Civic Report No. 17 July 2001Gaining Ground? Measuring the Impact of Welfare Reform on Welfare and Work
Appendix B:Data Issues: Caseload Data Versus The CPS
In this study we use the Current Population Survey (CPS) as our primary data source. Other studies have relied on state administrative data based on records of the welfare caseload. How do these different sources compare? We find that trends in the number of single mothers receiving welfare benefits are similar under different definitions and data sources. Figure B-1 displays these trends for the entire Basic family caseload and for the subset of Basic cases after the “child-only” cases have been subtracted. (The Basic caseload includes families with a deceased, disabled or absent father and child-only cases; the only group excluded is two-parent families with an able-bodied but unemployed father present.) Both series are caseload data from the administrative records of the individual states. The child-only cases refer to those in which only a child or children are receiving benefits, and such cases are exempt from the federal regulations concerning work requirements or time limits. Child-only cases can be created when a parent is present but does not qualify for TANF benefits because of alien status, a sanction, or receipt of SSI benefits. In other child-only cases the parent is not present and the child is under the care of a grandparent or other adult. The number of child-only cases increased throughout the 1990s and in 1999 accounted for about 30 percent of the Basic caseload.
Figure B-1 also compares the two series of caseload data with the estimated number of single mothers on welfare, ages 18–44, as counted in the March CPS. Although the patterns are quite similar, the caseload data indicate a larger count of female family heads on welfare than the CPS, even when the comparison is with the series omitting the child-only cases (which is not only closer for the CPS data but also more nearly parallel). In part this is because the caseload data include a wider range of ages of family heads than the 18–44-year-olds in the CPS data included in Figure 7. It is also possible that the caseload data double count some beneficiaries who have moved from one state to another or who may have gone off and then gone back on welfare within a month in the same state. On the other hand the CPS is believed to undercount welfare recipients, particularly those who were on welfare for a short time during the year. We use CPS data for most of the analysis in this report because it contains information about both the general population and the welfare population and because it provides the necessary detail on the characteristics of each individual.
[figure 1] [figure 2] [figure 3] [figure 4] [figure 5] [figure 6][figure 7] [figure 8] [figure 9] [figure 10] [figure 11] [figure 12] [table 1] [table 2] [table 3] [table 4] [table 5] [table 6] [table A-1] [table A-2] [appendix B] [figure B-1] [appendix C] [table C-1] [table C-2]