Prior to the passage of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996, also known as the welfare reform bill, states modified their welfare programs under waivers granted by the federal government. These waivers allowed states to experiment with different welfare policies. For example, many states adopted time limits and job training programs to reduce welfare dependence and to encourage work. Changes in the welfare system, both under waivers and the welfare reform bill, have intensified the interest in information on the characteristics of people who participate in welfare programs.
This report examines the participation in major means-tested government programs in the years just before federal welfare reform.(1) For the pre-reform era, the report provides a set of baseline estimates of participation for these programs:
The data in this report come from the 1993 panel of the Survey of Income and Program Participation (SIPP), which covers the period from October 1992 to December 1995.(2) SIPP is a longitudinal survey that follows the same individuals over time.(3) The longitudinal nature of SIPP enables a dynamic analysis of the rates of program participation and of the amounts of benefits received among people of different demographic and socioeconomic characteristics.
Specifically, this report examines the similarities and differences among various groups in:
Program Participation: 1993 to 1994
Of the estimated 261 million civilians living in the United States, approximately 40 million or 15.2 percent participated in one or more major means-tested assistance programs, on average, during each month of 1994. As can be seen in Figure 1 , the average monthly participation rate has increased noticeably from about 11 percent in the 1987-1990 period to 15.2 percent in 1993 and 1994.(5)
Only a small proportion of the population, however, participated on a long-term basis, with about 10 percent of the population having participated in each month of the 1993-1994 period. About 16.5 percent of those under 18 years old participated each month of the 1993-1994 period, a figure that is statistically higher than the comparable proportion of 18 to 64 year old recipients, 6.9 percent, and higher than that of elderly recipients, 10.3 percent.(6)
Medicaid Has the Highest Participation Rate
As shown in Table A and Figure 2, individuals were more likely to participate in Medicaid than in any of the other programs examined here. In 1994, the average monthly participation rate for Medicaid was 11.3 percent, higher than that for AFDC, GA, food stamps, housing assistance, or SSI. A larger proportion of the population, about 6.6 percent, participated in all 24 months in Medicaid than in any other program.(7)
An estimated 29 million people received Medicaid benefits in 1994; almost 16 million of those recipients were children.(8) In fact, 22.1 percent of children under age 18 received Medicaid, compared with 7.1 percent of people 18 to 64 years old, and 8.0 percent of people over 65 years old.
Over Half of the Poor Receive Means-Tested Assistance
Figure 3 shows that 60.3 percent of the poor, those with family incomes under the poverty thresholds, received at least one type of major means-tested benefit in 1994, compared with 7.0 percent of the nonpoor.(9) Additionally, 75.1 percent of the poor received benefits during at least 1 month of 1994, compared with 10.5 percent of the nonpoor. The poor also tended to be long-term participants in means-tested programs: 57.9 percent of the poor, but only 3.4 percent of the nonpoor, participated in all 24 months of 1993-1994.
Program Participation Varies by Demographic Group
The likelihood of receiving means-tested assistance varied among race and ethnic groups. In 1994, the average number of Whites receiving assistance per month, 26 million, was far greater than the number of Blacks, 12 million. However, the average monthly participation rate was higher for Blacks (36.0 percent) than for Whites (11.8 percent), as shown in Figure 4 . In addition, Blacks were more likely than Whites to participate in each month of the period 1993-1994: 27.0 percent of Blacks, compared with 6.9 percent of Whites.
There was also a strong association between Hispanic origin(10) and the likelihood of receiving means-tested assistance. In fact, nearly 9 million people of Hispanic origin received means-tested assistance in 1994, compared with almost 31 million people not of Hispanic origin. However, during the same year, the average monthly participation rate for people of Hispanic origin was 31.7 percent, compared with 13.2 percent for those not of Hispanic origin. Hispanics were also more likely to receive means-tested assistance in each month of the 1993-1994 period: 22.5 percent of Hispanics, compared with 8.6 percent for those not of Hispanic origin. Since poverty and participation in major means-tested assistance programs are closely related (see Figure 3), the differences among the racial and Hispanic-origin groups in participation rates can, in part, be explained by the differences in poverty rates. In 1994, the average monthly poverty rate was 12.7 percent for Whites and 31.2 percent for Blacks.(11) Likewise, Hispanics had an average monthly poverty rate of 31.4 percent, compared with 13.5 percent for non-Hispanics (see Figure 5).(12)
Children Under 18 Years Old Are More Likely to Receive Means-Tested Assistance Than People in Other Age Groups
Children under 18 years of age were more than twice as likely as people 18 to 64 years old to receive means-tested benefits, according to Figure 6. In an average month during 1994, 19 million (26.5 percent) children received some type of means-tested benefit, compared with 17 million (10.8 percent) people age 18 to 64 years old and 4 million (11.7 percent) people 65 years and older.(13) Children also tended to be long-term participants, with 11 million (16.5 percent) collecting benefits in all 24 months of the period 1993-1994, compared with 11 million (6.9 percent) and 3 million (10.3 percent) for people ages 18 to 64 years old and 65 years and older, respectively.
Families Maintained by Women Have Higher Participation Rates
Reflecting their higher poverty rates and lower incomes, individuals in families maintained by women were much more likely to participate in major means-tested programs, in an average month of 1994, than were those in married-couple families--44.3 percent compared with 8.9 percent.(14) During 1994, households maintained by women had an average monthly poverty rate of 39.8 percent and an annual median income of $19,872, compared with an average monthly poverty rate of 8.4 percent and an annual median income of $45,041 for married-couple familes. Similarly, approximately half (50.1 percent) of individuals in families maintained by women participated in means-tested programs during at least one month of 1994, while 12.3 percent of individuals in married-couple families participated in at least one month during that period. Moreover, individuals in families maintained by women were more than five times as likely as individuals in married-couple families to receive benefits in all 24 months of the period 1993-1994--32.2 percent compared with 5.2 percent.
Women were more likely than men to receive means-tested benefits. In 1994, 23 million (17.3 percent) women participated in an average month, compared with nearly 17 million (13.0 percent) men. Women were also more likely than men to receive means-tested benefits in each month of the 1993-1994 period, 11.6 percent compared with 8.2 percent. (see Figure 7)
Recipients Have Lower Educational Levels
For people age 18 and over, lower educational attainment was associated with greater program participation (see Table A and Figure 8). In 1994, about one in four (25.6 percent) of those with less than 4 years of high school received means-tested benefits, compared with 1 in 10 (10.5 percent) for high school graduates and about 1 in 25 (4.5 percent) for those with at least one year of college. Individuals who did not graduate from high school also remained on means-tested programs longer than high school graduates. The median duration of receipt for those without a high school diploma was 11.0 months, compared with 7.2 months for high school graduates, and 7.1 months for those with at least 1 year of college (see Table B).(15)
The Unemployed and Those Out of the Labor Force Are More Likely to Receive Means-Tested Benefits Than the Employed
Unemployed people were much more likely to receive means-tested benefits in an average month of 1994 than were people with full-time jobs. For people 18 years and older, nearly 27 percent of the unemployed received means-tested benefits in an average month of 1994, compared with 21.3 percent of those out of the labor force, 3.8 percent of those employed with full-time jobs, and 9.2 percent of those employed with part-time jobs (see Figure 9).
The unemployed may receive unemployment compensation in addition to major means-tested benefits. In an average month of 1994, only 19.3 percent of the unemployed received unemployment compensation, while 11.3 percent received AFDC or GA, 17.0 percent received Medicaid, 1.5 percent received SSI, and 20.2 percent received food stamps.(16)
Accuracy and Reliability of the Data
Statistics from sample surveys are subject to sampling and nonsampling error. All comparisons presented in this report have taken sampling error into account and meet the Census Bureau's standards for statistical significance. Nonsampling errors in surveys may be attributed to a variety of sources, such as how the survey was designed, how respondents interpret questions, how able and willing respondents are to provide correct answers, and how accurately answers are coded and classified. The Census Bureau employs quality control procedures throughout the production process--including the overall design of surveys, testing the wording of questions, reviewing of the work of interviewers and coders, and conducting statistical review of reports.
The SIPP employs ratio estimation, whereby sample estimates are adjusted to independent estimates of the national population by age, race, sex, and Hispanic origin. This weighting partially corrects for bias due to undercoverage, but how it affects different variables in the survey is not precisely known. Moreover, biases may also be present when people who are missed in the survey differ from those interviewed in ways other than the categories used in weighting (age, race, sex, and Hispanic origin). All of these considerations affect comparisons across different surveys or data sources.
For further information on statistical standards and the computation and use of standard errors, contact Mark Gorsak, Demographic Statistical Methods Division, at 301-763-4228 or on the Internet at Mark.Gorsak@ccmail.census.gov.
Comments from Data Users
The Census Bureau welcomes the comments and advice of data users. If you have suggestions or comments, please write to:Daniel Weinberg
1. Means-tested programs are those that require income and/or assets of the individual or family to be below specified thresholds in order to qualify for benefits. These programs provide cash and non-cash assistance to eligible individuals and families.
2. For each person on the 1993 nine-wave longitudinal file, data are available for up to 36 continuous reference months; the exact number of months depends upon when the person entered or exited the sample. The time period of calendar months covered by these reference months depends on the person's rotation group. Data for all four rotation groups (the full sample) are available only for the calendar months of January 1993 through September 1995. The Census Bureau will continue to follow the families who participated in the 1992 and 1993 SIPP panels in the Survey of Program Dynamics to provide post-reform longitudinal data.
3. Efforts were made during the life of the panel to follow people who moved to ensure that the sample remained representative of the noninstitutional population of the United States.
4. The estimate for 1993 is 39 (±1.3) million and for 1994 is 40 (±1.3) million. These estimates are not statistically different.
5. SIPP average program participation rates from 1987 to 1992 were obtained from previous Census Bureau reports, specifically, Current Population Reports, Household Economic Studies, Series P70-31, P70-41, P70-46, and P70-58. The program participation rate for 1989 is not available. There is no statistical difference between the average monthly participation rates for 1987 and 1990.
6. There is no statistical difference between the proportion of people 18 to 64 years old and the proportion of the elderly participating.
7. There is no statistical difference between the average monthly participation rate for AFDC and housing assistance.
8. There is no statistical difference between the percentage of people 18 to 64 years old who received Medicaid and the percentage of the elderly who received Medicaid.
9. The poverty threshold for a family of three with one related child was $11,929 in 1994. Data on poverty thresholds by family size and number of related children under 18 years for 1994 can be obtained from the Current Population Reports, Consumer Income, P60-189.
10. Hispanics may be of any race. The information on the Hispanic population shown in this report was collected in the 50 states and the District of Columbia, and therefore, does not include residents of Puerto Rico.
11. The information on poverty rates using SIPP can be found in Current Population Reports, Household Economics Studies, Series P70-63.
12. There is no statistical difference between the average monthly participation rate and the average monthly poverty rate for Hispanics. There is no statistical difference between the average monthly participation rate and the average monthly poverty rate for non-Hispanics. There is no statistical difference between the average monthly poverty rate for Blacks and Hispanics.
13. There is no statistical difference between the percent of people age 18 to 64 and the percent of people 65 years and older who received means-tested benefits.
14. There is no statistical difference between the average monthly participation rate and poverty rate for married-couple families.
15. There is no statistical difference between the median spell duration of high school graduates and the median spell duration of people with some college.
16. There is no statistical difference between the percentage of the unemployed who received Medicaid and the percentage of unemployed who received food stamps.