Indicators of Welfare Dependence: Annual Report to Congress, 2008

U.S. Department of Health and Human Services

---

RESEARCH BRIEF

---

This report was written and compiled by Gil Crouse, Susan Hauan, and Annette Waters
Rogers of the Office of Human Services Policy under the direction of Melissa Pardue,
Deputy Assistant Secretary for Human Services Policy, Office of the Assistant Secretary
for Planning and Evaluation.

---

A limited number of copies of this year’s report also are available from:

Office of Human Services Policy
Office of the Assistant Secretary for Planning and Evaluation
U.S. Department of Health and Human Services
Hubert H. Humphrey Building, Room 404E
200 Independence Ave., S.W.
Washington, D.C. 20201
Fax: 202-690-6562

The Welfare Indicators Act of 1994 requires the Secretary of Health and Human Services to prepare an annual report to Congress on indicators welfare dependence. The Indicators of Welfare Dependence report is prepared within the Office of Human Services Policy and delivered to Congress each spring. As mandated under the Congressional act, the report addresses the rate of welfare dependency, the degree and duration of welfare recipiency and dependence, and predictors of welfare dependence. Further, analyses of means-tested assistance in the report include benefits under the Temporary Assistance for Needy Families (TANF) program; the Food Stamp Program, and the Supplemental Security Income (SSI) program. The report also includes risk factors related to economic security, employment, and nonmarital births, as well an appendix with data related to the above programs.

"

Executive Summary

Contents

• Economic Security Risk Factors
  • ECON 1.  Poverty Rates
  • ECON 2.  Deep Poverty Rates
  • ECON 3.  Experimental Poverty Measures
  • ECON 4.  Poverty Rates with Various Means-Tested Benefits Counted as Income
  • ECON 5.  Poverty Spells
  • ECON 6.  Child Support
  • ECON 7.  Food Insecurity
  • ECON 8.  Lack of Health Insurance
• Employment and Work-Related Risk Factors
  • WORK 1.  Labor Force Attachment
  • WORK 2.  Employment among the Low-Skilled
  • WORK 3.  Earnings of Low-Skilled Workers
  • WORK 4.  Educational Attainment
  • WORK 5.  High School Dropout Rates
  • WORK 6.  Adult Alcohol and Substance Abuse
  • WORK 7.  Adult and Child Disability
  • WORK 8.  Labor Force Participation of Women with Children under 18
• NonMarital Birth Risk Factors
  • BIRTH 1.  Nonmarital Births
  • BIRTH 2.  Nonmarital Teen Births
  • BIRTH 3.  Nonmarital Teen Births Rates
  • BIRTH 4.  Never-Married Family Status

The Welfare Indicators Act of 1994 requires the Department of Health and Human Services to prepare annual reports to Congress on indicators and predictors of welfare dependence. The 2008 Indicators of Welfare Dependence, the eleventh annual report, provides welfare dependence indicators through 2005, reflecting changes that have taken place since enactment of the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) in August 1996. As directed by the Welfare Indicators Act, the report focuses on benefits under the Temporary Assistance for Needy Families (TANF) program, formerly the Aid to Families with Dependent Children (AFDC) program; the Food Stamp Program; and the Supplemental Security Income (SSI) program.

Welfare dependence, like poverty, is a continuum, with variations in degree and in duration. Families may be more or less dependent if larger or smaller shares of their total resources are derived from welfare programs. The amount of time over which families depend on welfare might also be considered in assessing their degree of dependence. Although recognizing the difficulties inherent in defining and
measuring dependence, a bipartisan Advisory Board on Welfare Indicators proposed that: A family is dependent on welfare if more than 50 percent of its total income in a one-year period comes from TANF (formerly AFDC), food stamps and/or SSI, and this welfare income is not associated with work activities. Given data limitations, we follow the Board’s proposal by adopting the following definition of welfare dependence among individuals in families¹ for use in this report:

• Welfare dependence is the proportion of all individuals in families that receive more than half of their total family income in one year from TANF, food stamps and/or SSI.

This report uses data from the Current Population Survey (CPS) and administrative data for the TANF (formerly AFDC), Food Stamp and SSI programs to provide updated measures through 2005 for several dependence indicators. Other measures are based on the Survey of Income and Program Participation (SIPP), the Panel Study of Income Dynamics (PSID) and other data sources. Based on these data, this report provides a number of key indicators of welfare recipiency, dependence and labor force attachment. Highlights from the report include the following:

• In 2005, 3.8 percent of the total population was dependent in that they received more than half of their total family income from TANF, food stamps and/or SSI (see Indicator 1). While higher than the 3.7 percent dependency rate measured in 2004, the 2005 rate is lower than the 5.2 percent rate measured in 1996. Overall, 2.59 million fewer Americans were dependent on welfare in 2005 compared with 1996.
• Trends in dependency are similar to the more well-known changes in TANF and food stamp caseloads. For example, the percentage of individuals receiving TANF cash assistance (formerly AFDC) fell from 5.4 percent to 1.6 percent between 1993 and 2006 (see Indicator 3). Food stamp recipiency rates fell from 10.4 percent in 1993 to 6.1 percent in 2000 and 2001. Since then, the food stamp recipiency rate has increased to 8.9 percent in 2006. This increase in food stamp recipiency may explain the increase in overall dependency since 2000.
• In an average month in 2005, more than half (52.3 percent) of TANF recipients lived in families with at least one family member in the labor force. Comparable figures for food stamp and SSI recipients were 55.3 and 38.9 percent, respectively (see Indicator 2). Although there was a decline in labor force participation among TANF families from 2002 to 2004, full-time employment increased considerably among TANF families during much of the last decade.
• Spells of TANF receipt in the early 2000s were much shorter than spells of AFDC receipt in the early 1990s. Half (49.6 percent) of TANF spells for individuals entering the program between 2001 and 2003 lasted 4 months or less, compared to 30.4 percent of AFDC spells beginning between 1992 and 1994 (See Indicator 7).
• Longer-term welfare receipt was much less common during the late 1990s and early 2000s compared to earlier periods. Less than 3 (2.6) percent of those with some TANF (or AFDC) assistance between 1995 and 2004 received assistance in nine or ten years of the period, compared to 10.5 percent and 14.6 percent of AFDC recipients in the earlier two time periods (See Indicator 9).

Since the causes of welfare receipt and dependence are not clearly known, the report also includes a larger set of risk factors associated with welfare receipt. The risk factors are organized into three categories: economic security measures, measures related to employment and barriers to employment and measures of nonmarital childbearing. The economic security risk factors include measures of poverty and well-being that are important not only as potential predictors of dependence, but also as a supplement to the dependence indicators, ensuring that dependence measures are not assessed in isolation. As such, the report includes data on the official poverty rate, one of the most common measures of economic wellbeing:

• As the dependency rate decreased after 1993, the poverty rate for all individuals fell also, from 15.1 percent in 1993 to 11.3 percent in 2000. Between 2000 and 2004, the poverty rate increased, but still remained lower than any year between 1980 and 1997. Between 2005 and 2006, the poverty rate decreased from 12.6 percent to 12.3 percent of all individuals (see Economic Security Risk Factor 1).

Finally, the report has four appendices that provide additional data on major welfare programs, alternative measures of dependence and nonmarital births, as well as background information on several data and technical issues.

---

¹ Appendix D provides more information on the use of individuals, rather than families or households, as the unit of analysis for most of the statistics in this report.

Chapter I. Introduction and Overview

Contents

• Organization of Report
• Measuring Welfare Dependence
• Measuring Economic Well-Being
• Data Sources

The Welfare Indicators Act of 1994 (Public Law 103-432) directed the Secretary of Health and Human Services (HHS) to publish an annual report on welfare dependency. This 2008 report, the eleventh annual indicators report, gives updated data on the measures of welfare recipiency, dependency, and predictors of welfare dependence developed for previous reports. Much of this report reflects changes that have taken place since enactment of the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) in August 1996.

The purpose of this report is to address questions concerning the extent to which American families depend on income from welfare programs. Under the Welfare Indicators Act, HHS was directed to address the rate of welfare dependency, the degree and duration of welfare recipiency and dependence, and predictors of welfare dependence. The Act further specified that analyses of means-tested assistance should include benefits under the Temporary Assistance for Needy Families (TANF) program (formerly the Aid to Families with Dependent Children (AFDC) program), the Food Stamp Program, and the Supplemental Security Income (SSI) program. Throughout the report we use AFDC/TANF to refer to cash assistance benefits received under these two programs because the AFDC program preceded the TANF program.

This 2008 report provides updated measures through 2005 for dependency measures based on the Current Population Survey (CPS), Annual Social and Economic Supplement. Although more recent administrative data provide some information on recipiency through 2006, the survey data needed to examine overall welfare recipiency are not available past 2005 for the CPS-based measures, 2003 for the SIPP-based measures, and 2004 for the Panel Study of Income Dynamics measures. As in the 2007 report, measures updated annually are presented at the front of each chapter, followed by the figures that are derived from data sources that are updated less frequently.

 

Organization of Report

This introductory chapter provides an overview of the specific summary measure of welfare dependence proposed by a bipartisan Advisory Board¹ and how this measure was adopted for use in this annual report series. Also it discusses summary measures of poverty, following the Advisory Board’s recommendation that dependence measures not be assessed in isolation from other measures of economic well-being. The introduction concludes with a discussion of data sources used for the report.

Chapter II of the report, Indicators of Dependence, presents ten indicators of welfare dependence and recipiency. These indicators include dependence measures based on total income from all three programs – AFDC/TANF, SSI and food stamps – as well as measures of recipiency for each of the three programs considered separately. Labor force participation among families receiving welfare and benefit receipt across multiple programs also are shown. The second half of the chapter includes longitudinal data on transitions on and off welfare programs and spells of program recipiency, including spells of TANF receipt among persons in families that have no attachment to the labor market. Also, this section includes a measure of long-term program receipt of up to 10 years, and a measure of events associated with the beginning and ending of program spells.

Chapter III, Predictors and Risk Factors Associated with Welfare Receipt, focuses on predictors of welfare dependence – risk factors believed to be associated with welfare receipt. These predictors are shown in three different groups:

1. Economic security – including various measures of poverty, receipt of child support, food insecurity and health insurance coverage – is important in predicting dependence because families with fewer economic resources are more likely to rely on welfare programs for their support.
2. Measures of the work status and potential barriers to employment of adult family members also are critical, because families must generally receive an adequate income from employment in order to avoid dependence without severe deprivation.
3. Finally, data on nonmarital births are important since a high proportion of long-term welfare recipients first became parents outside of marriage, frequently as teenagers.

Additional data and technical notes are presented in four appendices. Appendix A provides basic program data on each of the main welfare programs and their recipients; Appendix B shows how dependence is affected by the inclusion of benefits from the SSI program; Appendix C includes additional data on nonmarital childbearing; and Appendix D provides background information on several data and technical issues. The main welfare programs in Appendix A include the following.

• The Temporary Assistance for Needy Families (TANF) program provides monthly cash benefits to eligible families with children and is run directly by the states. Prior to 1996 this program was known as the Aid to Families with Dependent Children (AFDC) program. Data on the AFDC and TANF programs are provided in Appendix A, with AFDC data provided from 1962 through June 1997, and TANF data from July 1997 through 2006.
• The Food Stamp Program provides monthly food stamp benefits to individuals living in families or alone, provided their income and assets are below limits set in federal law. It reaches more poor people over the course of a year than any other means-tested public assistance program. Appendix A provides historical data from 1962 to 2006.
• The Supplemental Security Income (SSI) program provides monthly cash payments to elderly, blind or disabled individuals or couples whose income and assets are below levels set in federal law. Though the majority of recipients are adults, disabled children also are eligible. Historical data from 1974 through 2006 are provided in Appendix A.

Measuring Welfare Dependence

As suggested by its title, this report focuses on welfare “dependence” as well as welfare “recipiency.” While recipiency can be defined fairly easily, based on the presence of benefits from AFDC/TANF, SSI or food stamps, dependence is a more complex concept.

Welfare dependence, like poverty, is a continuum, with variations in degree and in duration. Families may be more or less dependent if larger or smaller shares of their total resources are derived from welfare programs. The amount of time over which a family depends on welfare might also be considered in assessing its degree of dependence. Nevertheless, a summary measure of dependence to be used as an indicator for policy purposes must have some fixed parameters that allow one to determine which families should be counted as dependent, just as the poverty line defines who is poor under the official standard. The definition of dependence proposed by the Advisory Board for this purpose is as follows: A family is dependent on welfare if more than 50 percent of its total income in a one-year period comes from AFDC/TANF, food stamps and/or SSI, and this welfare income is not associated with work activities. In following the Board’s proposal, we adopt the following definition of welfare dependence among individuals in families² for use in this report:

• Welfare dependence is the proportion of all individuals in families that receive more than half of their total family income in one year from TANF, food stamps, and/or SSI.

Any definition of welfare dependence is not without its limitations. The Advisory Board recognized that no single measure could capture fully all aspects of dependence and that their proposed measure should be examined in concert with other indicators of well-being. While the Board’s proposal would count unsubsidized and subsidized employment and work required to obtain benefits as work activities, existing data sources do not permit distinguishing between welfare income associated with work activities and nonwork-related welfare benefits. As a result, the data shown in this report may overstate the incidence of dependence because welfare income associated with work required to obtain benefits is classified as welfare and not as income from work. This issue may be growing in importance under the increased work requirements of the TANF program. In FY 2006, 30.1 percent of welfare recipients were working (including employment, work experience and community service), compared to 7 percent in 1992.³

Any definition also represents an arbitrary choice of a percentage of income from welfare beyond which families will be considered dependent. But using a single point – in this case 50 percent – yields a relatively straightforward measure that can be tracked easily over time, and is likely to be associated with any large changes in total dependence, however defined. For example, dependence under the definition used in this report declined as policy changes under welfare reform moved more recipients into employment.

As shown in Figure SUM 1, 3.8 percent of the population would be considered “dependent” on welfare in 2005 in that they received more than half of their family income in 2005 from TANF, food stamps and/or SSI. This is one-quarter of the percentage that lived in a family receiving at least some TANF, food stamps, or SSI benefits during the year.

---

Figure SUM 1.
Recipiency and Dependency Rates: 1993-2005

Note:  Recipiency is defined as living in a family with receipt of any amount of AFDC/TANF, SSI or food stamps during the year.  Dependency is defined as having more than 50 percent of annual income from AFDC/TANF, SSI and/or food stamps. Dependency rates would be lower if adjusted to exclude welfare assistance associated with working.

Source:  Unpublished tabulations from the Current Population Survey, Annual Social and Economic Supplement, 1994-2006, analyzed using the TRIM3 microsimulation model.

---

Dependency and recipiency rates follow fairly similar trends, falling fairly dramatically during the 1990s to lows of 3.0 percent for dependency and 12.5 percent for recipiency in 2000.  While rates have increased somewhat between 2000 and 2005, the 2005 dependency and recipiency rates remain substantially lower than the peak rates of 5.9 and 17.2 percent, occurring in 1993 and 1994, respectively.  The overall drop in recipiency rates since the early 1990s is consistent with TANF administrative data showing declining TANF caseloads, especially after enactment of welfare reform in 1996.  What is not apparent from administrative records, but is shown in national survey data, is that dependency also declined after 1993, with the sharpest decline occurring after enactment of the 1996 welfare reform legislation.  While 13.74 million individuals were dependent in 1996, only 11.15 million were dependent in 2005 — representing a decline of 2.59 million people.

Table SUM 1.
Recipiency and Dependency Rates: Selected Years

Recipiency and dependency rates are higher for Non-Hispanic Blacks and Hispanics than for Non-Hispanic Whites, as shown in Table SUM 1.  Recipiency and dependence also are higher for young children than for adults, and for individuals in female-headed families than for those in married-couple families.  However, both recipiency and dependency rates are much lower for Non-Hispanic Blacks, Hispanics, children and individuals in female-headed families in 2005 compared to 1993.

Measures of welfare dependency also vary based upon which programs are counted as “welfare programs.”  Dependency would be much lower — 2.1 percent — if only AFDC/TANF and food stamp benefits were counted (as shown in Appendix B and as is done in some measures in this report).  Moreover, the drop in dependency is even larger under this alternative definition of dependence than usually reported.  For example, between 1995 and 2005, dependency declined from 3.6 percent to 2.1 percent under the alternative definition.

Another factor affecting dependence is the time period observed.  The summary measures shown in Figure and Table SUM 1 focus on recipiency and dependency rates measured on an annual, cross-sectional basis.  Longitudinal measures of program receipt (both annual and monthly) show that program spells are typically short and long-term recipiency is more rare (see Chapter II).  Indicator 9, for example, shows that among individuals receiving AFDC/TANF at some point over a ten-year period ending in 2004, 10.0 percent received some welfare during six or more years. Another quarter (24.7 percent) were recipients in three to five years, and nearly two-thirds (65.3 percent) received welfare in only one or two years during this period.

Measuring Economic Well-Being

To assess the social impacts of any change in dependence, changes in the level of poverty should be considered.  This chapter focuses on the official poverty rate, the most common poverty measure. Additional measures of poverty and need also are included under the Economic Risk Factors found in Chapter III.

The poverty rate in 2006 remains much lower than in 1993, when poverty reached its highest peak since the early 1980s. The official poverty rate for 2006 was 12.3 percent, compared to 15.1 percent in 1993. This difference in the poverty rate indicates that 2.8 million fewer people are in poverty and 2.9 million fewer children are in poverty in 2006 than in 1993.

---

Figure SUM 2.
Percentage of Total Population in Poverty with Various Means-Tested Transfers Counted as Income: 1979-2006

Note: The three measures of income are as follows: (1) “Before means-tested cash transfers” is earnings and other pre-transfer (“private” or “market”) cash income, plus social security, workers compensation, and other social insurance cash transfers.  It does not include means-tested cash transfers; (2) The “Official poverty measure” uses the official Census Bureau income definition, which includes means-tested cash transfers, primarily AFDC/TANF and SSI; (3) “After means-tested non-cash benefits and taxes” counts the cash value of means-tested food and housing benefits, adds the refundable Earned Income Tax Credit (EITC), and subtracts federal payroll and income taxes. The fungible value of Medicare and Medicaid is not included in any of the income measures.

Source: Unpublished tabulations from the Current Population Survey, Annual Social and Economic Supplement, 1980 – 2007, analyzed by the Congressional Budget Office. See ECON 4 in Chapter III for underlying table and further notes.

---

Figure SUM 2 shows poverty estimates under the official poverty rate and two other measures that adjust income by adding or subtracting means-tested cash transfers, means-tested non-cash benefits, and federal taxes.  While each of the three poverty measures in the graph uses a different definition of income, all three poverty measures use the Census Bureau’s official poverty thresholds.

The “Official poverty measure” trend line shows the official poverty rate based on total cash income, including means-tested cash transfers. The official poverty rate was 12.3 percent in 2006.

The “Before means-tested cash transfers” trend line shows what the poverty rate would be if means-tested cash transfers (primarily AFDC/TANF and SSI) were excluded from income.  Income in this measure includes earnings and other pre-transfer cash income, plus social security, workers compensation, and other social insurance cash transfers.  The poverty rate under this measure would be higher than under the official measure, or 13.0 percent in 2006.

The “After means-tested non-cash benefits and taxes” trend line shows that the poverty rate would be lower if the cash value of means-tested food and housing transfers and the effect of federal taxes were counted as income.⁴  Under this definition, the poverty rate in 2006 would be more than two percentage points lower than the official measure, or 10.0 percent.

Data Sources

The primary data sources for this report are the Current Population Survey (CPS), the Survey of Income and Program Participation (SIPP), the Panel Study of Income Dynamics (PSID) and administrative data for the AFDC/TANF, Food Stamp and SSI programs.  Beginning with the 2001 report, there was a shift to using CPS rather than SIPP data for several indicators and predictors of welfare recipiency and dependence.  This change was necessary because CPS data are updated annually, while SIPP updates are available much less frequently.

If it were not for the lags in data availability, the SIPP would be considered the most useful national survey for measuring welfare dependency.  It was used most extensively in the first three annual dependence reports.  Its longitudinal design, system of monthly accounting and detail concerning employment, income and participation in federal income-support and related programs, make the SIPP particularly effective for capturing the complexities of program dynamics.  It continues to be an important source of data in this report, particularly for measures related to AFDC/TANF and poverty spell duration, transitions in and out of program dependency and reasons for entering or leaving the AFDC/TANF program.

For measures of receipt, dependency and poverty at a single point in time, the report primarily uses the Annual Social and Economic Supplement to the CPS, which measures income and poverty over an annual accounting period.  As stated above, the CPS data are available on a timelier basis than the SIPP, and have been widely used to measure trends since the welfare reform legislation of 1996.  However, because the CPS does not collect income in the same detail as the SIPP, it has been subject to criticism for underreporting of income, particularly welfare income.  To address this concern, some of the indicators in this report are based on CPS data that have been analyzed by the Transfer Income Model (TRIM3), a microsimulation model developed by the Urban Institute under contract to the Office of the Assistant Secretary for Planning and Evaluation.  Although its primary purpose is to simulate program eligibility and the impact of policy proposals, the TRIM model also has been used to correct for underreporting of welfare receipt and benefits.  Welfare caseloads in TRIM3 are based on CPS data, adjusted upward to ensure that total estimates of recipients equal the total counts from administrative data.  To maintain consistency in data trends, we present estimates based on CPS data analyzed by TRIM3 beginning in 1993, the first year the TRIM3 microsimulation model became available.

As shown in Figure SUM 3, the overall measures of dependency and recipiency have not been greatly affected by the change in data sources. Both data sources show a decline in dependence between 1996 and 1999 and a small increase in dependence during the early 2000s.  Still, readers are cautioned against comparing measures for 1987-1995 from the SIPP data in the first three annual reports with the measures for 1993-2005 from the TRIM-adjusted CPS data.

---

Figure SUM 3.
Recipiency and Dependency Rates from Two Data Sources: 1987 – 2005.

Note:  Recipiency is defined as receipt of any amount of AFDC/TANF, SSI or food stamps during the year. Dependency is defined as having more than 50 percent of annual family income from AFDC/TANF, SSI and/or food stamps. Dependency rates would be lower if adjusted to exclude welfare assistance associated with working. While only affecting a small number of cases, General Assistance income is included within AFDC/TANF income and veterans pension benefits are included in means-tested assistance income for SIPP-based receipt and dependency estimates prior to 2001.

Source:  Unpublished tabulations from the Current Population Survey, Annual Social and Economic Supplement, 1994-2006, analyzed using the TRIM3 microsimulation model, and unpublished tabulations from the Survey of Income and Program Participation, 1987, 1990, 1992, 1993, 1996, 2001, and 2004 panels.

---

The Panel Study of Income Dynamics (PSID) is another source of data used in this report.  Like the SIPP it provides longitudinal data, but over a much longer time period than the three- to four-year time period of the SIPP.  With annual data on program receipt since 1968, the PSID provides vital data for measuring longer-term welfare use over periods of up to 10 years.  Because the PSID indicators cover time spans as long as a decade, they are updated less frequently than the CPS-based and SIPP-based measures.  In this 2008 annual report, the key PSID indicator for long-term welfare receipt, Indicator 9, has been updated.  Indicator 9 still includes estimates of welfare receipt over 10-year periods, but the years covered by each time period have shifted.  Thus, readers are cautioned against comparisons with estimates from prior reports.

The report also draws upon administrative data for the AFDC/TANF, Food Stamp and SSI programs.  These data are largely reported in Appendix A.  Like the CPS data, administrative data are generally available with little time lags; these data are generally available through fiscal year (FY) 2006.  To the extent possible, TANF administrative data are reported in a consistent manner with data from the earlier AFDC program, as noted in the footnotes to the tables in Appendix A.  The fact remains that assistance under locally designed TANF programs encompasses a diverse set of cash and non-cash benefits designed to support families in making a transition to work, and so direct comparisons between AFDC receipt and TANF receipt should be made with caution.  This issue also affects reported data on AFDC and TANF receipt in national data sets such as the CPS and SIPP.

For further technical information about the data presented in the report, specifically for information on race and ethnicity, unit of analysis and annual versus monthly measures, please see Appendix D.

Endnotes

¹ The first annual report was produced under the oversight of a bipartisan Advisory Board on Welfare Indicators, which assisted the Secretary in defining welfare dependence, developing indicators of welfare dependence, and choosing appropriate data.  Under the terms of the original authorizing legislation, the Advisory Board was terminated in October 1997, prior to the submission of the first annual report.

² Appendix D provides more information on the use of individuals, rather than families or households, as the unit of analysis for most of the statistics in this report.

³ This 30.1 percent includes 21.7 percent in unsubsidized employment and 8.5 percent in work preparation activities (including subsidized jobs, on-the-job training, work experience or community services). The earnings of those in unsubsidized employment would be correctly captured as income from work in national surveys.  Any welfare benefits associated with work experience, community service programs or other work activities, however, would be counted as income from welfare in most national surveys, a classification incompatible with the Advisory Board’s proposed definition.

⁴ The effects of food and housing benefits are shown separately from the effect of federal taxes in Figure ECON 4 in Chapter III.  Prior to 1993, including the effect of federal taxes increased poverty. Since 1993, federal taxes and tax credits (including refunds through the Earned Income Tax Credit) have had the net effect of reducing poverty rates.

Chapter II. Indicators of Dependence

Contents

• Indicator Summary
• Indicator 1: Degree of Dependence
• Indicator 2: Receipt of Means-Tested Assistance and Labor Force Attachment
• Indicator 3: Rates of Receipt of Means-Tested Assistance
• Indicator 4: Rates of Participation in Means-Tested Assistance Programs
• Indicator 5: Multiple Program Receipt
• Indicator 6: Dependence Transitions
• Indicator 7: Program Spell Duration
• Indicator 8: Welfare Spell Duration with No Labor Force Attachment
• Indicator 9: Long Term Receipt
• Indicator 10: Events Associated with the Beginning and Ending of Program Spells

Following the format of the previous annual reports to Congress, Chapter II presents summary data related to indicators of dependence.  These indicators differ from other welfare statistics because of their emphasis on welfare dependence, rather than simply welfare receipt.

As discussed in Chapter I, the Advisory Board on Welfare Indicators suggested that families be considered dependent if more than 50 percent of their total income in a one-year period comes from cash assistance through the Temporary Assistance for Needy Families (TANF) program (formerly the Aid to Families with Dependent Children (AFDC) program), food stamps and Supplemental Security Income (SSI) benefits. Furthermore, this welfare income was not to be associated with work activities.  Existing data from administrative records and national surveys, however, do not generally distinguish welfare benefits received in conjunction with work from benefits received without work.  Thus, it was not possible to construct one single indicator of dependence that captured fully the Advisory Board’s recommendation; that is, one indicator based on the percentage of income from means-tested assistance only if this income is not associated with work activities. As discussed in Chapter I, we adopt the following definition of welfare dependence among individuals in families¹ for use in this report:

• Welfare dependence is the proportion of all individuals in families that receive more than half of their total family income in one year from TANF, food stamps and/or SSI.

The ten indicators in Chapter II were selected to provide information about the range and depth of dependence as proposed by the Advisory Board, including indicators that measure the presence of employment activities.  This chapter focuses on recipients of three major means-tested cash and nutritional assistance programs: cash assistance through the AFDC and TANF programs, benefits under the Food Stamp Program, and SSI benefits for elderly and disabled recipients. For some indicators, summary data and characteristics are provided for all recipients, not just those defined as welfare-dependent. While a number of indicators focus on the percentage of recipients’ income from means-tested assistance, other indicators focus on presence of work activities at the same time as welfare receipt.

---

¹ Appendix D provides more information on the use of individuals, rather than families or households, as the unit of analysis for most of the statistics in this report.

Indicator Summary

Indicator 1: Degree of Dependence.  This indicator focuses most closely on those individuals who meet the Advisory Board’s proposed definition of “dependence.”  In addition to examining individuals with more than 50 percent of their annual family income from AFDC/TANF cash assistance, food stamps and/or SSI benefits, it shows various levels of dependence by examining those with more than 0 percent, 25 percent and 75 percent of their family income from these sources (Indicators 1a and 1b).  This indicator also shows the average percentage of income from means-tested assistance and earnings received by families with various levels of income relative to the poverty level (Indicators 1c and 1d).

Indicator 2: Receipt of Means-Tested Assistance and Labor Force Attachment.  This indicator looks further at the relationship between receipt of means-tested assistance and participation in the labor force. This is an important issue because of the significant number of low-income individuals that use a combination of means-tested assistance and earnings from the labor force.

Indicator 3: Rates of Receipt of Means-Tested Assistance.  This indicator paints yet another picture of dependence by measuring recipiency rates, that is, the percentage of the population that receives AFDC/TANF, food stamps or SSI in an average month.  Administrative data for the AFDC/TANF, Food Stamp and SSI programs make these figures readily available over time, allowing a better sense of historical trends than is available from the more specialized indicators of dependence.

Indicator 4: Rates of Participation in Means-Tested Assistance Programs.  While means-tested public assistance programs are open to all that meet their requirements, not all eligible individuals and households participate in the programs.  This indicator uses AFDC/TANF, Food Stamp and SSI administrative data and microsimulation models to reflect “take-up rates” — the number of families that actually participate in the programs as a percentage of those who are estimated to be legally eligible.

Indicator 5: Multiple Program Receipt.  Depending on their circumstances, individuals may choose a variety of different means-tested assistance “packages.”  This indicator looks at the percentage of individuals receiving AFDC/TANF, food stamps and SSI in a month, examining how many rely on just one of these programs, and how many rely on a combination of two programs.

Indicator 6: Dependence Transitions.  This indicator uses data from the Survey of Income and Program Participation (SIPP) to look at whether individuals dependent on welfare in one year make the transition out of dependence in the following year.

Indicator 7: Program Spell Duration.  One critical aspect of dependence is how long individuals receive means-tested assistance. This indicator provides information on short, medium and long spells of welfare receipt for each of the three major means-tested programs — AFDC/TANF, the Food Stamp Program, and SSI.

Indicator 8: Welfare Spell Duration with No Labor Force Attachment.  This indicator is concerned with dynamics of welfare receipt among persons in families with no attachment to the labor market.  It differs from Indicator 7 in that it provides information on spells of TANF receipt during months where no one in the family worked or was officially unemployed.

Indicator 9: Long Term Receipt.  Many individuals who leave welfare programs cycle back on after an absence of several months.  Thus it is important to look beyond individual program spells, measured in Indicator 7, to examine the cumulative amount of time individuals receive assistance over a period of several years.

Indicator 10: Events Associated with the Beginning and Ending of Program Spells.  To gain a better understanding of welfare dynamics, it is important to go beyond measures of spell duration and examine information regarding the major events in people’s lives that are correlated with the beginnings or endings of program spells. This measure focuses on receipt of TANF.

INDICATOR 1. Degree of Dependence

Figure IND 1a.
Percentage of Total Income from Means-Tested Assistance Programs: 2005

Note:  Means-tested assistance includes TANF, SSI, and food stamps. Total >50% includes all persons with more than 50 percent of their total annual family income from these means-tested programs. Income includes cash income and the value of food stamps.

Source:  Unpublished tabulations from the Current Population Survey, Annual Social and Economic Supplement, 2006, analyzed using the TRIM3 microsimulation model.

---

• Figure IND 1a shows the percentage of persons in families with varying degrees of dependence on means-tested assistance programs in 2005.
• The majority of persons (84.7 percent) lived in families that received no income from means-tested assistance programs in 2005.
• Fifteen (15.3) percent of persons lived in families that receive income supports from means-tested assistance programs.  Almost 4 (3.8) percent of persons lived in families that received more than half of their income from means-tested assistance programs.  These persons would be considered welfare dependent under the definition of dependence used in this report.¹
• Table IND 1a shows the percentage of persons in families with varying degrees of reliance on income from means-tested assistance programs by demographic characteristics.  Welfare dependence varies across demographic groups. Among racial and ethnic groups, Non-Hispanic Blacks were more likely to be welfare dependent (10.2 percent) than were Non-Hispanic Whites (2.2 percent) or Hispanics of any race (5.7 percent).
• Among age categories, children from birth to 5 years of age were more likely to live in families that were welfare dependent than were children of other age categories.
• Among family types, persons living in female-headed families were more likely to be welfare dependent than those in other family categories.
• Table IND 1b shows trends in welfare dependence between 1993 and 2005.  Welfare dependence was highest in 1993 at 5.9 percent.  Welfare dependence declined between 1993 and 2000, with notable drops occurring between 1996 and 2000.  After 2000, the downward trend in welfare dependence reversed, with dependence increasing from 3.0 percent in 2000 to 3.8 percent in 2005.

Table IND 1a.
Percentage of Total Income from Means-Tested Assistance Programs by Selected Characteristics: 2005

Table IND 1b.
Percentage of Total Income from Means-Tested Assistance Programs:  1993-2005

---

Figure IND 1b.
Percentage of Total Income from Various Sources by Poverty Status: 2005

Note:  Total income is total annual family income, including the value of food stamps. Other income is non-means-tested, non-earnings income such as child support, alimony, pensions, Social Security benefits, interest and dividends. Poverty status categories are not mutually exclusive.

Source:  Unpublished tabulations from the Current Population Survey, Annual Social and Economic Supplement, 2006, analyzed using the TRIM3 microsimulation model.

---

• Figure IND 1b shows sources of income by poverty status in 2005. There is an association between poverty status and receiving income from means-tested assistance programs.
• Persons in families with incomes below the poverty line received 46.6 percent of their income from earnings and 32.5 percent from means-tested assistance programs.  Persons in families with incomes at 200 percent or more of the poverty line received 86.6 percent of their income from earnings and .2 percent of their income from means-tested assistance programs.
• The percentage of family income that comes from earnings is inversely proportional to overall family income relative to the poverty line.  For example, the percentage of income received from earnings for persons in families living in deep poverty (below 50 percent of the poverty line) was 25.3 percent compared to 46.6 percent for all poor persons in 2005.
• Table IND 1c shows sources of income by poverty status for various demographic groups.  On average, persons in married-couple families rely on earnings more and on means-tested assistance programs less than persons in other family categories at all income levels.
• Table IND 1d shows the percentage of income from various sources across selected years.  The percentage of income received from earnings for persons in families with incomes below the poverty line increased from 40.4 percent in 1995 to a high of 49.5 percent in 2000.  After 2000, the rate decreased to 46.6 percent in 2005.
• Over the same time period, the percentage of income from means-tested programs among persons in poor families decreased substantially from 41.3 percent in 1995 to 30.3 percent in 2000.  After 2000, the rate increased to 32.5 percent in 2005

Table IND 1c.
Percentage of Total Income from Various Sources by Poverty Status and Selected Characteristics: 2005

Table IND 1d.
Percentage of Total Income from Various Sources: Selected Years

---

¹ For a discussion on defining welfare dependence, please see “Measuring Welfare Dependence” in Chapter I.

INDICATOR 2. Receipt of Means-Tested Assistance and Labor Force Attachment

Figure IND 2.
Percentage of Recipients in Families with Labor Force Participants by Program: 2005

Note:  Recipients are limited to those individuals or family members directly receiving benefits in a month. Full-time workers are those who usually work 35 hours or more per week. Part-time labor force participation includes part-time workers and those who are unemployed, laid off and/or looking for work. This indicator measures, on an average monthly basis, the combination of individual benefit receipt and labor force participation by any family member in the same month.

Source:  Unpublished tabulations from the Current Population Survey, Annual Social and Economic Supplement, 2006, analyzed using the TRIM3 microsimulation model.

• Figure IND 2 shows the percentage of recipients in families with labor force participants by program.¹  In 2005, SSI recipients were more likely to live in families with no labor force participants (61.1 percent) than were TANF recipients (47.7 percent) or food stamp recipients (44.6 percent).
• Table IND 2a shows the percentage of recipients in families with labor force participants by program and demographic characteristics.
• Among TANF recipients, Hispanics of any race were more likely to live in families with at least one full-time worker (34.3 percent) than were Non-Hispanic Whites (25.9 percent) or Non-Hispanic Blacks (22.5 percent).
• Among TANF recipients, 49.2 percent of persons in married-couple families lived with at least one full-time worker compared to 19.6 percent of persons in female-headed families, and 28.3 percent of persons in male- headed families.
• Table IND 2b shows the percentage of AFDC/TANF recipients living in families with labor force participants by year.  The percentage of recipients living in families with full-time workers increased from 18.8 percent in 1993 to 35.3 percent in 2001 and then declined to 26.9 percent in 2005.

Table IND 2a.
Percentage of Recipients in Families with Labor Force Participants by Program and Selected Characteristics: 2005

Table IND 2b.
Percentage of AFDC/TANF Recipients in Families with Labor Force Participants: 1993-2005

---

¹ Note that lower family employment rates are reported in TANF administrative data, which are limited to the employment of family members in the TANF assistance unit and employment reported to welfare agencies (see Table TANF 7 in Appendix A).

INDICATOR 3. Rates of Receipt of Means-Tested Assistance

Figure IND 3a.
Percentage of the Total Population Receiving AFDC/TANF by Age: 1970-2006

Note:  See Appendix A, Tables TANF 2, TANF 12 and TANF 14, for more detailed data on recipiency rates, including recipiency rates by calendar year. Recipients are expressed as the fiscal year average of monthly caseloads from administrative data, excluding recipients in the territories. Tribal TANF recipients are also excluded.  Child recipients include a small number of dependents ages 18 and older who are students. The average number of adult and child recipients in 1998 and 1999 are estimated using data from the National Emergency TANF Data Files and thereafter using the National TANF Data Files. Beginning in 2000, the data include both TANF and SSP recipients who have comprised as much as 11 percent of total recipients.

Source:  U.S. Department of Health and Human Services, Administration for Children and Families, Office of Family Assistance. Population denominators for the percents in each category are from the U.S. Census Bureau (available online at http://www.census.gov).

---

• Figure IND 3a shows the percentage of the population who received income from the AFDC program or TANF program by age group from 1970 to 2006.
• Table IND 3a shows the number and percent of the population receiving AFDC/TANF by age between 1970 and 2006.  In 1993, 5.4 percent of the population received income from AFDC.  In 2006, the percentage was 1.6.  The 2006 rate of TANF receipt was the lowest since 1970.
• AFDC/TANF recipiency rates have been higher with more pronounced changes over time for children than for adults.  Between 1993 and 2006, AFDC/TANF receipt among children decreased from 13.9 percent to 4.8 percent, the most rapid decline in the time period shown.

Table IND 3a.
Number and Percentage of the Total Population Receiving AFDC/TANF by Age:
1970-2006

Figure IND 3b.
Percentage of the Total Population Receiving Food Stamps by Age: 1975-2006

Note:  See Appendix A, Tables FSP 1 and FSP 6 for more detailed data on recipiency rates. Recipient totals exclude the territories and are the fiscal year averages of monthly caseloads from administrative data. From 1975 to 1983 the number of participants includes the Family Food Assistance Program (FFAP) that was largely replaced by the Food Stamp Program in 1975. From 1975 to 1983 the number of FFAP participants averaged only 88 thousand.

Source:  Recipient data by age from U.S. Department of Agriculture, Food and Nutrition Service, Office of Analysis, Nutrition and Evaluation, Characteristics of Food Stamp Households, Fiscal Year 2006 and earlier reports (available online at http://www.fns.usda.gov/pd/snapmain.htm), and unpublished data from the Food Stamps National Data Bank. Population denominators for the percents in each category are from U.S. Census Bureau (available online at http://www.census.gov).

---

• Figure IND 3b shows the percentage of the population who received food stamps by age category from 1975 to 2006.
• The food stamp recipiency rate increased to 8.9 percent in 2006 from a low of 6.1 percent in 2000 and 2001, the lowest rate since the Food Stamp Program became available nationwide. While the 2006 recipiency rate is higher than the 2005 rate, it is still lower than the peak rate of 10.4 percent experienced in 1993 and 1994.
• As with AFDC/TANF, food stamp recipiency rates have been higher over time for children than for adults.  Between 1980 and 2006, the percentage of all children who received food stamps was at least double the percentage for all adults ages 18 to 59.
• Table IND 3b shows the number and percentage of the population receiving food stamps by age from 1975 to 2006.  Trends in food stamp recipiency across all age groups are similar over the time period.  The trends may largely reflect changes in the rate of unemployment and programmatic changes.
• The percentage of all persons receiving food stamps declined between 1984 and 1988 and then increased in the early 1990s reaching a peak rate in 1993 (10.4 percent). The percentage then declined through 2000 and since then has risen to 8.9 percent in 2006.

Table IND 3b.
Number and Percentage of the Total Population Receiving Food Stamps
by Age: 1975-2006

Figure IND 3c.
Percentage of the Total Population Receiving SSI by Age: 1975-2006

Note:  December population figures used as the denominators are obtained by averaging the U.S. Census Bureau's July 1 population estimates for the current and the following year. See Appendix A, Tables SSI 2, SSI 8 and SSI 9 for more detailed data on SSI recipiency rates.

Source:  Social Security Administration, Office of Research, Evaluation and Statistics, SSI Annual Statistical Report, 2006, (available online at http://www.ssa.gov/policy). Population denominators for the percents in each category are from the U.S. Census Bureau (available online at http://www.census.gov).

---

• Figure IND 3c shows the percentage of the population who received income assistance from the SSI program by age category from 1975 through 2006.
• Unlike the recipiency rates for AFDC/TANF and food stamps, overall recipiency rates for SSI show less variation over time.  After decreasing from 1975 to the early 1980s, the proportion of the population that received SSI increased from 1.7 percent in 1985 to 2.5 percent in 1996.  The percentage then declined to 2.4 percent in 2006.  The total number of recipients has increased by 72 percent over the same period, from 4.1 million in 1985 to 7.2 million people in 2006.
• Table IND 3c shows the percentage of the population and number of persons receiving SSI by age between 1975 and 2006.
• Elderly adults (ages 65 and older) have higher recipiency rates than any other age group.  The gap, however, has narrowed as the percentage of adults aged 65 and older receiving SSI has been cut in half, declining from 10.9 percent in 1975 to 5.3 percent in 2006.
• The proportion of children receiving SSI increased gradually between 1975 and 1990, and grew more rapidly in the early and mid-1990s, reaching 1.4 percent in 1996.  The rate then fell through 2000 before rising to 1.5 percent in 2006.

Table IND 3c.
Number and Percentage of the Total Population Receiving SSI
by Age: 1975-2006

INDICATOR 4. Rates of Participation in Means-Tested Assistance Programs

Figure IND 4.
Participation Rates in the AFDC/TANF¹, Food Stamp and SSI Programs: Selected Years

Note:  AFDC/TANF and SSI participation rates are estimated by an Urban Institute model (TRIM3) that uses CPS data to simulate program eligibility and participation for an average month, by calendar year.  There have been small changes in estimating methodology over time, due to model improvements and revisions to the CPS.  Most notably, since 1994 the model has been revised to more accurately estimate SSI participation among children, and in 1997 and 1998 the model was adjusted to more accurately exclude ineligible immigrants.  For TANF, in contrast to editions prior to 2004, this table includes families receiving assistance under Separate State Programs (SSPs).  Note that families subject to full-family sanctions are counted as nonparticipating eligible families due to modeling limitations.  Although the coverage rate estimates take into account the number of families who lost aid due to the time limit (and do not count such families in the denominator of the coverage rate estimate), they do not make any allowance for families staying off of TANF to conserve their time-limited assistance months.  Also, the numbers of eligible and participating families include the territories and pregnant women without children, even though these two small groups are excluded from the TRIM model.  The numbers shown here implicitly assume that participation rates for the territories and for pregnant women with no other children are the same as for all other eligibles.  In 2004 the methods for identifying potential child-only units capture the fact that non-parent caretakers generally have a choice of whether or not to be included in the TANF unit.  TRIM now excludes those caretakers whose income would make the unit ineligible, increasing the number of potential child-only units.

Food Stamp Eligible households are estimated from a Mathematica Policy Research, Inc. model that uses CPS data to simulate the Food Stamp Program.  Food Stamp caseload data are from USDA, FNS program operations caseload data.  There have been small changes in the methodology over time, due to model improvements and revisions to the CPS.  Notably, the model was revised in 1994 to produce more accurate and lower estimates of eligible households. The estimates for previous years show higher estimates of eligibles and lower participation rates relative to the revised estimate for 1994 and estimates for subsequent years. The two estimates for 1999 are due to re-weighting of the March 2000 – 2003 CPS files to Census 2000 and revised methodologies for determining food stamp eligibility.  The original estimate (September 1999) is consistent methodologically with estimates from September 1994 – September 1998, while the revised estimate (FY 1999) is consistent with the estimates for FY 2000 – FY 2005.

Source:  U.S. Department of Agriculture, Food and Nutrition Service, Trends in Food Stamp Program Participation Rates: 1999-2005 (available online at http://www.fns.usda.gov/pd/snapmain.htm), and unpublished tabulations from the TRIM3 microsimulation model.

---

• Figure IND 4 shows the participation rates of means-tested assistance programs for selected years.  This indicator examines participating families or households as a percentage of the estimated eligible population.  It is a contrast to Indicator 3, which examines participants as a percentage of the total population (recipiency rates).
• Forty (40.4) percent of families estimated as eligible for TANF cash assistance, 59.1 percent of households estimated as eligible for food stamps, and 67.7 percent of adults estimated as eligible for SSI are estimated to have enrolled and received benefits in an average month in 2005.

Table IND 4a.
Number and Percentage of Eligible Families Participating in the AFDC/TANF Cash Assistance Program: Selected Years

• Table IND 4a shows the number and percentage of eligible families participating in the cash AFDC/TANF program in selected years.
• Between 1981 and 1996, participation rates in the AFDC program ranged from 76.7 percent (in 1987) to 85.7 percent (in 1992).
• After 1996, participation rates in the cash TANF program decreased from 78.9 percent of families estimated to be eligible for AFDC/TANF cash benefits in 1996 to 40.4 percent of families estimated to be eligible for TANF cash benefits in 2005.
• Note that TANF is a flexible program with a flexible funding stream.  As such, states provide substantial “non assistance” services and benefits that would not be included in these cash assistance estimates.
• Families also may receive cash benefits or other services through general assistance and other solely state-funded programs² that are separate from the TANF program and are not shown here.

Table IND 4b.
Number and Percentage of Eligible Households Participating in the Food Stamp Program: Selected Years

• Table IND 4b shows the number and percentage of eligible households participating in the Food Stamp Program for selected years.  Between fiscal years 2004 and 2005, the participation rate for food stamps increased from 55.5 percent in 2004 to 59.1 percent in 2005.
• Between fiscal years 1999 and 2005 there was a 24.8 percent increase in households eligible for the Food Stamp Program (from 14.5 to 18.1 million households).  Caseloads grew by 42.6 percent over the same period, with notable increases occurring in both 2004 and 2005.  Subsequently, the estimated participation rate increased from 51.4 percent in 1999 to 59.1 percent in 2005.
• While there were 10.7 million households participating in the Food Stamp Program in 2005, the caseload is still lower than the 1993 peak (10.9 million households).  During the mid to late 1990s, there was a 33 percent drop in food stamp caseloads, from a peak of 10.9 million households in 1993 to 7.3 million households in 1999.  This decline in caseloads occurred during a time when both the eligible population and the program participation rates were generally decreasing.

Table IND 4c.
Percentage of Eligible Adult Units Participating in the SSI Program by Selected Characteristics: 1993-2005

• Table IND 4c shows the percentage of eligible adult units participating in the SSI program by select demographic categories. After rising to 75.8 percent of adults estimated to be eligible for SSI in 2000, the SSI participation rate decreased to 67.7 percent of those estimated to be eligible for SSI in 2005.  This rate remains higher than recent TANF and food stamp participation rates (see Tables IND 4a and IND 4b).
• Between 2004 and 2005, for aged adults in one-person units, the estimated SSI participation rate increased from 57.0 percent in 1993 to a high of 70.9 percent in 2000.  After some declines in the early 2000s, the estimated SSI participation rate among aged one-person units increased from 61.9 percent in 2002 to 63.4 percent in 2005.

---

¹ Unlike the Food Stamp and SSI programs, TANF is a block grant program for which there is no individual entitlement.  One of the main goals of TANF is to move people from cash assistance to self-sufficiency, which may be inconsistent with achieving a higher coverage rate.

² As discussed in the note to Table IND 4a above, the model for estimating participation in the TANF cash assistance program does take into account benefits from Separate State Programs (SSPs) that are used to meet Maintenance of Effort (MOE) requirements.

INDICATOR 5. Multiple Program Receipt

Figure IND 5.
Percentage of Recipients Receiving Assistance from Multiple Programs — TANF, Food Stamps and SSI: 2005

Note:  Categories are mutually exclusive.  SSI receipt is based on individual receipt; AFDC/TANF and food stamp receipt are based on the full recipient unit. In practice, individuals do not tend to receive both AFDC/TANF and SSI; hence, no individual receives benefits from all three programs. The percentage of individuals receiving assistance from any one program in an average month (shown here) is lower than the percentage residing in families receiving assistance at some point over the course of a year (shown in Table SUM 1 in Chapter I and Table IND 1a in Chapter II). Spouses are not present in the female-headed and male-headed family categories

Source:  Unpublished tabulations from the Current Population Survey, Annual Social and Economic Supplement, 2006, analyzed using the TRIM3 microsimulation model.

---

• Figure IND 5 shows the percentage of those receiving benefits from TANF, food stamps, or SSI or a combination of benefits from these programs in 2005.  About three-quarters (74.3 percent) of persons in families receiving TANF, food stamps, or SSI benefits in an average month received assistance from only one program. Most received food stamps or SSI benefits only.  Two percent of persons in families received only TANF benefits.
• Table IND 5a shows the percentage of the population receiving assistance from multiple means-tested assistance programs by demographic characteristics.  Among age categories, children were more likely than persons in other age categories to live in families that received support from multiple means-tested assistance programs.  For example, 5.3 percent of children from birth to 5 years lived in families that received both TANF and food stamps as compared with 1.2 percent of women aged 16 to 64.
• Among family categories, persons in female-headed families were more likely than those living in other types of families to receive support from multiple means-tested assistance programs.  Among persons in female-headed families, 7.4 percent received support from TANF and food stamps, as compared to .5 percent of those in married-couple families, and 1.9 percent of those in male-headed families.
• Table IND 5b shows the percentage of the population receiving assistance from multiple means-tested assistance programs between 1993 and 2005.  Reliance on multiple means-tested programs has decreased over time.  In 1993, 4.8 percent of the population received AFDC and food stamps.  In 2005, the percent who received both TANF and food stamps decreased to 1.5 percent.

Table IND 5a.
Percentage of Population Receiving Assistance from Multiple Means-Tested Assistance Programs
by Selected Characteristics: 2005

Table IND 5b.
Percentage of Population Receiving Assistance from Multiple Means-Tested Assistance Programs: 1993-2005

INDICATOR 6. Dependence Transitions

Figure IND 6.
Dependency Status in 2003 of Persons Who Received More than 50 Percent of Income from Means-Tested Assistance in 2002
by Race and Ethnicity

Note:  Means-tested assistance is defined as AFDC/TANF, food stamps and SSI.  While only affecting a small number of cases, General Assistance income is included within AFDC/TANF income.  Individuals are defined as dependent if they reside in families with more than 50 percent of total annual family income from these means-tested programs.

Persons of Hispanic ethnicity may be of any race. Due to small sample size, American Indians/Alaska Natives, Asians and Native Hawaiians/Other Pacific Islanders are included in the total for all persons but are not shown separately.

Source:  Unpublished tabulations from the Survey of Income and Program Participation, 2001 panel.

---

• Figure IND 6 shows the 2003 dependency status of persons who were welfare dependent in 2002 by race and ethnicity.  Welfare dependence is defined as receiving more than half of one’s total family income in the year from TANF, food stamps, and/or SSI.  For further discussion of defining welfare dependency, see Chapter I.
• Of the recipients who received more than 50 percent of their total family income from AFDC/TANF, food stamps and/or SSI in 2002, 70.8 percent of Non-Hispanic Whites, 72.6 percent of Non-Hispanic Blacks, and 69.1 percent of Hispanics were welfare dependent in 2003.
• Table IND 6a shows the 2003 dependency status of persons who were welfare dependent in 2002 by demographic groups.  Men ages 16 to 64 who received more than half of their total income from means-tested assistance programs in 2002 remained dependent in 2003 in higher percentages than women.
• Table IND 6b shows the dependency status of all persons who received more than 50 percent of their income from means-tested assistance programs in the previous year.  Recipients of means-tested assistance programs were more likely to move out of welfare dependency in the early 2000s than in the early 1990s.

Table IND 6a.
Dependency Status in 2003 of Persons Who Received More than 50 Percent of Income
from Means-Tested Assistance in 2002 by Selected Characteristics

Table IND 6b.
Dependency Status of All Persons Who Received More than 50 Percent of Income
from Means-Tested Assistance in Previous Year

INDICATOR 7. Program Spell Duration

Figure IND 7.
Percentage of TANF, Food Stamp and SSI Spells for Persons Entering Programs during the 2001-2003 Period
by Length of Spell

Note:  Spell length categories are mutually exclusive.  Spells separated by only 1 month are not considered separate spells.  Due to the length of the observation period, actual spell lengths for spells that lasted more than 20 months cannot be observed.  Program spells are defined as those starting during the 2001 SIPP panel.  For certain age categories, data are not available (NA) because of insufficient sample size.

Source:  Unpublished tabulations from the Survey of Income and Program Participation, 2001 panel.

---

• Figure IND 7 shows the percentage of TANF, food stamp, and SSI spells by spell length categories for persons entering programs in the early 2000s.  Between 2001 and 2003, spells lasting four months or less accounted for 49.6 percent of TANF spells, 35.9 percent of food stamp spells, and 27.9 percent of SSI spells.
• Approximately three-fourths of all TANF spells (73.3 percent) and three-fifths of food stamp spells (60.3 percent) lasted one year or less compared to 49.3 percent of SSI spells.
• Table IND 7a shows the percentage of program spells for persons entering programs during the 2001 – 2003 period by length of spell and demographic characteristics.  For TANF spells, a smaller percentage of Non-Hispanic White recipients (11.9 percent) had long spells lasting more than 20 months compared to Non-Hispanic Blacks (19.1 percent) and Hispanics (19.8 percent).
• Table IND 7b shows how the percentage of program spells of varying lengths for persons entering programs during selected periods has changed.  Spells of welfare receipt were shorter in the early 2000s than in the early 1990s.  For instance, 16.8 percent of TANF
• spells for persons entering TANF between 2001 and 2003 lasted 20 months or longer as compared to 34.4 percent of AFDC spells beginning between 1992 and 1994.

Table IND 7a.
Percentage of TANF, Food Stamp and SSI Spells for Persons Entering Programs during the 2001-2003 Period
by Length of Spell and Selected Characteristics

Table IND 7b.
Percentage of AFDC/TANF, Food Stamp, and SSI Spells for Persons Entering Programs
during Selected Periods by Length of Spell

INDICATOR 8. Welfare Spell Duration with No Labor Force Attachment

Figure IND 8.
Percentage of TANF Spells with No Family Labor Force Attachment for Persons Entering Programs during the 2001 – 2003 Period
by Length of Spell

Note:  Spell length categories are mutually exclusive.  Spells separated by only 1 month are not considered separate spells.  Due to the length of the observation period, actual spell lengths for spells that lasted more than 20 months cannot be observed.  TANF spells with no family labor force attachment are defined as those spells starting during the 2001 SIPP panel for persons who received TANF and lived in families with no labor force participants in each month.

Source:  Unpublished tabulations from the Survey of Income and Program Participation, 2001 panel.

---

• Figure IND 8 shows the percentage of TANF spells with no family labor force attachment for persons entering the TANF program between 2001 and 2003 by length of spell.
• Welfare spells with no family labor force attachment are measured as consecutive months that a person received TANF benefits and lived in a family with no labor force participants.  Welfare spells with no family labor force attachment may end when a person leaves the TANF program or when a person remains on TANF but at least one person in the family enters the labor market.
• Fifty-six (56.1) percent of welfare spells with no family labor force attachment lasted less than four months as measured in the SIPP.
• Table IND 8a shows the percentage of TANF spells with no family labor force attachment by spell length for different demographic groups. The percentage of spells ending in four months or less was larger for Non-Hispanic Whites (61.2 percent) than it was for Non-Hispanic Blacks (52.8 percent) and Hispanics (59.9 percent).

Table IND 8a.
Percentage of TANF Spells with No Family Labor Force Attachment for Persons Entering Programs
during the 2001 – 2003 Period by Length of Spell and Selected Characteristics

Table IND 8b.
Percentage of TANF Spells with No Family Labor Force Attachment
for Persons Entering Programs during Selected Years

• Table IND 8b shows the percentage of TANF spells with no family labor force attachment for persons entering the program during selected periods by spell length.  In the early 2000s, 56.1 percent of TANF spells with no family labor force attachment ended within four months and 79.1 percent ended within a year.
• The percentage of spells with no family labor force attachment lasting more than 20 months was higher in the early 1990s than in the early 2000s (22.5 percent compared to 10.2 percent, respectively).
• Indicators 7 and 8 provide similar information; however, the percentages of spell lengths differ because the two Indicators are computed differently.  Indicator 7 shows spells for all recipients while Indicator 8 restricts welfare spells to recipients in families without any labor force participants.  This difference results in a higher percentage of spells longer than 20 months in Indicator 7, where TANF and employment may be combined and compared to Indicator 8 where no one in the family may be in the labor force.

INDICATOR 9. Long Term Receipt

Figure IND 9.
Percentage of AFDC/TANF Recipients
by Years of Receipt during the 1995 – 2004 Period

Note:  The base for the percentages consists of mothers who received at least $1 of AFDC/TANF in any year in the ten-year period. Child recipients are defined by age in the first year of the 10-year period. This indicator measures years of recipiency over the specified ten-year time periods and does not take into account years of recipiency that may have occurred before or after each ten-year period.

Source:  Unpublished tabulations from the Panel Study of Income Dynamics, public release data files, 1996-2005.

---

• Figure IND 9 shows the percentage of AFDC/TANF recipients by years of receipt between 1995 and 2004.  Among all persons receiving AFDC/TANF at some point within the ten-year period, 65.3 percent received assistance in only one or two of these years. In contrast, 2.6 percent received assistance in 9 or 10 of the years.
• Table IND 9 shows the percentage of AFDC/TANF recipients with varying years of receipt across three ten-year time periods by demographic characteristics.  Long spells of welfare receipt were more common in earlier time periods than they were in later time periods. For example, for the 1975 - 1984 time period, 14.6 percent of AFDC recipients received benefits in at least 9 of the 10 years as compared to 2.6 percent of AFDC/TANF recipients for the 1995 – 2004 time period.
• Short spells of TANF were more prevalent in the 1995 to 2004 period compared to earlier periods.  Between 1995 and 2004, 65.3 percent of TANF recipients received benefits in only one or two years compared to 44.5 percent in the 1985 to 1994 period and 46.2 percent in the 1975 to 1984 period.
• Among child recipients, for the 1975 – 1984 time period, 36.3 percent of children birth to age 5 lived in families that received AFDC/TANF in only 1 or 2 of the years as compared to 66.3 percent for the 1995 – 2004 time period.
• Among racial groups, the percentage of Non-Hispanic Black recipients receiving TANF benefits in only one or two years during a ten-year period increased by 25.6 percentage points between the 1975 to 1984 period and the 1995 to 2004 period.  In comparison, this same percentage for Non-Hispanic White recipients increased by 15.8 percentage points across the same two time periods.

Table IND 9.
Percentage of AFDC/TANF Recipients across Three Ten-Year Time Periods
by Years of Receipt and Selected Characteristics

INDICATOR 10. Events Associated with the Beginning and Ending of Program Spells

Figure IND 10a.
Events Associated with Single Mother TANF Exits during the 2001-2003 Period

Note:  Welfare exits are defined as moving from receipt to non-receipt between two successive SIPP interviews (conducted 4 months apart); an event was associated with a welfare transition if the event was observed within two interviews (i.e., 8 months) of the interview marking the welfare exit.  In general, events are neither mutually exclusive nor exhaustive, and transition events may sum to more than 100 percent.  Two exceptions are that “Increase in other Household Earnings” was limited to cases when there were increases in household earnings without an increase in recipient earnings, and “Increase in Adults (not marriage)” was limited to cases where the adult joining the household was not marrying the head of the household.  While only affecting a small number of cases, General Assistance income is included within AFDC/TANF income. Other government benefits include Unemployment Insurance, Foster Care, Railroad Retirement, veterans payments and Workers Compensation.  An increase in earnings must be an increase of at least $50 per month.  A work limitation is defined as a condition that limits the kind or amount of work.  The category "None of above in recent past" represents the percentage of all spell beginnings during the period that were not associated with any of the events measured.

Spells of welfare receipt and associated events are measured using monthly data from the SIPP.  In the 2003 Indicators of Welfare Dependence volume (and earlier volumes), events associated with the beginning and ending of program spells were measured using annual data from the Panel Study of Income Dynamics (PSID).  Thus, the estimates shown above are not comparable to estimates reported in volumes prior to 2004.

Events sum to more than 100 percent because the same household could experience more than one event associated with a specific welfare entry or exit.

Source:  Unpublished tabulations from the Survey of Income and Program Participation, 2001 panel.

---

• Figure IND 10a shows events associated with single mother TANF exits during the 2001 – 2003 time period.  Welfare exits were most often associated with an increase in recipient earnings.  Thirty-four (34.1) percent of welfare spells that ended during the 2001 to 2003 time period were associated with an increase in the recipient’s earnings.  Twelve (12.1) percent of welfare exits were associated with an increase in the earnings of other household members.
• Thirty-seven percent of welfare exits during the 2001 – 2003 time period were not associated with any of the events listed above within the time period observed.
• Table IND 10a shows the events associated with welfare exits among single mother recipients for selected years.  Exits associated with an increase in recipient earnings have decreased over time.  For the 1993 – 1995 time period, 54.8 percent of exits were associated with an increase in recipient earnings yet for the 2001 - 2003 time period, 34.1 percent were associated with increases in recipient earnings¹.

Table IND 10a.
Percentage of Single Mother AFDC/TANF Spell Exits Associated with Specific Events:
Selected Periods

• Welfare exits associated with changes in household composition have also decreased over time.  For the 1993 – 1995 time period, 5.6 percent of welfare exits were related to the last child in a household leaving home or turning 19 years old as compared to 1.5 percent for the 2001 – 2003 time period.  Welfare exits associated with marriage also declined over the two time periods.  For the 1993 – 1995 time period, 5.4 percent of exits were related to marriage, for the 2001 – 2003 time period, the rate was 2.2 percent.
• Thirty-seven (37.4) percent of welfare exits were not associated with any of the events listed above within the time period observed.

---

Figure IND 10b.
Events Associated with Single Mother TANF Entries during the 2001-2003 Period

Note:  Welfare entries are defined as moving from non-receipt to receipt between two successive SIPP interviews (conducted 4 months apart); an event was associated with a welfare transition if the event was observed within two interviews (i.e., 8 months) of the interview marking the welfare entry.  In general, events are neither mutually exclusive nor exhaustive, and transition events may sum to more than 100 percent.  Two exceptions are that “Other Household Earnings Decreased” was limited to cases when there were decreases in household earnings without a decrease in recipient earnings, and “Decrease in Number of Adults (not divorce)” was limited to cases where the adult leaving the household was not married to the head of the household.  While only affecting a small number of cases, General Assistance income is included within AFDC/TANF income. Other government benefits include Unemployment Insurance, Foster Care, Railroad Retirement, veterans payments and Workers Compensation.  A decrease in earnings must be a decrease of at least $50 per month.  A work limitation is defined as a condition that limits the kind or amount of work.  The category "None of above in Recent Past" represents the percentage of all spell beginnings during the period that were not associated with any of the events measured.

Spells of welfare receipt and associated events are measured using monthly data from the SIPP.  In the 2003 Indicators of Welfare Dependence volume (and earlier volumes), events associated with the beginning and ending of program spells were measured using annual data from the Panel Study of Income Dynamics (PSID).  Thus, the estimates shown above are not comparable to estimates reported in volumes prior to 2004.

Events sum to more than 100 percent because the same household could experience more than one event associated with a specific welfare entry or exit.

Source:  Unpublished tabulations from the Survey of Income and Program Participation, 2001 panel.

---

• Figure IND 10b shows the events associated with the beginning of TANF spells among single mother recipients in the 2001 – 2003 time period. A decrease in earnings was the most common event associated with welfare entries.  For spells beginning between 2001 and 2003, 50.3 percent were associated with a decrease in the recipient’s earnings and 20.0 percent were associated with a decrease in the earnings of other household members.
• Changes in household composition also were associated with the beginning of welfare spells.  Twenty (20.2) percent of welfare entries were associated with a new child joining the family while 4.2 percent of welfare entries were associated with divorce or separation.
• Seventeen (16.9) percent of welfare entries were not associated with any of the events listed above within the time period observed.

Table IND 10b.
Percentage of Single Mother AFDC/TANF Spell Entries Associated with Specific Events:
Selected Periods

• Table IND 10b shows the events associated with the beginning of welfare spells among single mother recipients by selected time periods.
• For the 1993-1995 time period, 57.1 percent of AFDC spell entries were associated with a decrease in recipient earnings.  The percentage was 50.3 percent for the 2001-2003 time period.
• A decrease in other household members’ earnings also was related to the beginning of welfare spells.  For the 1993-1995 time period, 24.0 percent of welfare entries were associated with a decrease in other household members’ earnings.  For the 2001-2003 time period, 19.8 percent of welfare entries were associated with a decrease in other household members’ earnings.

---

¹ The percentage point difference between the two time periods in exits associated with increases in earnings may be related to the larger share of the welfare caseload combining welfare and work.  Some recipients with welfare exits in more recent years may have experienced increases in earnings before the 5- to 8-month time period used to observe “associated” events.

Chapter III. Predictors and Risk Factors Associated with Welfare Receipt

The Welfare Indicators Act challenges the U.S. Department of Health and Human Services to identify and set forth not only indicators of welfare dependence and welfare duration but also predictors and causes of welfare receipt. However, welfare research has not established clear and definitive causes of welfare receipt and dependence.  Instead, it has identified a number of risk factors associated with welfare use. For the purposes of this report, the terms “predictors” and “risk factors” are used somewhat interchangeably.

Following the recommendation of the Advisory Board, this chapter includes a wide range of possible predictors and risk factors.  As research advances, some of the “predictors” included in this chapter may turn out to be simply correlates of welfare receipt, some may have a causal relationship, some may be consequences, and some may have predictive value.

The predictors/risk factors included in this chapter are grouped into three categories:  economic security risk factors, employment-related risk factors, and risk factors associated with nonmarital childbearing.

Economic Security Risk Factors (ECON)

The first group includes eight measures associated with economic security.  This group encompasses five measures of poverty, as well as measures of child support receipt, food insecurity, and lack of health insurance.  The tables and figures illustrating measures of economic security are labeled with the prefix ECON throughout this chapter.

Poverty measures are important predictors of dependence, because families with fewer economic resources are more likely to be dependent on means-tested assistance.  In addition, poverty and other measures of deprivation, such as food insecurity, are important to assess in conjunction with the measures of dependence outlined in Chapter II.

Reductions in caseloads and dependence can reduce poverty, to the extent that such reductions are associated with greater work activity and higher economic resources for former welfare families.  However, if former welfare families are left with fewer economic resources, reductions in welfare caseloads may not lead to decreases in poverty.

Several aspects of poverty are examined in this chapter.  Those that can be updated annually using the Current Population Survey include: overall poverty rates (ECON 1); the percentage of individuals in deep poverty (ECON 2), and poverty rates using alternative definitions of income (ECON 3 and 4).  The chapter also includes data on the length of poverty episodes or spells (ECON 5).

This chapter also includes data on child support collections (ECON 6), which can play an important role in reducing dependence on government assistance and thus serve as a predictor of dependence.  Household food insecurity (ECON 7) is an important measure of deprivation that, although correlated with general income poverty, provides an alternative measure of tracking the incidence of material hardship and need, and how it may change over time.  Finally, lack of health insurance (ECON 8) is tied to the income level of the family, and may be a precursor to future health problems among adults and children.

Economic Security Risk Factor 1. Poverty Rates

Figure ECON 1.
Percentage of Persons in Poverty by Age: 1959-2006

Note:  Last data point is 2006.  All persons under 18 include related children (own children, including stepchildren and adopted children, plus all other children in the household who are related to the householder by birth, marriage, or adoption), unrelated individuals under 18 (persons who are not living with any relatives), and householders or spouses under age 18.

Source:  U.S. Census Bureau, “Income, Poverty, and Health Insurance Coverage in the United States: 2006,” Current Population Reports, SeriesP60-233, and data published online at http://www.census.gov/hhes/www/poverty.html.

---

• Figure ECON 1 shows the percentage of persons in poverty by age from 1959 to 2006.  The official poverty rate was 12.3 percent in 2006.  The percentage of persons living in poverty in 2006 was lower than poverty rates during all of the 1980s and most of the 1990s.
• Children under 18 had a poverty rate of 17.4 percent in 2006.  As in past years, the child poverty rate is higher than the overall poverty rate.
• Table ECON 1 shows the percentage of persons in poverty by age and family type for selected years.
• The poverty rate for the elderly (persons ages 65 and over) was 9.4 percent and the poverty rate for other adults (persons ages 18 to 64) was 10.8 percent in 2006.
• Related children from birth to age five have had the highest poverty rate among all age groups throughout the 1980s, 1990s, and into the 2000s.  In 2006, 20.0 percent of related children from birth to age 5 lived below the poverty line.
• The poverty rates for persons in both married-couple families and female-headed families have decreased over time.  In 1959, 18.2 percent of persons in married-couple families and 49.4 percent of persons in female-headed families were poor.  By 2006, 5.7 percent of persons in married-couple families and 30.5 percent of persons in female-headed families were poor.

Table ECON 1.
Percentage of Persons in Poverty by Age and Family Type:  Selected Years

Economic Security Risk Factor 2. Deep Poverty Rates

Figure ECON 2.  Percentage of Total Population below 50, 100, and 125 Percent of Poverty Level

Source:  U.S. Census Bureau, “Income, Poverty, and Health Insurance Coverage in the United States:  2006,” Current Population Reports, Series P60-233, and data published online at http://www.census.gov/hhes/www/poverty.html.

---

• Figure ECON 2 shows the percentage of the population below 50, 100, and 125 percent of the poverty level over time.  The percentage of the population in “deep poverty” (with incomes below 50 percent of the federal poverty level) was 5.2 percent in 2006, compared to an overall poverty rate of 12.3 percent.
• Five (4.5) percent of the population was “near-poor;” they had incomes at or above 100 percent but below 125 percent of the federal poverty level in 2006.
• Table ECON 2 shows the number and percentage of the population below 50, 75, and 125 percent of the poverty level for selected years.  In general, the percentage of the population with incomes below 50 percent of the poverty level has followed a pattern that reflects the trend in the overall poverty rate.
• The percentage of people below 50 percent of the poverty level rose in the late 1970s and early 1980s to 5.9 percent, and then after falling, rose to a second peak of 6.2 percent in 1993.  The rates for 100 percent and 125 percent of the poverty level followed a somewhat similar pattern with more pronounced peaks and valleys.
• Over the past two decades, the proportion of the poverty population in “deep poverty” has increased.  From a low of 28 percent of the poverty population in 1976, this population rose to just over 42 percent in 2006.
• The total number of poor people in 2006 was 36.5 million.  This number was 2.8 million lower than the recent peak of 39.3 million in 1993.

Table ECON 2.
Number and Percentage of Total Population below 50, 75, 100, and 125 Percent of Poverty Level:  Selected Years

Economic Security Risk Factor 3. Experimental Poverty Measures

Figure ECON 3.
Percentage of Persons in Poverty Using Various Experimental Poverty Measures by Age:  2006

Note:  These measures use versions of 1999 CE-based poverty thresholds that are adjusted for inflation using the CPI-U.

These experimental poverty measures implement changes recommended by a 1995 NAS panel, including: counting certain non-cash income as benefits; subtracting from income certain work-related, health and child care expenses; introducing new poverty thresholds; and adjusting those thresholds for geographic differences in housing costs.  The three alternative measures are similar, except that each accounts for medical out-of-pocket expenses (MOOP) differently.  The first alternative (MOOP subtracted from income or MSI) subtracts out-of-pocket medical expenses from income.  The second alternative (MOOP in the threshold or MIT) increases the poverty thresholds to take MOOP expenses into account.  The third measure, CMB for combined methods, combines attributes of the previous two measures.  Each of the three measures is calculated with and without accounting for geographic adjustments (GA and NGA).

Source:  U.S. Census Bureau, “Alternative Poverty Estimates Based on National Academy of Sciences Recommendations, by Geographic and Inflationary Adjustments,” available online at http://www.census.gov/hhes/www/povmeas/altmeas06/nas_measures_2005_2006_comparison.xls, and unpublished CPS data from the U.S. Census Bureau.

---

• Figure ECON 3 shows the percentage of persons in poverty using various experimental poverty measures by age in 2006.  Three experimental measures of poverty (developed by the U.S. Census Bureau in response to the recommendation of a 1995 panel of the National Academy of Sciences) yield poverty rates that are similar to the official poverty measure overall, but differ by age and other characteristics.
• Experimental measures generally show lower poverty rates among children than the official measure, partly because they take into account non-cash benefits that many children receive. Conversely, experimental measures show higher rates of poverty among the elderly than the official measure, in part due to taking into account certain out-of-pocket health costs for these measures.
• All three alternative measures shown in Figure ECON 3 are versions that do not take into account geographic adjustments for housing costs (NGA); there also are versions that do take into account those geographic adjustments (GA), as shown in Tables ECON 3a and 3b.

Table ECON 3a.
Percentage of Persons in Poverty Using Various Experimental Poverty Measures by Selected Characteristics:  2006

Table ECON 3b.
Percentage of Persons in Poverty Using Various Experimental Poverty Measures: 1999-2006

Economic Security Risk Factor 4. Poverty Rates with Various Means-tested Transfers Counted as Income

Figure ECON 4.
Percentage of Total Population in Poverty with Various Means-Tested Transfers Counted as Income: 1979-2006

Note:  The four measures of income are as follows: (1) “Pre-transfer cash income plus social insurance cash transfers” is earnings and other pre-transfer (“private” or “market”) cash income, plus social security, workers compensation, and other social insurance cash transfers.  It does not include means-tested cash transfers; (2) “Plus means-tested cash transfers” is the official Census Bureau income definition, which includes means-tested cash transfers, primarily AFDC/TANF and SSI; (3) “Plus food and housing benefits” counts the cash value of means-tested food and housing benefits as income; and (4) “Plus EITC and federal taxes” is the most comprehensive income measure used.  It adds the refundable Earned Income Tax Credit (EITC) to income, while subtracting federal payroll and income taxes.  The fungible value of Medicare and Medicaid is not included in any of the income measures.

Source:  Unpublished tabulations from the Current Population Survey, Annual Social and Economic Supplement, 1980-2007, analyzed by the Congressional Budget Office.

---

• Figure ECON 4 shows the percentage of the population in poverty with various means-tested transfers counted as income for the years 1979 to 2006.  The official poverty rate – using the official income definition, which includes means-tested cash transfers (primarily TANF and SSI) in addition to pre-transfer cash income and social insurance cash transfers – was 12.3 percent in 2006.  Without cash welfare, the 2006 poverty rate would be 13.0 percent.
• Adding non-cash, means-tested transfers to the official income definition has the effect of lowering the percentage of people with incomes below the official poverty line.  Including the value of food and housing benefits in total income would reduce the poverty rate to 11.0 percent in 2006.
• When income is defined to include the Earned Income Tax Credit (EITC) and the effect of federal taxes, the percentage of people in poverty would decrease to 10.0 percent in 2006.  Federal taxes and the EITC have had the net effect of reducing poverty rates following the EITC expansions in 1993 and 1995.
• Table ECON 4 shows the percentage of the population in poverty with various means-tested transfers counted as income for selected years.  The combined effect of means-tested cash transfers, food and housing benefits, the EITC, and federal taxes was to reduce the poverty rate in 2006 by 3 percentage points.  Net reductions in poverty rates were smaller during the 1981 to 1982 recession, and higher in the mid-1990s, largely due to expansions in the EITC.

Table ECON 4.
Percentage of Total Population in Poverty with Various Means-Tested Transfers Counted as Income:  Selected Years

Economic Security Risk Factor 5. Poverty Spells

Figure ECON 5.
Percentage of Poverty Spells for Persons Entering Poverty during the 2001 – 2003 Period by Length of Spell

Note:  Spell length categories are mutually exclusive.  Spells separated by only 1 month are not considered separate spells.  Due to the length of the observation period, actual spell lengths for spells that lasted more than 20 months cannot be observed.

Source:  Unpublished tabulations from the Survey of Income and Program Participation, 2001 panel.

---

• Figure ECON 5 shows the percentage of poverty spells that are of various lengths for persons who became poor during the 2001 to 2003 period.  Nearly half (49.2 percent) of poverty spells that began between 2001 and 2003 ended within 4 months.  More than three-quarters (76.9 percent) of poverty spells during this period ended within one year while 15.5 percent of spells lasted more than 20 months.
• Table ECON 5a shows the percentage of poverty spells for persons entering poverty during the 2001 to 2003 period by length of spell and demographic characteristics.
• Among racial and ethnic groups, a larger percentage of Non-Hispanic Whites had short spells of poverty (52.3 percent) than Non-Hispanic Blacks (42.1 percent) or Hispanics of any race (45.7 percent).  For poverty spells greater than 20 months, a larger percentage of Non-Hispanic Blacks had longer poverty spells (21.1 percent) compared to Non-Hispanic Whites (13.5 percent) and Hispanics of any race (16.8 percent).
• Among age categories, the difference in the percentage of poverty spells among adults 65 years or older and other adults is notable.  Twenty-one (21.2) percent of adults ages 65 years and over had poverty spells that lasted more than 20 months as compared to 14.4 percent of women ages 16 to 64 and 12.1 percent of men ages 16 to 64.

Table ECON 5a.
Percentage of Poverty Spells for Persons Entering Poverty during the 2001-2003 Period by Length of Spell and Selected Characteristics

Table ECON 5b.
Percentage of Poverty Spells for Persons Entering Poverty during Selected Time Periods by Length of Spell

Economic Security Risk Factor 6. Child Support

Figure ECON 6.
Percentage of Families Receiving Child Support Collections by Receipt of IV-D Services and Other Public Assistance: 1993-2005

Note:  AFDC/TANF families are families who have reported receiving cash assistance for any month during the 12-month period.  Therefore, not all the child support reported received was necessarily received while the family was receiving cash assistance. Data limitations do not allow a month-by-month breakdown.  Families receiving SSI, food stamps, Medicaid or housing assistance are limited to families not receiving AFDC/TANF.  Families receiving services through the IV-D system are estimated according to the methodology described in technical appendices to the ASPE-published report Characteristics of Families Using Title IV-D Services in 1999 and 2001, available at: http://aspe.hhs.gov/hsp/CSE-Char04/index.htm and previous reports.

Source:  Unpublished tabulations from the Current Population Survey, Child Support Supplement, 1994-2006.

---

• Figure ECON 6 shows the percentage of all families that receive child support collections by receipt of title IV-D services and other public assistance between 1993 and 2005. Title IV-D of the Social Security Act authorizes state programs to assist custodial parents in establishing paternity and child support awards, and collecting child support payments. The total amount of child support received by custodial parents through the IV-D system in 2005 was $17.2 billion (constant 2005 dollars) or 65.9 percent of all child support payments received by custodial parents.
• In total for 2005, custodial parents reported receiving $26.1 billion in child support payments from non-resident parents.¹ Total child support collections have increased by 19.2 percent since 1993, after adjusting for inflation. 
• Table ECON 6 shows greater detail on child support collections by receipt of IV-D services and other assistance.  Child support payments received through IV-D by custodial parents who also received AFDC/TANF cash assistance, declined from $3.3 billion (constant 2005 dollars) in 1993 to $1.8 billion in 2005.²
• Child support payments to custodial parents who did not receive TANF but received another form of public assistance (food stamps, SSI, Medicaid or housing assistance) increased from $2.2 billion (in constant 2005 dollars) to $5.9 billion between 1993 and 2005.  This group of custodial parents includes former TANF recipients as well as those eligible for cash assistance.  The increased collections for this group offset the decline in payments to TANF families.

---

¹ This amount represents current year support received for a twelve-month period and does not include amounts paid for prior periods (arrearages) or amounts retained by the federal and state governments to recoup welfare costs.

² The decline partly reflects the decrease in AFDC/TANF caseloads.  Also, some states no longer “pass-through” any child support payments to custodial parents receiving TANF.  Prior to the enactment of PRWORA in 1996, states were required to pass-through the first $50 of any child support collected.

Economic Security Risk Factor 7. Food Insecurity

Figure ECON 7.
Percentage of Households Classified by Food Security Status: 2006

Note:  Food secure households had consistent access to enough food for active, healthy lives for all household members at all times during the year.  Households with low food security obtained enough food to avoid substantial disruptions in eating patterns and food intake, using a variety of coping strategies, such as eating less varied diets, participating in Federal food assistance programs, or getting emergency food from community food pantries or emergency kitchens.  Households with very low food security reported reduced food intake of some household members and their normal eating patterns were disrupted because of the lack of money and other resources.

Source:  U.S. Department of Agriculture, Economic Research Service, Household Food Security in the United States, 2006.

---

• Figure ECON 7 shows the percentage of households that were food secure, had low food security, and had very low food security in 2006.  The majority of U.S. households (89.1 percent) was food secure in 2006; that is, they showed little or no evidence of concern about food supply or reduction in food intake.
• Seven (6.9) percent of U.S. households experienced low food security and 4.0 percent were classified as having very low food security.  Very low food security is defined as having reduced food intake and having normal eating patterns disrupted because of financial constraints.
• Table ECON 7a shows the percentage of households classified by food security status by selected demographic characteristics.
• For households by age categories, households with elderly were more food secure (94.0 percent) than were households with children under six (83.3 percent) or households with children under 18 (84.4 percent).
• There is a relationship between poverty and food security.  Sixty-four (63.7) percent of poor households were food secure compared to 66.9 percent of households below 130 percent of the poverty level, 72.7 percent of households below 185 percent of the poverty level, and 92.9 percent of households at or above 185 percent of the poverty level.
• Married-couple households were less likely to experience food insecurity than female-headed households.  Ten (10.1) percent of married-couple households were food insecure in 2006 compared to 30.4 percent of female-headed households.
• Table ECON 7b shows the percentage of households classified by food security status between 1998 and 2006.  The percentage of households with food insecurity (both low and very low food insecurity) has fluctuated over time from a low of 10.1 percent in 1999 to a high of 11.9 percent in 2004.

Table ECON 7a.
Percentage of Households Classified by Food Security Status and Selected Characteristics: 2006

Table ECON 7b.
Percentage of Households Classified by Food Security Status: 1998-2006

Economic Security Risk Factor 8. Lack of Health Insurance

Figure ECON 8.
Percentage of Persons without Health Insurance by Poverty Status: 2006

Note:  "Poor persons" are defined as those with total family incomes at or below the federal poverty threshold.  Health insurance rates for the education categories include only adults age 18 and over.

Persons of Hispanic ethnicity may be of any race. Beginning in 2002, estimates for Whites and Blacks are for persons reporting a single race only. Persons who reported more than one race are included in the total for all persons but are not shown under any race category.  Due to small sample size, American Indians/Alaska Natives, Asians and Native Hawaiians/Other Pacific Islanders are included in the total for all persons but are not shown separately.  Some of the race categories presented for ECON 8 have been changed slightly from prior year reports to provide more internal consistency throughout this report; in reports prior to 2006, the race categories for Black and White included persons of Hispanic origin.

Source:  Unpublished tabulations from the Current Population Survey, Annual Social and Economic Supplement, 2007.

---

• Figure ECON 8 shows the percentage of persons without health insurance by race and ethnicity, educational attainment, and poverty status for 2006.  Thirty-two (31.6) percent of poor persons were without health insurance as compared to 13.6 percent of non-poor persons. 
• Among race and ethnic groups, poor Hispanics of any race had higher rates of being uninsured (42.9 percent) than did poor Non-Hispanic Whites (27.0 percent) and poor Non-Hispanic Blacks (28.1 percent).
• For non-poor persons, as education increases, the rate of being uninsured decreases.  Twenty-nine (28.5) percent of the non-poor who were not high school graduates were uninsured compared to 17.6 percent of high school graduates, and 6.6 percent of college graduates.
• Among the poor, 41.5 percent of persons who were not high school graduates, 39.9 percent of high school graduates, and 32.5 percent of college graduates were uninsured.
• Table ECON 8 shows the percentage of persons without health insurance by poverty status and demographic characteristics.  Across all demographic categories, poor persons were more likely than non-poor persons to be uninsured regardless of race and ethnicity, gender, educational attainment, age, or family category.
• For poor persons, 19.3 percent of children 17 years of age or less were without health insurance as compared to 51.3 percent of poor adults 25 to 34 years of age. The 25 to 34 year age category had the highest percentage of uninsured among poor persons.
• For non-poor persons, 10.0 percent of the children 17 years of age or less were without health insurance as compared to 26.2 percent of adults 18 to 24 years of age. The 18 to 24 year age category had the highest percentage of uninsured among non-poor persons.

Table ECON 8.
Percentage of Persons without Health Insurance by Poverty Status and Selected Characteristics:  2006

Employment and Work-Related Risk Factors (WORK)

The second grouping, labeled with the WORK prefix, includes eight factors related to employment and barriers to employment.  These measures include data on overall labor force attachment and employment and earnings for low-skilled workers, as well as data on barriers to work.  The latter category includes incidence of adult and child disabilities, adult substance abuse, and levels of educational attainment and school drop-out rates.

Employment and earnings provide many families with an escape from dependence.  It is important, therefore, to look both at overall labor force attachment (WORK 1), and at employment and earnings for those with low education levels (WORK 2 and WORK 3).  The economic condition of the low-skill labor market is a key predictor of the ability of men and women to support families without receiving means-tested assistance.

The next two measures in this group (WORK 4 and WORK 5) focus on educational attainment.  Individuals with less than a high school education have the lowest amount of human capital and are at the greatest risk of being poor, despite their work effort.

Measures of barriers to employment provide indicators of potential work limitations, which may be predictors of greater dependence.  Substance abuse (WORK 6) and disabling conditions among children and adults (WORK 7) all have the potential of limiting the ability of the adults in the household to work.  In addition, debilitating health conditions and high medical expenditures can strain a family’s economic resources.  The labor force participation of women with children (WORK 8) is also a predictor of dependence.

Employment and Work-related Risk Factor 1. Labor Force Attachment

Figure WORK 1.
Percentage of Persons in Families with Labor Force Participants by Race and Ethnicity: 2006

Note:  Full-time, full-year workers (FT/FY) are defined as those who usually worked for 35 or more hours per week, for at least 50 weeks in a given year. Part-time and part-year labor force participation includes part-time workers and individuals who are unemployed, laid off, and/or looking for work for part or all of the year.  This indicator represents annual measures of labor force participation, and thus cannot be compared to monthly measures of labor force participation in Indicator 2.  Persons of Hispanic ethnicity may be of any race. Beginning in 2002, estimates for Whites and Blacks are for persons reporting a single race only. Persons who reported more than one race are included in the total for all persons but are not shown under any race category.  Due to small sample size, American Indians/Alaska Natives, Asians and Native Hawaiians/Other Pacific Islanders are included in the total for all persons but are not shown separately.

Source:  Unpublished tabulations from the Current Population Survey, Annual Social and Economic Supplement, 2007.

---

• Figure WORK 1 shows the percentage of persons in families with labor force participants by race and ethnicity.  In 2006, Hispanics of any race were more likely to live in families with at least one full-time, full-year labor force participant (77.8 percent) than were Non-Hispanic Whites (72.3 percent) or Non-Hispanic Blacks (66.8 percent). 
• Table WORK 1a shows the percentage of persons in families with labor force participants by demographic characteristics.   In 2006, children ages 6 to 15 were more likely to live in families with at least one full- time, full-year labor force participant (80.7 percent) than were children from birth to 5 years of age (78.5 percent).   
• Among family types, persons living in married-couple families were more likely than persons living in other family types to live in families with at least one full-time, full-year labor force participant. 
• Table WORK 1b shows the percentage of persons in families with labor force participants for select years between 1990 and 2006.  The percentage of persons living in families with at least one full-time, full-year labor force participant has fluctuated over time.  The percentage increased from a low of 67.6 percent in 1992 to a high of 73.3 percent in 2000.  In 2006, 72.8 percent of persons lived in families with at least one full-time, full-year worker. 

Table WORK 1a.
Percentage of Persons in Families with Labor Force Participants by Selected Characteristics: 2006

Table WORK 1b.  Percentage of Persons in Families with Labor Force Participants: Selected Years

Employment and Work-related Risk Factor 2. Employment Among the Low-skilled

Figure WORK 2.
Percentage of Persons Ages 18 to 65 with No More than a High School Education Who Were Employed at Any Time during Year by Race and Ethnicity: 1968-2006

Note:  All data include both full and partial year employment for the given calendar year.  Persons of Hispanic ethnicity may be of any race. Beginning in 2002, estimates for Whites and Blacks are for persons reporting a single race only. Due to small sample size, American Indians/Alaska Natives, Asians and Native Hawaiians/Other Pacific Islanders are not shown separately. Hispanic origin was not available until 1975.

Source:  Unpublished tabulations from the Current Population Survey, Annual Social and Economic Supplement, 1969-2007.

---

• Figure WORK 2 shows the employment rate of low-skilled workers ages 18 to 65 (those with a high school education or less) by gender and race and ethnicity between 1968 and 2006.  This measure of low skill is based only on educational attainment and does not take other skills based on work experience, training or other credentials into account.¹
• In 1968, 65.8 percent of Non-Hispanic Black women and 55.8 percent of Non-Hispanic White women with a high school education or less were employed.  In the 1970s, however, Non-Hispanic White women reached parity with their Non-Hispanic Black counterparts and then surpassed them. 
• Employment rates for women with a high school education or less increased during the 1980s and 1990s.  By the 2000s, however, the employment rate for women with a high school education or less peaked and in 2006, the rate declined to 66.5 percent for Non-Hispanic White women, 63.2 percent for Non-Hispanic Black women, and 56.8 percent for Hispanic women of any race.
• In 1968, 92.8 percent of Non-Hispanic White men and 89.9 percent of Non-Hispanic Black men with a high school education or less were employed. 
• Beginning in the 1970s, the employment rates for men with a high school education or less declined and the employment rates between Non-Hispanic White and Non-Hispanic Black men with a high school education or less began to diverge.  In 2006, 80.6 percent of Non-Hispanic White men as compared to 65.6 percent of Non-Hispanic Black men with a high school education or less were employed.
• Over the time period, Hispanic men with a high school education or less have had employment rates similar to Non-Hispanic White men.  In 1998, among men with a high school education or less, the employment rate for Hispanic men surpassed the rate for Non-Hispanic White men.  In 2006, 86.4 percent of Hispanic men with a high school education or less were employed compared to 80.6 percent of Non-Hispanic White men.

Table WORK 2.
Percentage of Persons Ages 18 to 65 with No More than a High School Education Who Were Employed by Race and Ethnicity: 1968-2006

---

¹ This education–based measure of low skill is from the work of Rebecca Blank in “It Takes a Nation: A New Agenda for Fighting Poverty,” 1998.

Employment and Work-related Risk Factor 3. Earnings of Low-skilled Workers

Figure WORK 3.
Mean Weekly Wages of Women and Men Working Full-Time, Full-Year with No More than a High School Education
by Race and Ethnicity (2006 Dollars): 1980-2006

Note:  Last data point is 2006.  Full-time, full-year workers work at least 48 weeks per year and usually work 35 hours per week.

Persons of Hispanic ethnicity may be of any race. Beginning in 2002, estimates for Whites and Blacks are for persons reporting a single race only. Due to small sample size, American Indians/Alaska Natives, Asians and Native Hawaiians/Other Pacific Islanders are not shown separately.

Source:  Unpublished tabulations from the Current Population Survey, Annual Social and Economic Supplement, 1981-2007.

---

• Figure WORK 3 shows the mean weekly wages in 2006 dollars of low-skilled women and men (those with a high school education or less) working full-time, full-year by race and ethnicity for selected years.  This measure of low skill is based only on educational attainment and does not take other skills based on work experience, training or other credentials into account.¹
• In 2006, Non-Hispanic White women with a high school education or less working full-time, full-year earned $574 in an average week compared to $503 for similar Non-Hispanic Black women and $447 for similar Hispanic women of any race.  Among men working full-time, full-year with a high school education or less, Non-Hispanic White men earned $821 in an average week, compared to $639 for Non-Hispanic Black men and $578 for Hispanic men of any race.
• Table WORK 3 provides the detailed estimates used for Figure WORK 3.  In 2006, Non-Hispanic White women had the highest average weekly wages among women working full-time, full-year with a high school education or less at $574.  This represents a 17.4 percent increase in their mean weekly wages between 1980 and 2006.  Over the same time period, similar Non-Hispanic Black women experienced a 12.5 percent increase in their mean weekly wages while similar Hispanic women of any race experienced a 6.9 percent increase.
• Among men working full-time, full-year with a high school education or less, average weekly wages increased 1.5 percent among Non-Hispanic White men and 6.1 percent among Non-Hispanic Black men between 1980 and 2006.  Hispanic men working full-time, full-year with a high school education or less experienced a 5.2 percent decrease in average weekly wages over the same time period.

Table WORK 3.
Mean Weekly Wages of Women and Men Working Full-Time, Full-Year with No More than a High School Education by Race and Ethnicity (2006 Dollars):  1980-2006

---

¹ This education-based measure of low skill is from the work of Rebecca Blank in “It Takes a Nation:  A New Agenda for Fighting Poverty,” 1998.

Employment and Work-related Risk Factor 4. Educational Attainment

Figure WORK 4.
Percentage of Adults Ages 25 and over by Level of Educational Attainment: 1960-2006

Note:  Completing the GED is not considered completing high school for this table.  Beginning with data for 1992, a new survey question results in different categories than for prior years.  Data shown as “High school graduate, no college” were previously from the category “High school, 4 years” and are now from the category “High school graduate.”  Data shown as “One to three years of college” were previously from the category “College 1 to 3 years” and are now the sum of the categories: “Some college” and two separate “Associate degree” categories.  Data shown as “Four or more years of college” were previously from the category “College 4 years or more,” and are now the sum of the categories: “Bachelor's degree,” “Master's degree,” “Doctorate degree” and “Professional degree.”

Source:  U.S. Census Bureau, “Educational Attainment in the United States, 2006,” Current Population Reports and earlier reports.

---

• Figure WORK 4 shows educational attainment for adults 25 years and older between 1960 and 2006.  Table WORK 4 shows the corresponding point estimates for select years. 
• The percentage of the population without at least a high school education has declined over the past 45 years, from 59.0 percent in 1960 to 14.5 percent in 2006.
• The percentage of the population receiving a high school education (with no post secondary education) was 24.6 percent in 1960 and rose to 38.9 percent in 1988.  Since 1988, this figure has fallen to 31.7 percent in 2006.
• Between 1960 and 1990, the percentage of the population with some college (one to three years) doubled, from 8.8 percent to 17.9 percent.  The increase in 1992 is partially the result of a change in survey methodology, but the trend continued upward reaching 25.7 percent in 2006.
• The percentage of the population completing four or more years of college has more than tripled between 1960 and 2006, rising from 7.7 percent to 28.0 percent.

Table WORK 4.
Percentage of Adults Ages 25 and over by Level of Educational Attainment: Selected Years

Employment and Work-related Risk Factor 5. High School Dropout Rates

Figure WORK 5.
Percentage of Students Enrolled in Grades 10 to 12 in the Previous Year Who Were Not Enrolled and Had Not Graduated in the Survey Year
by Race and Ethnicity: 1995-2005

Note:  Beginning in 1987, the U.S. Census Bureau instituted new editing procedures for cases with missing data on school enrollment. Beginning in 1992, the data reflect new wording of the educational attainment item in the Current Population Survey (CPS).

Persons of Hispanic ethnicity may be of any race.  Due to small sample size, American Indians/Alaska Natives, Asians and Native Hawaiians/Other Pacific Islanders are included in the total but are not shown separately.

Source:  U.S. Department of Education, National Center for Education Statistics, Dropout Rates in the United States: 2005 and earlier years (based on Current Population Survey data from the October supplement).

---

• Figure WORK 5 shows the percentage of students who were enrolled in grades 10 through 12 in the previous year but were not enrolled and had not graduated in the survey year by race and ethnicity for the time period 1995 to 2005.  With the exception of Non-Hispanic Blacks, there has been a general downward trend in dropout rates.
• In 2005, the dropout rate was 5.0 percent for Hispanic students of any race, 7.3 percent for Non-Hispanic Black students, and 2.8 percent for Non-Hispanic White students.
• Between 2003 and 2005, Non-Hispanic Blacks experienced an increase in the percentage of students dropping out of school, from 4.8 percent in 2003 to 7.3 percent in 2005.
• Table WORK 5 provides trend data on dropout rates beginning in 1972.  The dropout rate for all races was highest in 1978 and 1979 (6.7 percent) and then declined to 3.6 percent in 2002, a 30-year low.  Since 2002, the dropout rate for students of all races has risen somewhat to 3.8 percent in 2005. 
• Dropout rates among Hispanic students of any race have fluctuated since 1972.  Despite this fluctuation, Hispanic dropout rates were higher than rates for Non-Hispanic White students in all years since 1972 and higher than rates for Non-Hispanic Black students in all reported years except 2005.

Table WORK 5.
Percentage of Students Enrolled in Grades 10 to 12 in the Previous Year Who Were Not Enrolled and Had Not Graduated in the Survey Year
by Race and Ethnicity: 1972 - 2005

Employment and Work-related Risk Factor 6. Adult Alcohol and Substance Abuse

Figure WORK 6.
Percentage of Adults Who Used Cocaine or Marijuana or Abused Alcohol by Age: 2006

Note:  Cocaine and marijuana use is defined as use during the past month.  “Binge alcohol use” is defined as drinking five or more drinks on the same occasion on at least one day in the past 30 days.  “Heavy alcohol use” is defined as drinking five or more drinks on the same occasion on each of five or more days in the past 30 days; all heavy alcohol users are also binge alcohol users.

Source:  U.S. Department of Health and Human Services, National Survey on Drug Use and Health, 2007.

---

• Figure WORK 6 shows the percentage of adults who used cocaine, the percentage who used marijuana, and the percentage who abused alcohol by age group in 2006.  Adults 18 to 25 years of age were more likely than older adults to report cocaine, marijuana, binge alcohol or heavy alcohol use in the prior month.  For example, 16.3 percent reported using marijuana in the past month during 2006, compared with 8.5 percent of adults 26 to 34 years of age and 3.2 percent of adults 35 years and over.
• The percentage of adults reporting binge alcohol use was larger than the percentages for all other reported behaviors across all age groups.  Among those reporting binge alcohol use, however, this behavior was more prevalent among adults 18 to 25 years of age than among adults in other age categories.
• Table WORK 6 shows trend data for cocaine, marijuana, binge alcohol and heavy alcohol use for the years 1999 to 2006.
• For adults in all age groups, alcohol abuse increased between 2005 and 2006.  Cocaine use increased for adults 26 to 34 years of age, and marijuana use increased for adults ages 35 and over during the same two-year period.

Table WORK 6.
Percentage of Adults Who Used Cocaine or Marijuana or Abused Alcohol by Age:  1999-2006

Employment and Work-related Risk Factor 7. Adult and Child Disability

Figure WORK 7.
Percentage of the Non-Elderly Population Reporting an Activity Limitation by Selected Characteristics: 2006

Note:  Work disability is defined as limitations in or the inability to work as a result of a physical, mental or emotional health condition.  Individuals are identified as having long-term care needs if they need the help of others in handling either personal care needs (eating, bathing, dressing, getting around the home) or routine needs (household chores, shopping, getting around for business or other purposes).  Disability program recipients include persons covered by Supplemental Security Income (SSI), Social Security Disability Insurance (SSDI), Special Education Services, Early Intervention Services and/or disability pensions.

Respondents were defined as having an activity limitation if they answered positively to any of the questions regarding:  (1) work disability (see definition above; (2) long-term care needs (see definition above); (3) difficulty walking; (4) difficulty remembering; (5) for children under 5, limitations in the amount of play activities they can participate in because of physical, mental or emotional problems; (6) for children 3 and over, receipt of Special Educational or Early Intervention Services; and, (7) any other limitations due to physical, mental or emotional problems.

Persons of Hispanic ethnicity may be of any race. Due to small sample size, American Indians/Alaska Natives, Asians and Native Hawaiians/Other Pacific Islanders are included in the total for all persons but are not shown separately.

Source:  Unpublished tabulations from the National Health Interview Survey, 2007.

---

• Figure WORK 7 shows the percentage of non-elderly adults and children reporting an activity limitation by race and ethnicity in 2006.  Non-elderly adults were more likely than children to have an activity limitation, 10.9 percent compared to 7.6 percent.
• Table WORK 7 shows the percentage of the non-elderly population reporting a disability by selected demographic characteristics.  While non-elderly adults were more likely than children to report an activity limitation, a higher percentage of children (6.5 percent) than adults (4.9 percent) were actually recipients of disability program benefits in 2006.
• For both non-elderly adults and children, the percentage of Non-Hispanic Blacks with an activity limitation was higher than the percentages for Non-Hispanic Whites and Hispanics.
• Among non-elderly adults, rates of work disability and long-term care needs were lower for Hispanics (5.4 and 1.3 percent, respectively) than for Non-Hispanic Whites (8.9 and 2.3 percent, respectively) and Non-Hispanic Blacks (10.4 and 2.9 percent, respectively).

Table WORK 7.
Percentage of the Non-Elderly Population Reporting a Disability by Selected Characteristics: 2006

Employment and Work-related Risk Factor 8. Labor Force Participation of Women with Children Under 18

Figure WORK 8.
Labor Force Participation of Women with Children under 18: 1975-2006

Note:  The labor force participation rate includes all women who are employed, laid off or unemployed but looking for work. The employment rate includes only those women who are employed. The population of mothers with children under age 18 includes those 16 years of age and older.

Source:  U.S. Department of Labor, Bureau of Labor Statistics, unpublished tabulations from the Current Population Survey, Annual Social and Economic Supplement, 1976-2007.

---

• Figure WORK 8 shows the labor force participation rates for mothers with children under 18 years of age by marital status between 1975 and 2006.  In 2006, regardless of marital status, the majority of mothers in the U.S. were engaged in the labor force. 
• Historically, divorced, widowed and separated mothers have had the highest rates of labor force participation among mothers.  In 1975, 62.8 percent of divorced, widowed or separated mothers were in the labor force as compared to 44.9 percent of married mothers with spouses present and 42.2 percent of never-married mothers.  In 2006, divorced, widowed and separated mothers remained more likely than other mothers to participate in the labor force.
• Between 1992 and 2002, labor force participation rates for never-married mothers with children under 18 markedly increased—rising from 52.5 percent in 1992 to 75.3 percent in 2002.  Since 1998, labor force participation rates for never-married mothers have exceeded the rates for married mothers.
• The labor force participation rate of married mothers with children under 18 followed an upward trend from 1975 until 1997 when it peaked at 71.1 percent.  In 2006, 68.4 percent of married mothers with spouses present were in the labor force.
• Table WORK 1 shows both the labor force participation rate and the employment rate of mothers with children under 18 years of age between 1975 and 2006.
• The employment rate for all mothers increased over the time period.  The employment rate for married mothers with a spouse present was 40.5 percent in 1975; in 2006 the employment rate was 66.2 percent.  The employment rate for divorced, widowed and separated mothers was 54.9 percent in 1975; in 2006 the employment rate was 75.4 percent.  The employment rate for never- married mothers was to 32.1 percent in 1975; in 2006 the rate climbed to 62.5 percent.

Table WORK 8.
Employment Status of Women with Children under 18 Years of Age: 1975-2006

Nonmarital Birth Risk Factors (BIRTH)

The final group of risk factors addresses nonmarital childbearing. The tables and figures in this subsection are labeled with the BIRTH prefix.  This category includes long-term time trends in nonmarital births (BIRTH 1), nonmarital teen births (BIRTH 2 and BIRTH 3), and children living in families with never-married parents (BIRTH 4).  Children living in families with never-married mothers are at high risk of becoming dependent as adults, and it is therefore important to track changes in the size of this vulnerable population.

As noted above, the predictors/risk factors included in this chapter do not represent an exhaustive list of measures.  They are merely a sampling of available data that address in some way the question of how a family is faring on the scale of deprivation and well-being.  Such questions are a necessary part of the discussion on dependence as researchers assess the effects of welfare reform.

Nonmarital Birth Risk Factor 1. Nonmarital Births

Figure BIRTH 1.
Percentage of Births that are Nonmarital by Age: 1940-2006

Note:   Trends in non-marital births may be affected by changes in the reporting of marital status on birth certificates and in procedures for inferring non-marital births when marital status is not reported.

Source:  National Center for Health Statistics, “Nonmarital Childbearing in the United States, 1940-1999,” National Vital Statistics Reports, Vol. 48 (16), 2000; “Births: Preliminary Data for 2006,” National Vital Statistics Reports, Vol. 56 (7), December 2007, http://www.cdc.gov/nchs/data/nvsr/nvsr56/nvsr56_07.pdf.

---

• Figure BIRTH 1 shows the percentage of births that were nonmarital by age group from 1940 to 2006 and Table BIRTH 1 shows corresponding estimates for selected years.  Changes in nonmarital births reflect changes in the rate at which unmarried women have children, the rate at which married women have children and the rate at which women marry.  The percentage of children born outside of marriage to women of all ages has increased over the past 60 years.  In 1940, 3.8 percent of births were to unmarried women.  In 2006, the percentage increased to 38.5 percent.
• Teen births, as shown in Figure BIRTH 1 and Table BIRTH 1, show nonmarital teen births as a percentage of all teen births.  In 1940, 14.0 percent of births to teens were nonmarital.  While the percentage of all teen births that are nonmarital has increased since the mid-1960s, growth in the percentage slowed in the mid- to late- 1990s before rising to 84.4 percent in 2006.
• Over the past 10 years, the percentage of nonmarital births among all births to women 20 to 24 years of age increased by 27.0 percent from 45.6 percent in 1996 to 57.9 percent in 2006. This compares to an increase of 10.6 percent in the percentage of nonmarital births among teen births over the same period. 
• Since 1994, the percentage of births that are nonmarital remains steady among Black teens and all Black women. Among White teens and all White women, the trend continues upward (see Table C-1 in Appendix C for nonmarital birth data by age and race).

Table BIRTH 1.
Percentage of Births that are Nonmarital by Age: Selected Years

Nonmarital Birth Risk Factor 2. Nonmarital Teen Births

Figure BIRTH 2.
Percentage of All Births to Unmarried Teens Ages 15 to 19 by Race and Ethnicity: 1940-2005

Note:  Trends in nonmarital births may be affected by changes in the reporting of marital status on birth certificates and in procedures for inferring nonmarital births when marital status is not reported.  Beginning in 1980, data are tabulated by the race of the mother. Prior to 1980, data are tabulated by the race of the child.  Teens are defined as people ages 15 to 19.

Race categories include those of Hispanic ethnicity.  Persons of Hispanic ethnicity may be of any race. Due to small sample size, American Indians/Alaska Natives, Asians and Native Hawaiians/Other Pacific Islanders are included in the total for all persons but are not shown separately.

Prior to 1969, race data were available for Whites and Non-Whites only.

Source:  National Center for Health Statistics, “Nonmarital Childbearing in the United States, 1940 - 1999,” National Vital Health Statistics Reports, Vol. 48 (16), 2000; “Births: Final Data for 2005,” National Vital Statistics Reports, Vol. 56 (6), December 2007 http://www.cdc.gov/nchs/data/nvsr/nvsr56/nvsr56_06.pdf.

---

• Figure BIRTH 2 shows the percentage of all births to unmarried teens 15 to 19 years of age by race and ethnicity, and Table BIRTH 2 shows corresponding estimates for selected years between 1940 and 2005.  Unlike BIRTH 1, which showed nonmarital teen births as a percentage of all teen births, BIRTH 2 shows births to unmarried teens as a percentage of births to all women.  This percentage is affected by several factors: the age distribution of women, the marriage rate among teens, the birth rate among unmarried teens and the birth rate among all other women.
• The percentage of all births that were to unmarried teens fell over the last eight years, from 9.7 in 1997 to 8.3 percent in 2005.   
• Among Black women, the percentage of all births that were nonmarital teen births fell to 15.8 percent in 2005.  This is the lowest percentage since 1969, the first year in which data on Black women were collected.
• Among White women, the percentage of all births that were to unmarried White teens ages 15 to 19 remained virtually unchanged between 2002 and 2005 at approximately 7.2 percent.
• Among Hispanic women, the percentage of all births that were to unmarried teens increased from a low of 9.8 percent in 1990 to a high of 12.1 percent in 1998 before declining to 11.0 percent in 2005.

Table BIRTH 2.
Percentage of All Births to Unmarried Teens Ages 15 to 19 by Race and Ethnicity:  Selected Years

Nonmarital Birth Risk Factor 3. Nonmarital Teen Birth Rates

Figure BIRTH 3a.
Births per 1,000 Unmarried Teens Ages 15 to 17 by Race: 1960-2005

Figure BIRTH 3b.
Births per 1,000 Unmarried Teens Ages 18 and 19 by Race: 1960-2005

Note:  Rates are per 1,000 unmarried women in specified group. Trends in non-marital births may be affected by changes in the reporting of marital status on birth certificates and in procedures for inferring non-marital births when marital status is not reported.  Beginning in 1980, data are tabulated by the race of the mother.  Prior to 1980, data are tabulated by the race of the child.

Race categories include those of Hispanic ethnicity. Due to small sample size, American Indians/Alaska Natives, Asians and Native Hawaiians/Other Pacific Islanders are included in the total for all persons but are not shown separately.

Prior to 1969, race data were available for Whites and Non-Whites only.

Source:  National Center for Health Statistics, “Nonmarital Childbearing in the United States, 1940-1999,” National Vital Statistics Reports, Vol. 48 (16), 2000; “Births: Final Data for 2005,” National Vital Statistics Reports, Vol. 56 (6), December 2006.  Birthrates for 1950 to 1965 computed by ASPE staff from NCHS birth and Census population estimates.

---

• Figures BIRTH 3a and 3b show births per thousand unmarried teens between the ages of 15 to 17 and 18 to 19 from 1960 to 2005.  Table BIRTH 3 shows corresponding estimates for selected years between 1950 and 2005.  The birth rate per thousand unmarried teens ages 15 to 17 fell in 2005 for both Black and White teens.  The rate for Black teens ages 15 to 17 has been cut by more than half from 79.9 per thousand in 1991 to 35.4 per thousand in 2005.  This 2005 rate of 35.4 per thousand is lower than in any other year since 1969, the first year in which data on Black women were collected.
• The birth rates of unmarried teens in the older age group (18 and 19 years) showed a slight increase between 2002 and 2005.  For Black teens ages 18 and 19, the birth rate fell from a high of 147.7 per thousand in 1991 to a low of 100.4 per thousand in 2003 before increasing to 101.6 births per thousand in 2005.
• Prior to 1994, birth rates among unmarried White teens in both age groups rose steadily for over four decades.  For White teens 15 to 17 years of age, the birth rate increased from 3.4 births per thousand unmarried teens in 1950 to 23.9 births per thousand unmarried teens in 1994. For the 18 to 19 year olds, the rate increased from 8.5 births per thousand unmarried teens in 1950 to 55.7 births per thousand unmarried teens in 1994.  Since 1994, rates for both age groups have generally followed a downward trend.
• While birth rates among unmarried Black teens remain high compared to rates for unmarried White teens, the gap between Black and White teens narrowed during the 1990s and 2000s.

Table BIRTH 3.
Births per Thousand Unmarried Teen Women by Age and Race: 1950-2005

Nonmarital Birth Risk Factor 4. Never-married Family Status

Figure BIRTH 4.
Percentage of All Children Living in Families with a Never-Married Female Head by Race and Ethnicity: 1982-2007

Note:  Data are for all children under 18 who are not family heads (excludes householders, subfamily reference persons and their spouses).  Inmates of institutions also are excluded. Children who are living with neither of their parents are excluded from the denominator.  Based on Current Population Survey (CPS) data.

Race categories include those of Hispanic ethnicity. Persons of Hispanic ethnicity may be of any race. Beginning in 2002, estimates for Whites and Blacks are for persons reporting a single race only. Persons who reported more than one race are included in the total for all persons but are not shown under any race category.  Due to small sample size, American Indians/Alaska Natives, Asians and Native Hawaiians/Other Pacific Islanders are included in the total for all persons but are not shown separately.

Source:  U.S. Census Bureau, “Marital Status and Living Arrangements, ”Current Population Reports, Series P20-212, 287, 365, 380, 399, 418, 423, 433, 445, 450, 461, 468, 478, 484, 491, 496, 506, 514 and “America’s Families and Living Arrangements,” Current Population Reports, Series P20-537, 547, 553 and ASPE tabulations of the CPS for 2007.

---

• Figure BIRTH 4 shows the percentage of all children living in families with a never-married female head of household by race and ethnicity from 1982 to 2007.  Table BIRTH 4 shows corresponding estimates for selected years between 1960 and 2007.  The percentage of children living in families with never-married female heads increased from 4.6 percent in 1982 to 11.0 percent in 2007.
• The percentage of White children living in families headed by never-married women has increased fourfold over the past 25 years, from 1.6 percent in 1982 to 6.6 percent in 2007. 
• Among Hispanics of all races, the percentage of children living with a never-married female head of household tripled over the past 25 years, from 5.7 percent in 1982 to 12.9 percent in 2007.
• The percentage of Black children living in families with a never-married female head of household has been higher than the percentages for other groups throughout the time period.  In 2007, 34.6 percent of Black children lived in families with a never-married female head of household compared to 6.6 percent for White children and 12.9 percent for Hispanic children.

Table BIRTH 4.
Number and Percentage of All Children Living in Families with a Never-Married Female Head by Race and Ethnicity: Selected Years

Appendices

Appendix A. Program Data

Contents

• Temporary Assistance for Needy Families (TANF) and Aid to Families with Dependent Children (AFDC)
• Food Stamp Program (FSP)
• Supplemental Security Income (SSI)

The Welfare Indicators Act of 1994 specifies that the annual welfare indicators reports shall include analyses of families and individuals receiving assistance under three means-tested benefit programs:

• The Temporary Assistance for Needy Families (TANF) program authorized under part A of title IV of the Social Security Act (which replaced the Aid to Families with Dependent Children (AFDC) program in 1996);
• The Food Stamp Program under the Food Stamp Act of 1977, as amended;
• The Supplemental Security Income (SSI) program under title XVI of the Social Security Act.

This chapter includes information on these three programs, derived primarily from administrative data reported by state and federal agencies instead of the national survey data presented in previous chapters.  National caseloads and expenditure trend information on each of the three programs is included, as well as state-by-state trend tables and information on the characteristics of program participants.

Temporary Assistance for Needy Families (TANF) and Aid to Families with Dependent Children (AFDC)

The Aid to Families with Dependent Children (AFDC) program — originally named the Aid to Dependent Children program — was established by the Social Security Act of 1935 as a grant program to enable states to provide cash welfare payments for needy children who had been deprived of parental support or care because their fathers or mothers were absent from the home, incapacitated, deceased, or unemployed. All 50 states, the District of Columbia, Guam, Puerto Rico, and the Virgin Islands operated an AFDC program.  States defined “need,” set their own benefit levels, established (within federal limitations) income and resource limits, and administered the program or supervised its administration.  States were entitled to unlimited federal funds for reimbursement of benefit payments, at “matching” rates that were inversely related to state per capita income.  States were required to provide aid to all persons who were in classes eligible under federal law and whose income and resources were within state-set limits.

During the 1990s, the federal government increasingly used its authority under section 1115 of the Social Security Act to waive portions of the federal requirements under AFDC. This allowed states to test such changes as expanded earned income disregards, increased work requirements and stronger sanctions for failure to comply with them, time limits on benefits, and expanded access to transitional benefits such as child care and medical assistance.  As a condition of receiving waivers, states were required to conduct rigorous evaluations of the impacts of these changes on the welfare receipt, employment, and earnings of participants.

Public Law 104-193, the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA), replaced AFDC, AFDC administration, the Job Opportunities and Basic Skills Training (JOBS) program and the Emergency Assistance (EA) program with a block grant called the Temporary Assistance for Needy Families (TANF) program. Key elements of TANF include a lifetime limit of five years (60 months)¹ on the amount of time a family with an adult can receive assistance funded with federal funds, increasing work participation rate requirements that states must meet, and broad state flexibility on program design.  Spending through the TANF block grant is capped and funded at $16.5 billion per year, slightly above FY 1995 federal expenditures for the four component programs.  States also must meet a “maintenance of effort (MOE) requirement” by spending on needy families at least 75 percent of the amount of state funds used in FY 1994 on these programs (80 percent if they fail work participation rate requirements).

TANF gives states wide latitude in spending both federal TANF funds and state MOE funds.   Subject to a few restrictions, TANF funds may be used in any way that supports one of the four statutory purposes of TANF:  to provide assistance to needy families so that children can be cared for at home; to end the dependence of needy parents on government benefits by promoting job preparation, work and marriage; to prevent and reduce the incidence of out-of-wedlock pregnancies; and to encourage the formation and maintenance of two-parent families.

Legislative Changes

The current legislative authority for the TANF block grant is from the Deficit Reduction Act of 2005 (Public Law 109-171).  Enacted in February 2006, the Act reauthorizes the original 1996 legislation at an annual funding level of $16.5 billion and continues to require each state to have at least 50 percent of its work eligible families participating in meaningful work activities.  However, prior to this Act, a caseload reduction credit allowed states to reduce their work requirement by their caseload declines since 1995.  As most states experienced dramatic caseload declines, the credit had virtually eliminated the work participation requirements for most states.  Starting with FY 2007, the Deficit Reduction Act recalibrates the base year for calculating the caseload reduction credit to 2005, effectively re-implementing a meaningful performance guideline.

Also starting in FY 2007, the Deficit Reduction Act expands the work participation calculations to include adults in certain welfare programs funded out of state funds countable toward the maintenance of effort (MOE) requirement.  Under the original legislation, these adults were excluded from the calculations.  This change was implemented because there was some concern that states were moving work-eligible TANF adults into non-TANF programs with similar program structures, in part, to avoid federal work participation standards.² In addition, new regulations from the Department of Health and Human Services create consistent definitions of the activities that count toward federal work requirements and provide new flexibility for states to count adults who miss scheduled hours due to holidays and excused absences.  The new regulations also provide more detailed instructions to states as to which families they are to include in their work participation rate calculations.  In some circumstances states are required to include adults that have been removed from the assistance unit because of failing to comply with program rules.  In addition, the new regulations allow states to include adults receiving federal disability benefits on a case by case basis, and to exclude parents caring for disabled family members.

The Deficit Reduction Act also provides $100 million per year to support programs designed to promote healthy marriages, and up to $50 million annually for programs designed to encourage responsible fatherhood.  In addition, the new law increased mandatory child care funding to states to $2.9 billion annually.

Data Issues Relating to the TANF Program and the AFDC-TANF Transition

States had the option of beginning their TANF programs as soon as PRWORA was enacted in August 1996, and a few states began TANF programs as early as September 1996.  All states were required to implement TANF by July 1, 1997.  Because states implemented TANF at different times, the FY 1997 data reflect a combination of the AFDC and TANF programs.  In some states, limited data are available for FY 1997 because states were given a transition period of six months after they implemented TANF before they were required to report data on the characteristics and work activities of TANF participants.

Because of the greatly expanded range of activities allowed under TANF, a substantial portion of TANF funds are being spent on activities other than cash payments to families. Table TANF 4 in this Appendix which tracks overall expenditure trends includes only those TANF funds spent on “cash and work-based assistance” and “administrative costs,” not on work activities, supportive services, or other allowable uses of funds.  Spending on these other activities is detailed in Table TANF 5.  Note that TANF administrative costs include funds spent administering all activities, not just cash and work-based assistance. (Administrative costs under AFDC had included a small amount of funds for administering AFDC child care programs; such programs, and the costs of administering them, were transferred to the Child Care and Development Fund as part of PRWORA.)

There also is potential for discontinuity between the AFDC and the TANF caseload figures.  For example, under TANF there is no longer a separate “Unemployed Parent” (UP) program, as there was under AFDC.  While a separate work participation rate is calculated for two-parent families, this population is not identical to the UP caseload under AFDC.  It is also possible that a limited number of families will be considered recipients of TANF assistance, even if they do not receive a monthly cash benefit.  The vast majority of families receiving “assistance”³are, in fact, receiving cash payments.

Another data issue concerns the treatment of families who receive cash and other forms of assistance under Separate State Programs (SSPs), funded out of MOE dollars rather than federal TANF funds.  Under TANF, some states use SSP programs to serve specific categories of families (e.g., two-parent families, families who have exhausted their time limits).  From 1996-2005, such families were not subject to federal time limits.  States did not have to include them in the calculation of their work participation rates.  As of October 2006, such families are included in the work participation rate calculation, but may still be excluded from the application of the federal time limits on receipt of assistance.  Starting with the 2004 edition, this Indicators report adds recipients in SSPs into the caseload totals⁴ (the split between TANF and SSP caseloads is shown in Table TANF 3, nationally, and in Table TANF 15, by state).  Native Americans served through state TANF and SSP programs are included in these caseload counts, but families served through TANF programs operated by Tribal governments are excluded. Expenditures for SSPs are shown in Table TANF 5.

AFDC/TANF Program Data

The following tables and figures present data on caseloads, expenditures, and recipient characteristics of the AFDC and TANF programs.  Trends in national caseloads and expenditures are shown in Figures TANF 1 and TANF 2, and the first set of tables (Tables TANF 1 through 6).  These are followed by information on characteristics of AFDC/TANF families (Table TANF 7)⁵ and a series of tables presenting state-by-state data on trends in the AFDC/TANF program (Tables TANF 8 through 15).  These data complement the data on trends in AFDC/TANF recipiency and participation rates shown in Tables IND 3a and IND 4a in Chapter II.

AFDC/TANF Caseload Trends (Tables TANF 1 through TANF 3 and Figure TANF 1).  Welfare caseloads have stabilized over the past few years after declining dramatically during the 1990s.  In FY 2006, the average monthly number of TANF recipients was 4.7 million persons, down 7 percent from FY 2005.  Moreover, this was 62 percent lower than the average monthly AFDC caseload in FY 1996 and the smallest number of people on welfare since 1967. From the peak of 14.2 million in FY 1994, the number of AFDC/TANF recipients dropped by 67 percent to 4.7 million in FY 2007.⁶ Over four-fifths of the reduction in the caseload since FY 1994 has occurred following the passage of PRWORA in FY 1996.  These are the largest welfare caseload declines in the history of U.S. welfare programs.

Several studies have attempted to explain the unprecedented decline in caseloads and, specifically, to disentangle the effects of PRWORA and welfare reform from the simultaneous growth in the U.S. economy.  Separating these effects is difficult, however, because PRWORA was enacted at a time when the economy was expanding dramatically, offering a uniquely conducive environment within which to move many recipients off the welfare rolls and into the labor market.  Other policy changes, most notably expansions in the Earned Income Tax Credit, add further complexity.

In general, studies have found that both economic conditions and welfare reform policies have played important roles in the recent caseload decline.  A review of a dozen studies concluded that roughly 15 to 30 percent of the caseload decline prior to 1996 was attributed by most studies to welfare policies under waivers to the AFDC rules with approximately 30 to 45 percent of the decline explained by economic conditions (Schoeni and Blank, 2000).  A study by the Council of Economic Advisers (1999) of the post-PRWORA period finds that just over one-third of the caseload decline can be explained by welfare reform policy, while 8 to 10 percent is due to the economy.  A more recent study estimates that over half the decline in caseloads after enactment of PRWORA was attributable to welfare reform (O’Neill and Hill, 2001).

AFDC/TANF Expenditures (Tables TANF 4 through TANF 6 and Figure TANF 2). Tables TANF 4 and 5 show trends in expenditures on AFDC and TANF.  Table TANF 4 tracks both programs, breaking out the costs of benefits and administrative expenses.  It also shows the division between federal and state spending.  Table TANF 5 shows the variety of activities funded under the TANF program.

Figure TANF 2 and Table TANF 6 show that inflation has had a significant effect in eroding the value of the average monthly AFDC/TANF benefit.  In real dollars, by 2006 the average monthly benefit per recipient had declined to 65 percent of what it was at its peak in the late 1970s.

AFDC/TANF Recipient Characteristics (Table TANF 7). With the dramatic declines in the welfare rolls since the implementation of TANF, there has been a great deal of speculation regarding how the composition of the caseload has changed.  Two striking trends are the increases in the proportion of families with no adult in the assistance unit and in employment among adult recipients.

One of the most dramatic trends is the increase in the proportion of adult recipients who are working. In FY 2006, 22 percent of TANF adult recipients were employed, down from 26 percent in 2000, but up from 11 percent in FY 1996 and 7 percent in FY 1992, as shown in Table TANF 7.  Adding in those in work experience and community service positions, the percentage working was 30 percent in FY 2006⁷ (data not shown).  Similar trends are shown in data on income from earnings. These trends likely reflect the effects of expanded earnings disregards, welfare-to-work programs, and the economy.  One can also see a relationship between employment of welfare recipients and broader trends in labor force participation. (For example, see Table WORK 8 in Chapter III for trends in employment rates for women with children under age 18.)

Another dramatic change in the caseload is the increasing fraction of cases without an adult recipient. Such cases occur when the adults are ineligible (because they are a caretaker relative, SSI parent, immigrant parent, or sanctioned parent).  Families with no adults in the assistance unit have climbed from 15 percent of the caseload in FY 1992 to 47 percent in FY 2006.⁸ This dramatic growth has been due to an increase in the number of cases without recipient adults during the early 1990s, followed by a decline in the number of cases that included adults in the assistance unit.  The number of cases without an adult in the assistance unit has fallen by about 127,000 since 1996 — between 1996 and 1998 they decreased by 254,000 but subsequently increased by 127,000.

In other areas, TANF administrative data show fewer changes in composition than might have been expected.  There has been widespread anecdotal evidence that the most job ready recipients — those with the fewest barriers to employment — have already exited the welfare caseload and have stopped coming onto the welfare rolls, leaving a more disadvantaged population remaining.  However, as the expectations for welfare recipients have increased, and fewer recipients are totally exempted from work requirements, others have speculated that the most disadvantaged recipients may also have been sanctioned off the rolls or terminated for failure to comply with administrative requirements.  In fact, analyses of program data have not found much evidence of an increase or decrease in readily observed barriers to employment in the current caseload.

The question of whether the caseload has become more disadvantaged cannot be answered simply through TANF administrative data provided by the states, which do not contain detailed information on such barriers to employment as lack of basic skills, alcohol and drug abuse, domestic violence, and disabilities.  A few recent studies have found very high levels of these barriers among the TANF population.  These studies also have found that the more barriers a recipient faces, the less likely she is to find a job and maintain consistent employment over a period of time.

AFDC/TANF State-by-State Trends (Tables TANF 8 through TANF 15).  There is a great deal of state-to-state variation in the trends discussed above.  For example, as shown in Table TANF 10, while every state has experienced a caseload decline since the 1990s, the percentage change between the state’s caseload peak and March 2007 ranges from 96 percent (Wyoming) to 44 percent (Nebraska).  Sixteen states have experienced caseload declines of 75 percent or more.  Table TANF 10 also shows that states reached their peak caseloads as early as May 1990 (Louisiana) and as late as June 1997 (Hawaii).

Table TANF 15 shows TANF and Separate State Program (SSP) families and recipients, by state.  Thirty-two states (including DC) had such programs.

---

Figure TANF 1.
AFDC/TANF Families Receiving Income Assistance

Note: “Basic Families” are single-parent families and “UP Families” are two-parent cases receiving benefits under AFDC Unemployed Parent programs that operated in certain states before FY 1991 and in all states after October 1, 1990. The AFDC Basic and UP programs were replaced by TANF as of July 1, 1997 under the Personal Responsibility and Work Opportunity Reconciliation Act of 1996. Shaded areas indicate NBER designated periods of recession from peak to trough. The decrease in number of families receiving assistance during the 1981-82 recession stems from changes in eligibility requirements and other policy changes mandated by OBRA 1981. Beginning in 2000, “Total Families” includes TANF and SSP families. Last data point plotted is March 2007.

Source: U.S. Department of Health and Human Services, Administration for Children and Families, Office of Planning, Research, and Evaluation.

---

Figure TANF 2.
Average Monthly AFDC/TANF Benefit per Recipient in Constant 2006 Dollars

Note: See Table TANF 6 for underlying data.  Comparison of trends in the average monthly AFDC/TANF benefit per recipient in constant 2006 dollars with the weighted average maximum benefit in constant 2006 dollars since 1988 indicates that the primary cause of the decline in the average monthly benefit has been the erosion of the real value of the maximum benefit due to inflation.  This is due to the fact that the current value of the maximum benefits has increased less than the cost of living in most states since the late1980s.

Source: U.S. Department of Health and Human Services, Administration for Children and Families, Office of Family Assistance, Quarterly Public Assistance Statistics, 1992 & 1993 and earlier years along with unpublished data.

---

Table TANF 1.
Trends in AFDC/TANF Caseloads: 1962-2006

Table TANF 2.
Number of AFDC/TANF Recipients, and Recipients as a Percentage of Various Population Groups: 1970-2006

Table TANF 3.
TANF and Separate State Program (SSP) Families and Recipients: 2000-2006
[In thousands]

Table TANF 4.
Total AFDC/TANF Expenditures on Cash Benefits and Administration: 1970 – 2006
[In millions of dollars]

Table TANF 5.
Federal and State TANF Program and Other Related Spending: 1997 – 2006
[In millions of dollars]

Table TANF 6.
Trends in AFDC/TANF Average Monthly Payments: 1962 – 2006

Table TANF 7.
Characteristics of AFDC/TANF Families: Selected Years 1969 – 2006

Table TANF 8.
AFDC/TANF Benefits by State: Selected Fiscal Years 1978 – 2006
[In millions of dollars]

Table TANF 9.
Comparison of Federal Funding for AFDC and Related Programs and 2006 Family Assistance Grants Awarded under PRWORA
[In millions of dollars]

Table TANF 10.
AFDC/TANF Caseload by State: October 1989 to March 2007 Peak
[In thousands]

Table TANF 11.
Average Monthly AFDC/TANF Recipients by State: Selected Fiscal Years
[In thousands]

Table TANF 12.
AFDC/TANF Recipiency Rates for Total Population by State: Selected Fiscal Years
[In percent]

Table TANF 13.
Average Number of AFDC/TANF Child Recipients by State: Selected Fiscal Years
[In thousands]

Table TANF 14.
AFDC/TANF Recipiency Rates for Children by State: Selected Fiscal Years 1965 – 2006
[In percent]

Table TANF 15.
TANF and Separate State Program (SSP) Families and Recipients: 2006
[In thousands]

Endnotes

¹ Many states limit TANF assistance to less than the 60-month federal maximum.

² Separate State Programs (SSP) refer to programs funded by state MOE contributions.  Some states have additional programs that are similar to TANF, but are not funded by TANF or MOE sources.  These programs are sometimes called Solely State Funded programs and are excluded from any federal work standards and the 60-month limit on assistance.  Since States do not report data on these programs they are not included in any of the tables in this report.

³ States are allowed to use TANF funds on a variety of services, including employment and training services, domestic violence services, child care, transportation, and other support services. Families receiving such services, however, generally should not be counted as recipients of TANF “assistance.”  Under the final regulations for TANF, “assistance” primarily includes payments directed at ongoing basic needs.  It includes payments when individuals are participating in community service and work experience (or other work activities) as a condition of receiving payments (e.g., workfare).  In addition, the definition also includes certain child care and transportation benefits when families are not employed.   It excludes, however, such things as:  non-recurrent, short-term benefits; services without a cash value, such as education and training, case management, job search, and counseling; and benefits such as child care and transportation when provided to employed families.

⁴ States began submitting caseload data on SSPs in FY 2000.

⁵ Family characteristics in Table TANF 7 may differ from those reported in Chapter II because the administrative data focus on the assistance unit, whereas the survey-based data in Chapter II often use a broader family unit definition.  For example, grandparents, adult siblings, aunts, uncles, and other adult relatives living in the same household as the recipient children may be excluded from the assistance unit and thus the administrative data, yet be included in survey data on the family in which the TANF recipient resides.

⁶  Note that these figures include recipients in SSPs, who are sometimes omitted from TANF caseload statistics reported by the Department.

⁷ Not all of these adults are participating in enough hours to meet the TANF Work Participation Rate requirement.

⁸ The percentages in this paragraph do not include cases served by SSP programs.  In FY 2006, 14.2 percent of SSP caseloads funded by MOE did not have an adult in the assistance unit compared to 47.2 percent of families served through the main TANF programs.

Food Stamp Program

The Food Stamp Program (FSP), administered by the U.S. Department of Agriculture’s (USDA) Food and Nutrition Service, is the largest food assistance program in the country, reaching more poor individuals over the course of a year than any other public assistance program.  Unlike many other public assistance programs, FSP has few categorical requirements for eligibility, such as the presence of children, elderly, or disabled individuals in a household.  As a result, the program offers assistance to a large and diverse population of needy persons, many of whom are not eligible for other forms of assistance.

The Food Stamp Program was designed primarily to supplement the food purchasing power of eligible low-income households so they can buy a nutritionally adequate low-cost diet.  Participating households are expected to be able to devote 30 percent of their counted monthly cash income (after adjusting for various deductions) to food purchases.  Food stamp benefits then make up the difference between the household’s expected contribution to its food costs and an amount judged to be sufficient to buy an adequate low-cost diet. This amount, the maximum food stamp benefit level, is derived from USDA’s lowest-cost food plan, the Thrifty Food Plan (TFP).  

The federal government is responsible for virtually all of the rules that govern the program, and, with limited variations, these rules are nationally uniform, as are the benefit levels.  Nonetheless, states, the District of Columbia, Guam, and the Virgin Islands, through their local welfare offices, have primary responsibility for the day-to-day administration of the program.  They determine eligibility, calculate benefits, and issue food stamp allotments. The Food Stamp Act provides 100 percent federal funding of food stamp benefits.  States and other jurisdictions have responsibility for about half the cost of state and local food stamp agency administration.  

In addition to the regular Food Stamp Program, the Food Stamp Act authorizes alternative programs in Puerto Rico, the Northern Mariana Islands, and American Samoa. The largest of these, the Nutrition Assistance Program in Puerto Rico, was funded under a federal block grant of nearly $1.6 billion in 2006.  Unless noted otherwise, the food stamp caseload and expenditure data in this Appendix exclude costs for the Nutrition Assistance Program (NAP) in Puerto Rico.  (Prior to 2004, editions of this Appendix included NAP, but caseload and expenditure data in this Appendix are now limited to the Food Stamp Program, to be consistent with FSP data published by the USDA.)

The Food Stamp Program is available to nearly all financially needy households.  To be eligible for food stamps, a household must meet eligibility criteria for gross and net income, asset holdings, work requirements, and citizenship or immigration status.  The FSP benefit unit is the household.  Generally, individuals living together constitute a household if they customarily purchase and prepare meals together.  The income, expenses and assets of the household members are combined to determine program eligibility and benefit allotment.

Certain households are categorically eligible for food stamps and therefore not subject to income or asset limits.  Households are categorically eligible if all of their members receive SSI, cash or in-kind TANF benefits, or General Assistance.

Monthly income is the most important determinant of household eligibility.  Except for categorically-eligible households, or households containing elderly or disabled members, gross income cannot exceed 130 percent of poverty.  After certain amounts are deducted for living expenses, working expenses, dependent care expenses, excess shelter expenses, child support payment, and - for elderly/disabled households - medical expenses, net income cannot exceed 100 percent of poverty.  Non categorically-eligible households also must not have more than $2,000 in assets comprised of cash, savings, stocks and bonds, and in some states some vehicles; households with an elderly or disabled member can have up to $3,000 in countable assets.

All nonexempt adult applicants for food stamps must register for work.  To maintain eligibility, they must accept a suitable job, if offered one, and fulfill any work, job search, or training requirements established by the FSP office.  Nondisabled adults living in households without children can receive benefits for three months only, unless they work or participate in work-related activities.  Participation is restricted for certain groups, including students, strikers, and people who are institutionalized.  Legal immigrants who are disabled, under age 18, were admitted as refugees or asylees, or have at least five years of legal US residency are eligible; all other noncitizens are not.

Food stamp benefits are a function of a household’s size, its net monthly income, its assets, and maximum monthly benefit levels.  Allotments are not taxable and food stamp purchases may not be charged sales taxes.  Receipt of food stamps does not affect eligibility for or benefits provided by other welfare programs, although some programs use food stamp participation as a “trigger” for eligibility and others take into account the general availability of food stamps in deciding what level of benefits to provide.

Legislative Changes

Title IV and subtitle A of title VIII of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA) made major changes to the Food Stamp Program, including strong work requirements on able-bodied adults without dependent children, restricted eligibility of legal immigrants, and a reduction in maximum benefits.  These three provisions, and subsequent amendments, are discussed below; their impact on program participation and expenditures begins to appear in food stamp administrative data for 1997, with the fuller impact shown in data for 1998 and beyond.  

First, a work requirement was added for able-bodied adult food stamp recipients without dependents (ABAWDs).  Unless exempt, ABAWDs between the ages of 18 and 59 are not eligible for benefits for more than 3 months in every 36-month period unless they are:  (1) working at least 20 hours a week; (2) participating in and complying with a work program for at least 20 hours a week; or (3) participating in and complying with a workfare program.  Under the original legislation, the Department of Agriculture was authorized to waive application of the work requirement to any group of individuals at the request of the state agency, if a determination was made that the area where they reside has an unemployment rate over 10 percent or does not have a sufficient number of jobs to provide them employment.  The provision was further moderated under the Balanced Budget Act of 1997 (Public Law 105-33), which allowed states to exempt up to 15 percent of the ABAWD caseload (beyond those subject to waivers) and which increased funds for the food stamp employment and training program for the creation of job slots for able-bodied adults subject to time limits.

Separately, title IV of PRWORA (Public Law 104-193) made significant changes in the eligibility of noncitizens for food stamp benefits.  As first enacted, most qualified aliens, including legal immigrants (illegal aliens were already ineligible) were barred from receiving food stamps until citizenship. Subsequently, the Agriculture Research, Extension and Education Reform Act of 1998 (Public Law 105-185) restored food stamp eligibility to certain groups of qualified aliens who were legally residing in the United States before passage of PRWORA on August 22, 1996 and were over 65 years of age on that date or were under age 18 or disabled.

Finally, the 1996 legislation restrained growth in future program expenditures by making changes in the benefit structure for eligible participants, including a reduction in the maximum food stamp allotment.  Other provisions of the 1996 act disqualified from eligibility those convicted of drug-related felonies and gave states the option to disqualify individuals, both custodial and non-custodial parents, from food stamps when they do not cooperate with child support agencies or are in arrears in their child support.

Between 1996 and 2001, regulatory and legislative changes were made to increase access to food stamps among working poor families.  Regulatory changes announced in July 1999 and expanded in November 2000 allowed states to reduce reporting requirements and made it easier for working families to report income changes on a semiannual basis.  Under the November 2000 regulations, states also were given the option of providing a three-month transitional food stamp benefit to most families leaving TANF.  Regulations that went into effect in 2001 expanded categorical eligibility to those receiving noncash TANF benefits, excluded vehicles with little equity from the assets test, and eliminated the equity test for most vehicles.  In addition, the Agriculture Appropriations Bill for 2001 (Public Law 106-387) provided states with the option of liberalizing the treatment of vehicle assets to align with the states’ TANF rules on vehicle eligibility.  These changes were intended to address concerns that some of the decline in food stamp caseloads may be leaving poor families without nutritional assistance as they make the transition from welfare dependence to full self-sufficiency.

The Farm Security and Rural Investment Act of 2002 – also known as the 2002 Farm Bill – reauthorized the Food Stamp Program through fiscal year 2007. This law brought a number of significant changes to the program, including some that supersede earlier changes made through PRWORA and subsequent FSP legislation and regulations.  Specifically, the 2002 Farm Bill restores food stamp eligibility to legal immigrants who have lived in the country at least five years and to legal immigrants receiving disability benefits, regardless of entry date.  Children of legal immigrants also are eligible for food stamps regardless of entry date.  Effective in fiscal year 2004, the requirement that income and resources of an immigrant’s sponsor be counted in determining the eligibility and benefit amounts for immigrant children was eliminated. Each provision became effective at a different time, but all restorations were in effect by October 1, 2003.

The 2002 Farm Bill also increased the asset limit from $2,000 to $3,000 for households with a disabled member, making it consistent with the limit for households with elderly, and replaced the fixed standard deduction with a deduction that varies according to household size and is indexed to cost-of-living increases, in recognition of the higher expenses larger households incur.  For households in the 48 contiguous states and DC, Alaska, Hawaii and the Virgin Islands, the deduction is set at 8.31 percent of the applicable net income limit based on household size.  (Households in Guam will receive a slightly higher deduction.)  No household receives an amount less than the previous fixed standard deduction or more than the standard deduction for a household of six.

Other 2002 Farm Bill changes include the authorization of $5 million per year for education and outreach grants to help inform the low-income public of their eligibility for food stamps, and increased flexibility for states in spending Employment and Training program funds to promote work.  States also are now allowed to extend from three months to up to five months the period of time households may receive transitional food stamp benefits when they lose TANF cash assistance.  Benefits are equal to the amount the household received prior to termination of TANF with adjustments in income for the loss of TANF.  This change helps individuals moving off cash assistance to make the transition from welfare to work.

The 2002 Farm Bill also implemented a number of administrative reforms and program simplifications, including:

• changing the quality control system so that only those states with persistently high error rates will face liabilities;
• awarding bonuses to states that improve the quality and accuracy of their service;
• allowing states to exclude certain types of income and resources not counted under TANF or Medicaid, such as educational assistance, when determining food stamp eligibility;
• allowing states to deem child support payments as income exclusions rather than deductions as an incentive for parents to pay child support;
• allowing states to simplify the standard utility allowance (SUA) if the state elects to use the SUA rather than actual utility costs for all households, thus reducing administrative burden, costs and errors;
• permitting states to use a standard deduction from income of $143 per month for homeless households with some shelter expenses;
• allowing states to extend simplified reporting procedures to all households, not just households with earnings;
• eliminating the requirement that the Electronic Benefit Transfer (EBT) system be cost-neutral to the federal government to help support the EBT conversion process;
• allowing USDA to use alternative methods for issuing food stamp benefits during times of disaster when use of EBT is impractical;
• requiring food stamp applications be made available through the Internet; and
• combining Puerto Rico and American Samoa’s block grants into one grant and indexing both with inflation.

Food Stamp Program Data

The following six tables and accompanying figure provide information about the Food Stamp Program:

• Tables FSP 1 and FSP 2 and Figure FSP 1 present national caseload and expenditure trend data on the Food Stamp Program as discussed below;
• Table FSP 3 presents some demographic characteristics of the food stamp caseload; and
• Tables FSP 4 through FSP 6 present some state-by-state trend data on the FSP through fiscal year 2006.

Food Stamp Caseload Trends (Table FSP 1).  Average monthly food stamp participation was 26.7 million persons in fiscal year 2006, excluding the participants in Puerto Rico’s block grant.  This represents a significant increase over the fiscal year 2000 record-low average of 17.2 million participants.  It is, however, still below the peak of 27.5 million recipients in fiscal year 1994.  See also Table IND 3b and Table IND 4b in Chapter II for further data trends in food stamp caseload, specifically, food stamp recipiency and participation rates.  

Considerable research has demonstrated that the Food Stamp Program is responsive to economic changes, with participation increasing in times of economic downturns and decreasing in times of economic growth (see Figure FSP 1).  Economic conditions alone did not explain the caseload growth in the late 1980s and early 1990s, however.  Studies suggest that a variety of factors contributed to this caseload growth, including a weak economy and higher rates of unemployment, expansions in Medicaid eligibility, the legalization of 3 million undocumented immigrants, and longer participation spells (McConnell, 1991; Gleason, 1998).

The decline in participation from 1994 to 2000 was caused by several factors, according to studies of this period.  Part of the decline is associated with the strong economy in the second half of the 1990s.  However, participation fell more sharply than expected during this period of sustained economic growth.  Some of the decline reflected restrictions on the eligibility of noncitizens and time limits for unemployed nondisabled childless adults.  Participation fell most rapidly among the following three groups: noncitizens and their US-born children, unemployed nondisabled childless adults, and persons receiving cash welfare benefits.  As people left the welfare rolls, many also stopped participating in food stamps, even while remaining eligible (Genser, 1999; Wilde et al., 2000; Gleason et al., 2001; Kornfeld, 2002).

The increase in FSP participation from 2000 to 2005 occurred during a period when unemployment increased from four percent to five percent, eligibility was restored to many legal immigrants, states took advantage of opportunities to expand categorical eligibility to those receiving noncash TANF benefits and services and to liberalize the treatment of vehicles, and the Food and Nutrition Service was encouraging states to conduct outreach efforts and simplify the program.  In addition, the proportion of eligible households participating in the Food Stamp Program, increased from 50 percent in 2000 to 59 percent in 2005.  Between 2000 and 2005, food stamp participation increased by 3.6 million households (see Table IND 4b).  Part of this increase was associated with an increase in the number of eligible households and part was associated with an increased participation rate among those households that were eligible.  

Food Stamp Expenditures.  Total program costs, shown in Table FSP 2, were higher in 2005 and 2006 than in 2004, reflecting the increase in participation during that period as well as an increase in average benefits.  Total federal program costs were $32.8 billion in 2006, $32.2 billion in 2005, and $29.0 billion in 2004 (after adjusting for inflation).  Average monthly benefits per person, also shown in Table FSP 2, were $94.30 per person in 2006, $96.00 in 2005 and $92.10 in 2004 (after adjusting for inflation).  The personal monthly benefit decreased 1.8 percent between 2005 and 2006.

Food Stamp Household Characteristics.  As shown in Table FSP 3, the proportion of food stamp households with earnings has increased, from about 20 percent for most of the 1980s and early 1990s, to 30 percent in 2006.  At the same time, the proportion of households with income from AFDC/TANF has declined, from 42 percent in 1990 to 13 percent in 2006, following the dramatic decline in AFDC/TANF caseloads.  Over half of all food stamp households have children, although the proportion has declined from over 60 percent in most of the 1980s and early 1990s to 52 percent in 2006.  The majority (87 percent in 2006) of households have incomes below the federal poverty guidelines.

 

---

Figure FSP 1.
Persons Receiving Food Stamps: 1962–2006
(In millions)

Note: Shaded areas are periods of recession as determined by the National Bureau of Economic Research.

Sources: U.S. Department of Agriculture, Food and Nutrition Service, data published online at http://www.fns.usda.gov/pd/fssummar.htm and unpublished data from the Food Stamps National Data Bank.

---

Table FSP 1.
Trends in Food Stamp Caseloads:  Selected Years 1962–2006

Table FSP 2.
Trends in Food Stamp Expenditures: Selected Years 1975–2006

Table FSP 3.
Characteristics of Food Stamp Households: Selected Years 1980–2006

Table FSP 4.
Value of Food Stamps Issued by State: Selected Fiscal Years 1975–2006
[In millions]

Table FSP 5.
Average Number of Food Stamp Recipients by State: Selected Fiscal Years
[In thousands]

Table FSP 6.
Food Stamp Recipiency Rates by State: Selected Fiscal Years
[In percent]

Supplemental Security Income

The Supplemental Security Income (SSI) Program is a means-tested, federally administered income assistance program authorized by title XVI of the Social Security Act.  Established in 1972 (Public Law 92-603) and begun in 1974, SSI provides monthly cash payments in accordance with uniform, nationwide eligibility requirements to needy aged, blind and disabled persons.  To qualify for SSI payments, a person must satisfy the program criteria for age, blindness, or disability.  Children may qualify for SSI if they are under age 18 and meet the applicable SSI disability or blindness, income and resource requirements.  Individuals and married couples are eligible for SSI if their countable incomes fall below the federal maximum monthly SSI benefit levels of $623 for an individual and $934 for a married couple (if both are eligible) in fiscal year 2007.  SSI eligibility is restricted to qualified persons who have countable resources/assets of not more than $2,000, or $3,000 for a couple.

The Social Security Administration (SSA) administers the SSI program. Since its inception, SSI has been viewed as the “program of last resort.” Therefore, SSA helps recipients obtain any other public assistance that they are eligible to receive before providing SSI benefits.  After evaluating all other income, SSI pays what is necessary to bring an individual to the statutorily prescribed income “floor.”

Prior to the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA), no individual could receive both SSI payments and Aid to Families with Dependent Children (AFDC) benefits.  If eligible for both, the individual had to choose which benefit to receive.  Generally, the AFDC agency encouraged individuals to file for SSI and, once the SSI payments had started, the individual was removed from the AFDC filing unit.  Since states have the authority to set TANF eligibility standards and benefit levels under PRWORA, there is no federal prohibition against individuals receiving both TANF benefits and SSI.

With the exception of California, which converted food stamp benefits to cash payments that are included in the state supplementary payment, SSI recipients may be eligible to receive food stamps.  If all household members receive SSI, the household is categorically eligible for food stamps and does not need to meet the Food Stamp Program’s financial eligibility standards.  If SSI beneficiaries live in households in which other household members do not receive SSI benefits, the household must meet the net income eligibility standard of the Food Stamp Program to be eligible for food stamp benefits.

Legislative Changes

Public Law 104-121, the Contract with America Advancement Act of 1996, prohibited SSI eligibility to individuals whose drug addiction and/or alcoholism (DA&A) is a contributing factor material to the finding of disability.  This provision applied to individuals who filed for benefits on or after the date of enactment (March 29, 1996) and to individuals whose claims were finally adjudicated on or after the date of enactment.  It applied to current beneficiaries on January 1, 1997.

PRWORA made several changes designed to maintain the SSI program’s goal of limiting benefits to severely disabled children.   First, the act replaced the former “comparable severity” test with a new definition of disability specifically for children, based on a medically determinable physical or mental impairment that results in “marked and severe functional limitations.” Second, SSA discontinued use of the Individualized Functional Assessment (IFA) for children which it had implemented in 1991 following the Supreme Court's decision in Sullivan v Zebley, 493 U.S. 521 (1990).¹ Third, references to “maladaptive behaviors” in certain sections of the Listing of Impairments (among medical criteria for evaluation of mental and emotional disorders in the domain of personal/behavioral function) were eliminated.  The latter two provisions were effective for all new and pending applications upon enactment (August 22, 1996).  Beneficiaries who were receiving benefits due to an IFA or under the Listings because of limitations resulting from maladaptive behaviors received notice no later than January 1, 1997, that their benefits might end when their case was redetermined. Additional provisions of PRWORA with impact on enrollment are the requirement that eligibility be redetermined when beneficiaries reach age 18, using the adult disability standard; that "continuing disability reviews" be done for children; and that children who were eligible due to low birth weight have their eligibility redetermined at age one.

Title IV of Public Law 104-193 (PRWORA) also made significant changes in the eligibility of noncitizens for SSI benefits.  Some of the restrictions were subsequently moderated by Public Law 104-208, Public Law 106-169, and most notably by the Balanced Budget Act of 1997 (Public Law 105-33), which “grandfathered” immigrants who were receiving SSI at the time of enactment of the PRWORA.  Those immigrants who entered the U.S. after August 22, 1996, may be eligible to receive SSI after having been “lawfully admitted for permanent residence.”  In addition, Public Law 106-386, the Victims of Trafficking and Violence Protection Act of 2000, provides that noncitizens who are victims of “severe forms of trafficking in persons in the United States” shall be treated as refugees for purposes of SSI and be eligible for SSI benefits for the first 7 years they are in the United States.

Several provisions aimed at reducing SSI fraud and improving recovery of overpayments were enacted in 1999 as part of the Foster Care Independence Act of 1999 (Public Law 106-169). Other legislation enacted in 1999 (Public Law 106-170) provides additional work incentives for disabled beneficiaries of SSI (e.g., the Ticket to Work and Self-Sufficiency Program).

The Social Security Protection Act of 2004 (Public Law 108-203), enacted March 2, 2004, introduced program and beneficiary protections covering the use of representative payees and required documentation of changes in beneficiary status.  It also extended SSI eligibility to blind or disabled children living with a parent assigned to permanent U.S. military duty outside of the U.S. but who were not receiving SSI while in the U.S.  Furthermore, Public Law 109-163 provides that individuals who were made ineligible for SSI because of their spouses or parents being called to active military duty would not have to file a new application for SSI benefits if they again could be eligible for benefits before the end of 24 consecutive months of ineligibility.

The Deficit Reduction Act of 2005 (Public Law 109-171) included two SSI program reforms, designed to improve the accuracy of disability determinations and benefit awards, among other program goals.

SSI Program Data

The following tables and figures provide SSI program data:

• Tables SSI 1 through SSI 5 and Figure SSI 1 present national caseload and expenditure trend data on the SSI program; 
• Table SSI 6 presents demographic characteristics of the SSI caseload;
• Tables SSI 7 through SSI 9 present state-by-state trend data on the SSI program through fiscal year 2006.

SSI Caseload Trends (Tables SSI 1 and SSI 2 and Figure SSI 1). From 1990 to 1995, the number of SSI beneficiaries increased from 4.8 million to 6.5 million, an average growth rate of over 7 percent per year.  Between 1995 and 2000, the number of beneficiaries fluctuated between 6.5 and 6.6 million persons.  Between 2000 and 2006, the caseload increased from 6.6 to 7.2 million beneficiaries, an average annual growth rate of 1.5 percent. Table SSI 1 presents information on the total number of persons receiving SSI payments in December of each year from 1974 through 2006, and also presents recipients by eligibility category (aged, blind, and disabled) and by type of recipient (child, adults ages 18-64, and adults ages 65 or older).  See also Tables IND 3c and IND 4c in Chapter II for further data on trends in recipiency and participation.

The composition of the SSI caseload has been shifting over time, as shown in Table SSI 1.  The number of beneficiaries eligible because of age has been declining steadily, from a high of 2.3 million persons in December 1975 to a low of 1.2 million persons in December 2004 and has since remained essentially unchanged.  At the same time, there has been strong growth in blind and disabled beneficiaries, from 1.7 million in December 1974 to 6.0 million in December 2006.  Moreover, the number of disabled children has increased dramatically, particularly during the 1990s, when the number of disabled children receiving SSI increased from 309,000 in December 1990 to 955,000 in December 1996.  The number of disabled children fell over the next three years, but has been increasing since 2000, reaching just under 1.1 million children in 2006.

Several factors have contributed to the growth of the Supplemental Security Income program. Expansions in disability eligibility (particularly for mentally impaired adults and for children), increased outreach, overall growth in immigration, and transfers from state programs were among the key factors identified in a 1995 study by the Government Accountability Office (GAO).  GAO concluded that three groups – adults with mental impairments, children, and non-citizens – accounted for nearly 90 percent of the SSI program’s growth in the early 1990s.  The growth in disabled children beneficiaries is generally believed to be due to outreach activities, the Supreme Court decision in the Zebley case, expansion of the medical impairment category, and reduction in reviews of continuing eligibility.²

SSI Expenditures (Tables SSI 3 through SSI 5). While down slightly from 2004, the total amount of federally administered SSI benefits has increased over the past 5 years from $35.6 billion (inflation adjusted) in 2001 to over $38.9 billion in 2006, as shown in Table SSI 3. Average monthly federally administered benefits per person were $455 in 2006, up (4.4 percent) from 2001 inflation adjusted benefit level of $436. For more details see Table SSI 4.

SSI Recipient Characteristics (Table SSI 6). Over the last 20 years, the percentage of aged SSI recipients has dramatically decreased, while the percentage of disabled recipients has increased substantially. As shown in Table SSI 6, the proportion of SSI aged recipients has decreased dramatically, from 44 percent in 1980 to under 17 percent in 2006.  During the same period, the percentage of disabled recipients increased from 55 percent in 1980 to 82 percent in 2006.

 

---

Figure SSI 1.
SSI Recipients by Age: 1974 – 2006

Source:  Social Security Administration, Supplemental Security Income, Annual Statistical Report, 2007 (available online at http://www.ssa.gov/policy/docs/statcomps/supplement/2007/index.html).

---

Table SSI 1.
Number of Persons Receiving Federally Administered SSI Payments: 1974 – 2006
[In thousands]

Table SSI 2.
SSI Recipiency Rates: 1974 – 2006

Table SSI 3.
Federally Administered SSI Benefits and Administration: 1974 – 2006 ¹
[In millions of dollars]

Table SSI 4.
Average Monthly Federally Administered SSI Benefits: 1975 – 2006 ¹
[In millions of dollars]

Table SSI 5.
Number of Persons Receiving Federally Administered SSI Payments by Eligibility Category
[In thousands]

Table SSI 6.
Characteristics of SSI Recipients by Selected Characteristics: Selected Years 1980-2006

Table SSI 7.
Total Federally Administered SSI Payments by State: Calendar Year 2006
[In thousands]

Table SSI 8.
SSI Recipiency Rates by State And Program Type:  1979 and 2006
[In percent]

Table SSI 9.
SSI Recipiency Rates by State: Selected Fiscal Years 1975 – 2006
[In Percent]

Endnotes

Appendix B. Alternative Definition of Dependence Based on Income from TANF and Food Stamps

As directed by the Welfare Indicators Act of 1994 (Public Law 103-432), this annual report on Indicators of Welfare Dependence focuses on dependence on three programs: the Temporary Assistance for Needy Families (TANF) program, formerly the Aid to Families with Dependent Children (AFDC) program; the Food Stamp Program; and the Supplemental Security Income (SSI) program.  We adopt the following definition of welfare dependence for this report:

• Welfare dependence is the proportion of all individuals in families that receive more than half of their total family income in one year from TANF, food stamps and/or SSI.

This appendix examines an alternative definition of dependence that considers TANF and food stamps alone, excluding SSI.  As shown in Table B-1, the rate of dependency would have been much lower — only 2.1 percent — in 2005 if based on income from TANF and food stamps, as opposed to 3.8 percent when counting income from all three programs (TANF, food stamps and SSI).

There also is significant variation across age groups in the programs upon which individuals are dependent.  The elderly depend more on SSI than on TANF and food stamps; whereas 2.2 percent of elderly persons are dependent when counting the three major types of means-tested assistance, very few, 0.2 percent, are dependent when the definition is limited to TANF and food stamps.  In contrast, children are primarily dependent on TANF and food stamps.

Dependency on AFDC/TANF and food stamp receipt has declined since 1995, while dependency on SSI receipt alone has remained stable, as shown in Table B-2.  As a result, the difference between the standard definition (based on all three programs) and the alternative definition (based on TANF and food stamps only) has grown. In 1995, over two-thirds (68 percent) of individuals who were dependent under the standard definition also were dependent under the alternative definition shown in this appendix.  By 2005, the proportion had dropped to just over half (55 percent).  If this report had focused on the alternative definition of dependence, it would have shown an even larger decline in dependence than usually reported.  For example, between 1995 and 2005, dependency declined by 42 percent (3.6 percent to 2.1 percent) under the alternative definition, compared to a decline of 28 percent (5.3 percent to 3.8 percent) under the standard definition.

Table B-1.
Percentage of the Total Population with More than 50 Percent of Income from Various Means-Tested Assistance Programs
by Selected Characteristics: 2005

Table B-2.
Percentage of the Total Population with More than 50 Percent of Income
from Various Means-Tested Assistance Programs: 1995-2005

 

Appendix C. Additional Nonmarital Birth Data

Table C-1.
Percentage of Births to Unmarried Women within Age Groups
by Race and Ethnicity: Selected Years 1940-2005

Table C-2.
Percentage of Births that are to Unmarried Women by State: Selected Years 1960-2005

Table C-3.
Percentage of Births that are to Unmarried Women by Race and Ethnicity and State: 1994 and 2005

Table C-4.
Birth Rates of Teens 15-19 Years by State: Selected Years 1960-2005
[Births per 1,000 women in specified group]

Table C-5.
Birth Rates of Teens 15-19 Years by Race and Ethnicity and State: Selected Years
[Births per 1,000 women in specified group]

 

Appendix D. Technical Notes

Age Categories

Most of the indicators are shown by age categories, generally children ages 0 to 15, adults ages 16 to 64, and adults 65 and older. Youth 17 and 18 years of age are often classified with adults because they are considered potential members of the labor force in many labor force statistics. Many of the risk factors, however, use published data that define “children” to include all individuals less than 18 years of age.

Annual and Monthly Measures

There are differences between monthly and annual observation of benefit receipt. The measures of annual recipiency (that is, any receipt over the course of a year) shown in Figure and Table SUM 1 are higher than the more traditional measures of recipiency in an average month, as shown in several other indicators.

Note that annual measures are for calendar years except where explicitly noted as fiscal years.

Race and Ethnicity

Most of the data sources allow analysis of the indicators and predictors of welfare dependence across several age and racial/ethnic categories. Where the data are available, statistics are shown for three racial/ethnic groups — Non-Hispanic White, Non-Hispanic Black, and Hispanic. Due to small sample size, American Indians/Alaska Natives, Asians and Native Hawaiians/Other Pacific Islanders are included in the totals for all persons but are not shown under separate race categories. In some instances, however, data are shown for “Whites” and “Blacks,” rather than for “Non-Hispanic Whites” and “Non-Hispanic Blacks;” in such cases these racial categories include individuals of Hispanic Origin. Footnotes to the tables provide further documentation of issues related to race and ethnicity.

Estimates based on 2002 (and more recent) Current Population Survey (CPS) data are affected by a change in the CPS questionnaire that allows individuals to report one or more races. This change was implemented to comply with the 1997 Standards for Federal Data on Race and Ethnicity. In 2000, the Office of Management and Budget (OMB) published guidelines for implementing these new standards. To accommodate the race categories under the new standards, CPS estimates for racial/ethnic categories beginning in 2002 are for persons who are Non-Hispanic White (and no other race), Non-Hispanic Black (and no other race) and Hispanic (of any race). Persons who reported more than one race are included in the total for all persons but are not shown under any race category.

Family Structure Categories

For the primary measure of dependency, as well as selected indicators and measures, estimates are provided for individual persons by family structure (see SUM 1, IND 1, IND 2, IND 5, and ECON 7). For these measures, the entire population is subdivided into the following four groups:

• Persons in Married-Couple Families
• Persons in Female-Headed Families
• Persons in Male-Headed Families
• Unrelated Persons.

Two additional measures use a subset of the above categories (see IND 4, and ECON 1).

Spells

Spells of program recipiency (IND 7), spells of welfare receipt with no attachment to the labor market (IND 8) and spells of poverty (ECON 5) are limited to those spells that begin during the SIPP panel of observation. Spells separated by only 1 month are not considered separate spells. If an individual has 2 or more spells of dependency or receipt, each is counted separately in the analysis.

Unit of Analysis

The individual, rather than the family or household, is the unit of analysis for most of the statistics in this report. The individual’s dependency status, however, is generally based on total family income, taking into account means-tested assistance, earnings and other sources of income for all individuals in the family.¹ The introductory chapter of this report, for example, shows the percentage of individuals that are dependent (in SUM 1) or poor (in SUM 2) according to annual total family income. Recipiency status is also based on total annual family income in some instances; in SUM 1, for example, recipients are individuals in families receiving assistance at some point in the year. In most other indicators, however, recipiency is measured as the direct receipt of a benefit by an individual in a month. The difference between an individual and a family measure of recipiency is largest in the SSI program, which provides benefits to individuals and couples, not to families.

---

¹ Family is generally defined as following the broad U.S. Census Bureau definition of family — all persons residing together that are related by birth, marriage, or adoption.