	Individuals with disabilities represent the Medicaid enrollment group with the highest per-enrollee cost. See discussion in Appendix B.
26.	The population in the table is limited to examining the age of adults with disabilities because of the limitation of the ASEC. The ASEC does not ask children under the age of 15 questions about whether their work is limited by disability. Thus, the disabled population under the age of 15 is those who receive SSI benefits, need-tested benefits.
27.	See Kalman Rupp, Jeffrey Hemmeter, and Paul S. Davies, "Longitudinal Patterns of Disability Program Participation and Mortality Across Childhood SSI Award Cohorts," Social Security Bulletin, vol. 75, no. 1 (February 2015), http://www.ssa.gov/policy/docs/ssb/v75n1/v75n1p35.html. See also Michelle Stegman Bailey and Jeffrey Hemmeter, "Characteristics of Noninstitutionalized DI and SSI Program Participants, 2013 Update," Research and Statistics Note No. 2015-2, September 2015, http://www.ssa.gov/policy/docs/rsnotes/rsn2015-02.html.
28.	Social Security Administration, SSI Annual Statistical Report, SSA Publication No. 13-11827, July 2013, Table 35. As noted in the text, the SSI program classifies individuals based on their primary impairments. A recipient may have secondary conditions or comorbidities.
29.	Laura Wheaton, Underreporting of Means-Tested Transfer Programs in the CPS and SIPP, Urban Institute, 2008.
30.	Paul D. Johnson, Trudi Renwick, and Kathleen Short, Estimating the Value of Federal Housing Assistance for the Supplemental Poverty Measure, U.S. Census Bureau, December 2010.
31.	A discussion of TRIM3 can be found on the Urban Institute website at http://trim3.urban.org/T3Welcome.php.
32.	For more information on the SPM, see the "The Research Supplemental Poverty Measure" section of CRS Report RL33069, Poverty in the United States: 2013, by [author name scrubbed].
33.	For example, the EITC total estimated by the Census Bureau for 2012 was $45.4 billion.
34.	For more information about the Medicaid program, see CRS Report R43357, Medicaid: An Overview, coordinated by [author name scrubbed].
35.	This enrollment figure is measured according to "average monthly enrollment," which differs from "ever enrolled" counts that measure the number of people covered by Medicaid for any period of time during the year. Congressional Budget Office, Detail of Spending and Enrollment for Medicaid – CBO's March 2015 Baseline, March 9, 2015.
36.	Centers for Medicare & Medicaid Services (CMS), CMS-64 data as of March 30, 2015.
37.	LTSS refer to a broad range of health and health-related services and supports needed by individuals who lack the capacity for self-care due to a physical, cognitive, or mental disability or condition. These services include nursing facility services, home health, case management services, personal care services, and private duty nursing. For more information about LTSS, see CRS Report R43328, Medicaid Coverage of Long-Term Services and Supports, by [author name scrubbed].
38.	Some groups, such as people under the age of 26 who have aged out of foster care, are eligible for Medicaid coverage without regard to income and assets.
39.	The ACA was supposed to provide health coverage for all low-income individuals by providing Medicaid coverage to the individuals with the lowest incomes (up to 133% of FPL) and providing premium tax credits for coverage through the health insurance exchanges to low-income individuals with incomes above Medicaid eligibility levels (i.e., 100% through 400% of FPL). For more information about the ACA Medicaid expansion, see CRS Report R43564, The ACA Medicaid Expansion, by [author name scrubbed]. For more information about the ACA health insurance premium credits, see CRS Report R43945, Health Insurance Premium Credits in the Patient Protection and Affordable Care Act (ACA) in 2015, by [author name scrubbed].
40.	Medicaid and CHIP Payment and Access Commission, Report to the Congress on Medicaid and CHIP, MACStats, Figure 7, June 2014.
41.	Ibid.
42.	Home- and community-based waivers provide a range of LTSS for targeted populations of enrollees who require an institutional level of care.
43.	Medicaid and CHIP Payment and Access Commission, Report to the Congress on Medicaid and CHIP, Figure 7, MACStats, June 2014.
44.	In FY2011, 1.5 million Medicaid enrollees received institutional services. Medicaid and CHIP Payment and Access Commission, Report to the Congress on Medicaid and CHIP, MACStats, Figure 5, June 2014.
45.	For more information about the ACA Medicaid expansion, see CRS Report R43564, The ACA Medicaid Expansion, by [author name scrubbed].
46.	Ben Sommers, Rick Kronick, and Kenneth Finegold, et al., Understanding Participation Rates in Medicaid: Implications for the Affordable Care Act, Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation, Issue Brief, March 16, 2012; Benjamin D. Sommers, Meredith Roberts Tomasi, and Katherine Swartz, et al., "Reasons for the Wide Variation in Medicaid Participation Rates Among States Hold Lessons for Coverage Expansions in 2014," Health Affairs, vol. 31, no. 5 (May 2012).
47.	Ben Sommers, Rick Kronick, and Kenneth Finegold, et al., Understanding Participation Rates in Medicaid: Implications for the Affordable Care Act, Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation, Issue Brief, March 16, 2012.
48.	The State Children's Health Insurance Program (CHIP) is a means-tested program that provides health coverage to targeted low-income children and pregnant women in families that have annual income above Medicaid eligibility levels but have no health insurance. For more information about CHIP, see CRS Report R43627, State Children's Health Insurance Program: An Overview, by [author name scrubbed] and [author name scrubbed].
49.	Genevieve M. Kenney, Jennifer M. Haley, and Nathaniel Anderson, et al., "Children Eligible for Medicaid or CHIP: Who Remains Uninsured, and Why?," Academic Pediatrics, vol. 15, no. 3S (May-June 2015).
50.	Ibid.
51.	Rick Kronick and Kenneth Finegold, et al., Understanding Participation Rates in Medicaid: Implications for the Affordable Care Act, Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation, Issue Brief, March 16, 2012; Genevieve M. Kenney, Jennifer M. Haley, and Nathaniel Anderson, et al., "Children Eligible for Medicaid or CHIP: Who Remains Uninsured, and Why?," Academic Pediatrics, vol. 15, no. 3S (May-June 2015).
52.	Marianne P. Bitler and Madeline Zavodny, Medicaid: A Review of the Literature, Nation al Bureau of Economic Research, NBER Working Paper Series, May 2014; Genevieve M. Kenney, Jennifer M. Haley, and Nathaniel Anderson, et al., "Children Eligible for Medicaid or CHIP: Who Remains Uninsured, and Why?," Academic Pediatrics, vol. 15, no. 3S (May-June 2015).
53.	Benjamin D. Sommers, Genevieve M. Kenney, and Arnold M. Epstein, "New Evidence On The Affordable Care Act: Coverage Impacts of Early Medicaid Expansions," Health Affairs, vol. 33, no. 1 (January 2014); Ben Sommers, Genevieve M. Kenney, Jennifer M. Haley, and Nathaniel Anderson, et al., "Children Eligible for Medicaid or CHIP: Who Remains Uninsured, and Why?," Academic Pediatrics, vol. 15, no. 3S (May-June 2015); and Benjamin D. Sommers, Meredith Roberts Tomasi, and Katherine Swartz, et al., "Reasons for the Wide Variation in Medicaid Participation Rates Among States Hold Lessons for Coverage Expansions in 2014," Health Affairs, vol. 31, no. 5 (May 2012).
54.	Genevieve M. Kenney, Jennifer M. Haley, and Nathaniel Anderson, et al., "Children Eligible for Medicaid or CHIP: Who Remains Uninsured, and Why?," Academic Pediatrics, vol. 15, no. 3S (May-June 2015).
55.	Benjamin D. Sommers, Meredith Roberts Tomasi, and Katherine Swartz, et al., "Reasons for the Wide Variation in Medicaid Participation Rates Among States Hold Lessons for Coverage Expansions in 2014," Health Affairs, vol. 31, no. 5 (May 2012).
56.	Constance F. Citro and Robert T. Michael, ed., Measuring Poverty. A New Approach, ed. (Washington, DC: National Academy Press, 1995), pp. 223-224.
57.	Congressional Budget Office, The Distribution of Household Income and Federal Taxes, 2011, November 2014.
58.	Amy Finkelstein, Nathaniel Hendren, and Erzo F.P. Luttmer, The Value of Medicaid: Interpreting Results from the Oregon Health Insurance Experiment, National Bureau of Economic Research, June 2015.
59.	Amy Finkelstein, Sarah Taubman, Bill Wright, Mira Bernstein, Jonathan Gruber, Joseph P. Newhouse, Heidi Allen, Katherine Baicker, and the Oregon Health Study Group, "The Oregon Health Insurance Experiment: Evidence from the First Year" Quarterly Journal of Economics, August 2012, 127(3).
60.	The analysis of Finkelstein, Hendren, and Luttmer cited above found Medicaid coverage benefits providers of the low-income uninsured at about $0.6 per every dollar spent on Medicaid because the uninsured pay only a fraction of medical expenditures.
61.	If a household did not have enough income to meet basic needs, it was assumed the household would not spend anything on health insurance. Congressional Budget Office, The Distribution of Household Income and Federal Taxes, 2008 and 2009, July 2012.
62.	States can require certain beneficiaries to share in the cost of Medicaid services, but there are limits on (1) the amounts that states can impose, (2) the beneficiary groups that can be required to pay, and (3) the services for which cost sharing can be charged.
63.	For example, see Congressional Budget Office, Trends in the Distribution of Household Income Between 1979 and 2007, October 2011.
64.	See Medical Care Economic Risk: Measuring Financial Vulnerability from Spending on Medical Care , ed. Michael J. O'Grady and Gooloo S. Wunderlich (National Academies Press, 2012).