How to improve the welfare system has been the subject of intense debate in recent years and many States are modifying their programs substantially under waivers granted by Federal Government. These changes and proposed ones have intensified the interest in information on people who participate in welfare programs. This report uses data from the Survey of Income and Program Participation (SIPP) to examine who receives assistance from the major means-tested government programs--namely, Aid to Families with Dependent Children (AFDC), General Assistance, food stamps, Supplemental Security Income (SSI), Medicaid, and housing assistance--over the 28-month period from October 1991 through January 1994.
Because SIPP provides monthly information on the program participation of individuals, as well as on many demographic and socioeconomic characteristics that can vary over time such as family and labor force status, differences in patterns of participation can be analyzed2. Specifically, this report examines similarities and differences in: (1) average monthly program participation in 1993; (2) the percent of people who participated in at least one of these programs at some time during the 1992-1993 period; (3) the percent who participated in at least one program in all 24 months of 1992 and 1993; and (4) the length of time participants stayed in the programs (the duration of the spell).
Of the estimated 258 million civilians living in the United States, approximately 36 million or 14.0 percent participated in one or more of the major means-tested assistance programs. As shown in Figure 1, the average monthly program participation rate has increased noticeably, from 11.4 percent in 1987 to 14.0 percent in 19935.
A substantial proportion of the recipients, however, participated in major government programs only on a short-term basis. Only 8.6 percent of persons participated in these programs all 24 months of the 1992-1993 period. These long-term recipients were likely to be children or at least 65 years old. The proportions of children and the elderly that participated in these means-tested programs in each month of 1992 and 1993 were 14.1 percent and 9.7 percent, respectively,compared with 6.0 percent of people who were 18 to 64 years old.
Generally, program participation rates are related to poverty and business cycles--rising along with poverty rates during periods of economic contraction, and both falling during periods of economic expansion. During the expansionary period of 1987 and 1989, the official poverty rate dropped slightly from 13.4 percent to 12.8 percent. The poverty rate then rose to 13.5 percent in 1990 and reached 15.1 percent in 1993--a period when participation rates rose from 11.5 percent to 14 percent6.
As shown in table A and figure 2, individuals were more likely to participate in Medicaid than in any other program. In 1993, the average monthly participation rate for Medicaid, 10.3 percent, was higher than that of food stamps, AFDC or General Assistance, housing assistance, or SSI. A similar pattern existed for persons who were long-term participants, that is, who participated in these programs all 24 months of the 1992-1993 period7.
However, while a higher proportion participated in Medicaid, the length of stay on each of these programs was similar. Specifically, the median durations of participation for Medicaid, food stamps, and AFDC were not significantly different from one another (see table B). The median spells for SSI and housing assistance were not available for analysis, because more than half of their spells continued in the last month of data collection. This situation is especially likely to occur for elderly recipients whose incomes are likely to be stable over time.
In 1993, the average monthly number of Whites receiving means-tested assistance was far greater than of Blacks, 22.9 million compared with 11.6 million. However, Blacks and Hispanics had higher average participation rates than Whites and non-Hispanics, respectively, both overall and for the individual programs as well. More than one-third (35.5 percent) of Blacks participated in these means-tested programs, compared with only 10.6 percent of Whites (see table A). The proportion of Hispanics who received benefits was 28.9 percent, significantly higher than the 12.3 percent of non-Hispanics who participated.
Additionally, in the 1992-1993 period, the median number of months Blacks received benefits was larger than for Whites (see table B). However, the median duration for Hispanics was not significantly different from the medians for non-Hispanics, Whites, and Blacks8.
Blacks tended to receive higher monthly benefits than Whites, a reflection of their relatively lower incomes and larger families. As indicated in table A, the median monthly benefit for Black families ($526) was significantly higher than the median for White families ($399) in 1993, whereas the median benefit of Hispanics was not significantly different from that of non-Hispanics.
Program participation is closely associated with the age of an individual, as shown in figure 3. In 1993, nearly 1 in 4 (23.7 percent) children younger than 18 received some type of means-tested assistance, compared with only 1 in 10 (10.0 percent) persons aged 18 to 64 years and 1 in 8 (12.0 percent) of the elderly.
For people age 18 and over, lower educational attainment is associated with greater program participation (see table A). In 1993, about 1 in 4 (23.6 percent) of those with less than four years of high school received means-tested benefits, compared with 1 in 10 (10.1 percent) of those who completed high school but did not attend college, and only 4.1 percent of those with at least 1 year of college.
As expected, individuals who had not graduated from high school stayed in these programs longer than those with more education. The median duration of receipt for those without a high school degree (11.7 months) was higher than the medians for high school graduates (7.7 months) and persons with some college experience (7.4 months)9.
In 1993, on average, 25.2 percent of people 15 to 69 years old with a work disability received means-tested benefits, compared with only 7.8 percent of those with no work disability. Although SSI was designed for the disabled and automatically confers eligibility for Medicaid, people with work disabilities were also more likely than others without work disabilities to participate in General Assistance and housing assistance10.
When we examine program participation over time, those with disabilities were more likely to be long-term recipients than others--19.2 percent of them collected benefits in all 24 months, compared with 4.0 percent of those without work disabilities.
Not surprisingly, 57.3 percent of the poor received means-tested benefits in 1993, compared with 6.5 percent of the non-poor (see figure 4). Over half (53.5 percent) of the poor participated in all 24 months of 1992 and 1993, as did only 3.0 percent of the non-poor. In addition, the median duration of receipt for the poor was about twice that of the non-poor (11.5 months compared with 6.0 months)11.
Reflecting their relatively low incomes, individuals in families maintained by women were much more likely to participate in means-tested programs than those in married-couple families--42.9 percent compared with 7.7 percent in 1993. Moreover, over half (51.1 percent) of those in families maintained by women participated in means-tested programs during at least 1 month of 1992 and 1993, compared with 13.8 percent of those in married-couple families. Similarly, a higher proportion of families maintained by women received means-tested benefits in all 24 months than of married-couple families.
As shown in figure 5, in an average month of 1993, almost one-quarter of the unemployed received means-tested assistance, as did one in five individuals not in the labor force. In contrast, only 3.5 percent of those with full-time jobs and 8.6 percent of those with part-time jobs received these benefits at some time during 1993.
The unemployed and those who were not in the labor force received higher monthly benefits than people with full- or part-time jobs. The median monthly benefit from AFDC, General Assistance, food stamps, and SSI for families with an unemployed worker was $411, not significantly different from that of families with no members who were in the labor force. The median benefits for full-time and part-time workers were $231 and $282, respectively. In addition, the median unemployment compensation benefit among recipients was $618 in 1993, although only 28.7 percent of the unemployed received these benefits.
As shown in table A, people living outside metropolitan areas had a higher average program participation rate in 1993 than those living inside metropolitan areas. However, dividing metropolitan areas into central cities and non-central cities reveals a different picture, with those in central cities having participation rates significantly higher than either of the other two groups.
Regionally, the average monthly program participation rate was highest in the South (15.5 percent) and lowest in the Midwest (11.9 percent). The average participation rate in the West was similar to those of the Northeast and the South, but the average rates in the Northeast and the South were significantly different. However, disaggregation by type of program shows that people in the West were more likely to be covered by Medicaid than those in the South or the Midwest. People in the South were more likely to be covered by food stamps than those in any other region, but less likely to receive AFDC12.
All statistics are subject to sampling error, as well as nonsampling error such survey design flaws, respondent classification and reporting errors, data processing mistakes, and undercoverage. The Census Bureau has taken steps to minimize errors in the form of quality control and edit procedures to reduce errors made by respondents, coders, and interviewers. Ratio estimation to independent age-race-sex population controls partially corrects for bias attributable to survey under coverage. However, biases exist in the estimates when missed persons have characteristics different from those of interviewed persons in the same age-race-sex group.
Analytical statements in this report have been tested and meet statistical standards. However, because of methodological differences, use caution when comparing these data with data from other sources.
Contact Elaine Hock, Demographic Statistical Methods Division, at 301-763-4182 or on the internet at email@example.com for information on (1) the source of the data, (2) the accuracy of the estimates, (3) the use of standard errors, and (4) the computation of standard errors.
Note: All demographic surveys, including SIPP, are affected by under coverage of the population. This under coverage results from missed housing units and missed persons within sample households. Compared with the level of the 1980 decennial census, overall under coverage in SIPP is about 7 percent. Under coverage varies with age, sex, and race. For some groups, such as 20 to 24 year old Black men, the under coverage is as high as 27 percent compared with the census. It is important to note that the survey under coverage is in addition to the decennial census under coverage, which in 1980 was estimated to be about 1 percent overall and about 8.5 percent for Black men. The weighting procedures used by the Census Bureau partially correct for the bias due to under coverage, but it's final impact on estimates is not known.
The Census Bureau welcomes the comments and advice of data users. If you have suggestions or comments, please write to:Daniel H. Weinberg