CHARACTERISTICS AND FINANCIAL CIRCUMSTANCES
OF TANF RECIPIENTS

OCTOBER 1999 – SEPTEMBER 2000

This is a report of the demographic characteristics and financial circumstances of families who received assistance and families who no longer received assistance under the Temporary Assistance for Needy Families (TANF) program during the period of October 1999 – September 2000. It also includes the characteristics data of the Separate State Program - Maintenance of Effort (SSP-MOE) recipient families. This report is available on the Internet at “http://www.acf.hhs.gov/programs/opre/particip/index.htm#characteristics”.

The data referenced by this document were obtained from a statistically valid sample of TANF and SSP-MOE cases within the national TANF/SSP-MOE Database. Data are presented for all States, the District of Columbia, Guam, Puerto Rico, and the Virgin Islands.

Prepared by:
Administration for Children and Families
Office of Planning, Research and Evaluation
370 L’Enfant Promenade, S.W.
Washington, DC 20447

CONTENTS

DOWNLOADABLE FILES

INTRODUCTION

The Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA) established a new Temporary Assistance for Needy Families (TANF) program to replace the Aid to Families with Dependent Children (AFDC) program. The new welfare law also established State data reporting requirements for the TANF program. The Department of Health and Human Services (HHS) issued final regulations implementing data reporting and other aspects of PRWORA in April 1999.

The TANF statute requires States to collect TANF data on a monthly basis and report them to HHS on a quarterly basis. These data include disaggregated case record information on the families receiving assistance, families no longer receiving assistance and families newly-approved for assistance from programs funded with TANF funds. Effective with the October-December 1999 quarter, all States, the District of Columbia, Guam, Puerto Rico and the Virgin Islands were requested to implement the quarterly reporting requirements in the final TANF rules. All 54 States and Territories transmitted 8,391,790 active cases and 649,794 closed cases onto the national TANF database for fiscal year (FY) 2000. Oregon transmitted its closed cases after preparing this report. In addition, 24 States with separate State Maintenance of Effort (SSP-MOE) programs reported characteristics data on recipients covered by these separate State programs. Their SSP-MOE recipient characteristics are presented in Appendix I.

Under the TANF data reporting system, States have the option to submit either sample data or universe data to HHS. Thirty States submitted universe data, from which HHS randomly selected approximately 275 active cases and 100 closed cases each month to prepare this report. The remaining 24 States submitted sample data. A total sample of 198,415 active cases and 54,340 closed cases was used to compile 54 tables for TANF recipient characteristics. The statistical data in this report are estimates derived from samples and, therefore, are subject to sampling errors as well as non-sampling errors. Statistical specifications can be found in Appendix II, the reliability of estimates.

Implementation of the final rules of TANF/SSP-MOE data collection requirements posed significant initial challenges to State and HHS. Despite every effort to ensure the reliability of the data, HHS eliminated some problem/questionable data from this report. In cases where a few States submitted questionable data, the data from those States were eliminated from the tables. In cases where numerous States reported questionable data or unusually large numbers of "unknown" or "other" categories, HHS urges caution in drawing conclusions on the basis of the data.

back to top

SUMMARY

This summary describes TANF recipient characteristics in October 1999 - September 2000. Wherever possible, the recipient characteristics were compared with those in the preceding year, October 1998 - September 1999. As presented in Exhibit I, October 1999 - September 2000 TANF recipient characteristics were also compared between active cases, newly-approved cases, child-only cases, and closed cases. In Appendix I there are SSP-MOE recipient characteristics for 24 States who reported characteristics data on their SSP-MOE families. Because SSP-MOE eligible families may be quite different among these 24 States, it was not possible on a national level to compare characteristics of SSP-MOE recipients with those of TANF recipients.

The TANF Family

The average monthly number of TANF families was 2,269,000 in fiscal year (FY) 2000. The estimated total number of TANF recipients was 1,579,000 adults and 4,385,000 children. The average monthly number of TANF families decreased in almost all States and reflects an overall 14 percent decrease from 2,648,000 families in October 1998 - September 1999. During FY 2000, 147,000 applicant families were approved each month to receive TANF assistance, and 2,034,000 TANF families had their assistance terminated.

California had the largest number of TANF families with a monthly average of 501,000, which accounted for almost a quarter of the U.S. total TANF families. New York ranked second with a monthly average of 258,000. California and New York had a combined monthly average of 759,000, which accounted for a third of the U.S. totals. Of the $9.5 billion paid to TANF eligible families in cash assistance during FY 2000, California alone paid $2.9 billion, which accounted for 30 percent of the U.S. total cash payments. New York made total cash payments of $1.4 billion. The combined TANF cash payments of California and New York accounted for almost a half of the U.S. total TANF cash payments.

The average number of persons in TANF families was 2.6 persons. The TANF families averaged 2 recipient children, which remained unchanged. Two in five families had only one child. One in 10 families had more than three children. However, the average number of children in closed-case families was 1.8. Nearly one in two closed-case families had one child, and only 7 percent had more than three children.

Two-thirds of TANF families had only one adult recipient, and four percent included two or more adult recipients. In 15 States, there were no two-parent family cases to be used in calculating the work participation rate. These States placed two-parent families under separate State programs. About 35 percent of TANF families had no adult recipients, up about 5.4 percentage points when compared to the October 1998 - September 1999 period. Child-only cases were 21 percent of newly-approved families and 25 percent of closed-case families. Even though the percentage of child-only cases on the welfare rolls has continued to increase in the past several years, the total number of child-only cases has actually declined by about 200,000 since FY 1996.

Nearly half of child-only families had a parent in the household. Of those child-only families with a parent present, about 42 percent had a parent on SSI and 26 percent had a parent in unknown citizenship/alienage status. Only 11 percent had a parent in sanction status because of refusing to comply with work requirements, not attending school, or not cooperating with child support. The remaining 21 percent had a parent in unknown reasons for their not being in the assistant unit.

There was little change in the racial composition of TANF families. Black families comprised 39 percent of TANF families. White families comprised 31 percent of the families and 25 percent were Hispanic or Latino. Of TANF families, 1.6 percent were American Native, and 2.2 percent were Asian. It is notable that 28 percent of child-only families were Hispanics. Of all newly-approved families, 39 percent were black families, 37 percent were white families and 20 percent were Hispanics. Of all closed-case families, 36 percent were black families, 38 percent were white families and 21percent were Hispanics.

Of TANF families, 80 percent received Food Stamp assistance, which is consistent with previous levels. Those families received the average monthly Food Stamp assistance of $223. Of closed-case families, about 70 percent received Food Stamp assistance.

Almost every TANF family was eligible to receive medical assistance under the State plan approved under title XIX.

Reasons for which TANF families received a reduction in assistance for the reporting month were: sanction at 6.1 percent, recoupment of a prior overpayment at 8.7 percent and other at 8.8 percent. Over 50 percent of sanctions were due to a failure to comply with work requirements. "Other" could include reasons for a reduction in assistance, such as length of receiving assistance, or the application of a family cap.

The reasons for TANF families no longer receiving assistance were due to employment at 19.7 percent, State’s policy at 12.1 percent, family served by a Tribal program at 10.5 percent, failure to cooperate with eligibility requirements at 22.2 percent, sanction at 6.5 percent, and voluntary closure at 2.3 percent. States reported 26.4 percent of cases closing due to other unspecified reasons.

The TANF Adults

Most TANF adult recipients were women. Men represented 10 percent of adult recipients. The average age of TANF adult recipients was 31.3 years. Of TANF adult recipients, 7 percent were teenagers and 18 percent were 40 years of age or older. Nearly 93 percent of adult recipients were the head of the household. There were about 138,000 teen parents whose child was also a member of the TANF family. In other words, 14 percent of teen recipients were teen parents. Only 12 percent of adult recipients were married and living together.

Two of three TANF adult recipients were members of minority races or ethnic groups. Black adults comprised 38 percent of adult recipients. White adults comprised 33 percent of adult recipients and 24 percent were Hispanic or Latino. Of adult recipients, 1.8 percent were American Indian or Alaska Native, and 2.6 percent were Asian. The percentage of Asian adults significantly decreased from 5.0 percent in FY 1999.

There were about 700,000 non-recipient adults in child-only cases. Of all non-recipient adults living in the child-only case household, 55 percent were parents, 38 percent were caretakers and 7 percent were other persons. Of such non-recipient adults who may not be eligible to receive assistance, 30 percent received SSI benefits, 21 percent were in unknown citizenship/alienage including undocumented or non-qualified aliens and only 7 percent were in sanction status. About 42 percent were in unknown reasons for their not being in the assistant unit.

Most TANF adult recipients were U.S. citizens. Non-citizens residing legally in this country were 8 percent of TANF adults. Half of TANF adult recipients completed a formal education of 12 years or more. About 20 percent of adult recipients had less than 10 years of education.

Of TANF adult recipients, 26.4 percent were employed. There was little difference of the employment rate between the male recipients and the female recipients. Employment slightly decreased by about 4 percent when compared with 27.6 percent in FY 1999. A half of TANF adult recipients were in the labor force, i.e., seeking work but not employed, and a quarter of adult recipients were not in the labor force. On the other hand, 19.7 percent of adult recipients were employed in newly-approved families. In closed-case families, 37.9 percent were employed.

Work participation was mandatory for almost three of every five adult recipients. Of TANF adult recipients, about 7 percent were exempt from the work participation because they were single custodial parents with child under 12 months. Nearly 11 percent were exempt because of a sanction, part of an ongoing research evaluation or an approved welfare reform waiver. Ten percent were exempt from the work participation status because of a good cause exception, e.g., disabled, in poor health, or other. About 8 percent were teen parents who were required to participate in education.

Overall, 42 percent of all TANF adult recipients participated in some type of work activities during the reporting month. In other words, 26 percent worked in unsubsidized jobs, 5 percent did job search, and another 14 percent were engaged in subsidized employment, job skills training or work preparation activities. (Some TANF adults did two or three work activities. Also, some adults participated in under the work exemption status.) Their average number of hours per week participating in work activities was 25 hours.

In addition to 37 percent who were exempt from work participation, at least 20 percent of adult recipients did not participate in the required mandatory work activities.

The TANF Children

TANF recipient children averaged about 7.8 years of age. Thirteen percent of recipient children were under 2 years of age, while 39 percent were of preschool age under 6. Only 8 percent of the children were 16 years of age or older.

Most recipient children were children of the head of the household in TANF families, and only 8 percent were grandchildren of the head of the household. Of all TANF recipient children in child-only cases, 62 percent lived with parents and 23 percent with grandparents who did not themselves receive assistance.

The racial distribution of TANF recipient children was slightly changed in recent years. Black children continued to be the largest group of welfare children, comprising about 40 percent of recipient children. About 27 percent of TANF recipient children were white and also were Hispanic or Latino. The percentage of black children on TANF remains unchanged. However, the percentage of white or Hispanic children is up 1.0 percentage point, and down by 1.8 percentage points for Asian children when compared to October 1998 - September 1999.

Financial Circumstances

Of TANF families, 98 percent received cash and cash equivalents assistance with the average monthly amount of $349 under the State’s TANF program. The monthly cash payment to the TANF families was $285 for one child, $356 for two children, $423 for three children and $516 for four children or more. Some TANF families who were not employed received other forms of assistance such as child care, transportation and other supportive services.

One in every four TANF families had non-TANF income. The average monthly amount of non-TANF income was $580 per family. Sixteen percent of the TANF families had earned income with the average monthly amount of $675. Eight percent of the TANF families had unearned income with the average monthly amount of $277. Similarly, 22 percent of newly-approved-families had non-TANF income with the average monthly amount of $495. Of all closed-case families, 38 percent had non-TANF income with the average monthly amount of $832.

A quarter of TANF recipient adults had earned income. Their average monthly earned income increased to $668 from $598 in FY 1999, an increase of about 12 percent. Six percent of adult recipients had unearned income averaging about $291 per month. Three percent of recipient children had unearned income with the average monthly amount of $166.

Ten percent of TANF families received child support with the average monthly amount of $174.

One in 10 TANF families had some cash resources (e.g., cash on hand, bank accounts, or certificates of deposit) with the average amount of $252. Such family’s cash resources were defined by the State for determining eligibility for and/or amount of benefits.

NOTE: The work participation activity data may be somewhat different from those discussed in the work participation rates, because TANF recipient characteristics were prepared using (1) sample cases of 3,300 randomly selected for States who submitted universe data, and (2) the data transmitted by States as of May 23, 2001.

back to top

AFDC/TANF Trends in the 1990’s

Because of the rapid decline in the caseload beginning from a record high of 5.0 million families in 1994 and accelerating after the implementation of PRWORA in 1996 to 2.3 million families in 2000, the question has been raised as to whether the current caseload has changed significantly after PRWORA. For example, it has been suggested that the families now receiving cash assistance are more disadvantaged than prior to welfare reform. In order to address the question, an examination of longer-term trends is helpful in beginning to understand whether, and how, welfare reform has contributed to changes in the characteristics of welfare recipients.

A number of major changes in the characteristics of welfare recipients have occurred in the 1990’s including the number of child-only families, the racial composition of welfare families, the age of adult recipients, the age of the youngest child, and the employment rate of adults. These trends in AFDC/TANF recipient characteristics are presented in Exhibit II.

Child-only Families

In 2000, there were about 782,000 child-only cases, which accounted for 35 percent of the total caseload. The number of child-only families increased steadily throughout the middle 1990’s, reaching a peak of 978,000 such families in 1996. Through 1998 the number of child-only families decreased to 743,000, although their proportion of the caseload continued to increase slowly to 23 percent. However, in 2000 both the number and the proportion increased.

Racial/Ethnic Composition of Families

The racial composition of welfare families has changed substantially over the past ten years. In 1990, it was 38 percent whites, 40 percent blacks and 17 percent Hispanics. In 2000, however, it was 31 percent whites, 39 percent blacks and 25 percent Hispanics. Viewed over the decade there has been a shift from white to Hispanic families which is consistent with broader population trends. This shift has been accelerated since 1996 and is particularly pronounce in California, New York and Texas. Thus, in 2000, 70 percent of all Hispanic welfare families were in three large States (California, New York and Texas), as compared to 65 percent in 1996. In California, the proportion of Hispanic welfare families increased to 45 percent in 2000 from 38 percent in 1996. In addition, black families that had been a declining proportion of the caseload have trended up slightly since 1996. The upshot of these changes is that the proportion of welfare families that were minorities has increased from three-fifths to just over two-thirds over the decade, primarily driven by the growth in Hispanic families.

Age of Adult Recipients

Throughout the decade the average age of adults has gradually, but steadily, increased from 29.7 in 1990 to 31.3 in 2000. Between 1990 and 2000, the proportion of older adults over 39 years increased most dramatically from 13 to 18 percent of adult recipients. Slightly less than half of this growth occurred in the three years after TANF compared to the six years before.

Employment Rate

The employment rate of adult recipients has been increased significantly in the past five years. In 2000, 26 percent of adult recipients were employed, about 2.4 times the 1996 employment rate of 11 percent and four times the rate of the early 1990’s.

Age of the Youngest Child

Between 1990 and 2000 the proportion of families with a youngest child who was a toddler, i.e., aged 1 or 2, declined sharply from 30 to 20 percent. At the same time the proportion of families with a youngest child with no pre-schooler, i.e., no child under 6 years old, increased sharply from 36 to 45 percent. Furthermore, this trend was accelerated after 1996 with 63 percent of the growth occurring in the later three years.

Although the question of whether the caseload has become more disadvantaged cannot be answered simply through the administrative data provided by the States and will require analysis of data from national data sets as it becomes available, the trends identified above are certainly consistent with a more disadvantaged caseload. For example, older recipients who have older children comprise a larger share of the caseload, and that is consistent, but does not prove, that these may comprise longer-term recipients who are having a harder transition to independence. Similarly, the larger proportion of minorities, especially Hispanics who may have a higher probability of speaking English as a second language could also partly account for the sharper increase in minorities after 1996. However, it is important to note that these trends also represent longer-term trends. In addition, a number of studies suggest that minorities be helped into work as much as non-minorities by the kinds of welfare-to-work strategies that most States are employing. Longer-term observation of these trends along with examination of other national data when it becomes available will thus be necessary to better understand if the caseload is becoming more disadvantaged.

back to top

LIST OF TABLES

back to top
Last Updated on 06/06/2002
By ACF

APPENDIX II

Reliability of Estimates

All States, the District of Columbia, Guam, Puerto Rico and Virgin Islands were required to submit TANF data on the demographic characteristics and financial circumstances of families receiving assistance and families no longer receiving assistance under their TANF program. During FY 2000, all 54 States and Territories transmitted 8,391,790 active cases and 649,794 closed cases onto the national TANF database. Under the TANF data reporting system, States have the option to submit either sample data or universe data to HHS. Twenty-four States submitted sample data. The remaining 30 States submitted universe data, from which HHS randomly selected approximately 275 active cases and 100 closed cases each month. A total probability sample of 198,415 active cases and 54,340 closed cases was used in the TANF recipient characteristics study for FY 2000.

The statistical data are estimates derived from samples and, therefore, are subject to sampling errors as well as nonsampling errors. Sampling errors occur to the extent that the results would have been different if obtained from a complete enumeration of all cases. Nonsampling errors are errors in response or coding of responses and nonresponse errors or incomplete sample frames.

Standard (Sampling) Errors

For FY 2000, the average monthly caseload, annual sample sizes, average monthly sample sizes, sampling fractions and the percentage points by which estimates of the total caseload for each State might vary from the true value at the 95 percent confidence level are shown in Table 1 and 2.

Table 3 indicates the approximate standard error for various percentages for the U.S. total caseload. These standard errors are somewhat overstated because they are calculated assuming a sample of 14,855 cases out of a total of 2,269,131 cases or 0.65466546 percent of the average monthly caseload. California is the State with such a small sampling fraction. To obtain the 95 percent confidence level at each percent in Table 3, multiply the standard error by a factor of 1.96.

For example, national estimates of 50 percent should not vary from the true value by more than plus or minus 0.8036 percentage points (0.41 x 1.96) at the 95 percent confidence level. To obtain the 99 percent confidence level, multiply the standard errors by a factor of 2.58.

Nonsampling Errors

Every effort is made to assure that a list of the universe or the sample frame is complete. It is possible, however, that some cases receiving assistance for the reporting month are not included. There is no measure of the completeness of the universe.

Data entries are based on information in the case records. Errors may have occurred because of misinterpretation of questions and because of incomplete case record information. Errors may also have occurred in coding and transmitting the data. There are no measures of the reliability of the coded information. For some data elements, obviously incorrect or missing information was recoded as unknown in the data processing.

Standard Errors of Subsets

For tables based on subsets of the populations, e.g., one-adult or two-adult families, the approximate standard errors can be computed by the following method: (a) determine the assumed sample size of the subset by multiplying the number of cases in the subset by 0.0065466546; (b) divide the sample size of all families (14,855) by the assumed sample size of the subset; and (c) take the square root of the result and multiply it by the standard errors of the total caseload shown in Table 3.

For example, for child-only cases the approximate standard errors of percentages can be found by multiplying the data in Table 3 by the square root of 14,855/5,117 or 2.9031. The sample size of 5,117 is determined by 781,546 x 0.0065466546.

Standard Errors for State Estimates

The method used above can be adapted to calculating standard errors of State estimates. First, divide the national sample size of all families (14,855) by the State sample size shown in Table 1. Then take the square root of the result and multiply it by the standard errors shown in Table 3. For example, for Florida the approximate standard errors of percentages can be found by multiplying the data in Table 3 by the square root of 14,855/3,275 or 2.1298.

Statistically Significant Differences

Table 4 shows the percentage values at which differences between national and State estimates become significant at the 5 percent confidence level based on annual State samples of 3,000 active cases.

Table 5 shows the percentage values at which differences between State estimates become significant at the 5 percent confidence level based on annual State samples of 3,000 active cases.

back to top

back to top

back to top