Final Synthesis Report of Findings from ASPE "Leavers" Grants. Program Participation by Subgroup

11/27/2001

Individual reports find differences in results when comparing across subgroups such as employed and not employed or sanctioned and not sanctioned. Two types of subgroup comparisons for program participation were included in enough reports to be discussed here employment status and race/ethnicity.

Examining receipt of government assistance by employment status provides information on the extent to which those who are employed are combining work with some form of public assistance. Generally, one would expect that those who are working are less likely to be receiving government benefits. It is also interesting to examine whether working leavers benefit receipt declines over time, which would be consistent with increasing wages and more stable employment. All the figures described in this section with the exception of South Carolina are calculated from the public use data files for these studies. This allows the report to present program participation by employment status when it is not reported in the published studies.

Program participation by employment status is examined both in administrative data and survey data. Public-use data files including administrative data on food stamps and Medicaid have been provided by two states, Arizona and Iowa. For both of these studies we compare those who have worked at some point since exit and those who have never worked since exit (Table IV.11). Across both these states and programs, 44 to 63 percent of leavers who have worked since exit receive food stamps or Medicaid in any given quarter after exit. A higher percentage, almost three-quarters, receive food stamps or Medicaid at some point during the year after exit. These higher figures indicate that among working leavers there is some cycling on and off non-TANF benefit programs. Because we are examining the group who worked at any point since exit (not necessarily continually), this benefit receipt could be connected to movement on and off TANF.

Table IV.11:
Percent of Single-Parent Leavers Receiving Government Benefits by Employment Status: Administrative Data

Program

State/Study Exit Cohort Quarter Relative to Exit (%) Receipt Any Time in Year After Exit
Q1 Q2 Q3 Q4

Food Stamps

 

Arizona1

1Q98  
 

Worked Since Exit

  55 49 48 44 71
 

Never Worked Since Exit

  44 39 39 37 57
 

Iowa1

2Q99  
 

Worked Since Exit

  53 47 44 60 70
 

Never Worked Since Exit

  45 39 37 48 58

Medicaid

 

Arizona1

1Q98  
 

Worked Since Exit

  63 58 52 48 76
 

Never Worked Since Exit

  51 45 44 39 63
 

Iowa1

2Q99  
 

Worked Since Exit

  57 52 51 49 72
 

Never Worked Since Exit

  48 41 43 40 58

1Quarterly data calculated from public use files.
Source: See Appendix B for a complete listing of the leavers studies referenced.

The administrative data show that those who have never worked since exit are less likely to receive food stamps or Medicaid than those who have worked since exit. This is true in all quarters for Arizona and Iowa for both the food stamp and Medicaid programs. One possible explanation for this difference is that those who have never worked since exit are a heterogeneous group including leavers who have a working spouse or partner and may not qualify for benefits, as well as leavers who have few sources of income. Another possible explanation is that some leavers who do not show up as having earnings or food stamps in administrative databases have moved out of state or may have misspelled last names or missing social security numbers which hamper linkages to both earnings and food stamps databases. In fact, survey data discussed below show different results.

The administrative data can be used to examine the trend in participation over the four quarters for these subgroups. For those who worked at some point since exit and those who never worked since exit, both studies show a gradual decline in benefit receipt over the first three quarters. In Iowa, however, food stamp participation turns up sharply in the fourth post-exit quarter.

Another way of considering program participation by employment status is to examine current benefit receipt among those who are currently working or currently not working. Some of the studies with survey data allow us this comparison to be made. Six of the studies we information by employment status on receipt of food stamps, Medicaid, and SSI (Table IV.12). Program participation for food stamps and Medicaid is almost always higher for those not currently employed compared to those currently employed. The only exception is Massachusetts, where 80 percent of those not currently employed have Medicaid compared with 83 percent of those currently employed. While many of the differences are modest, in the District of Columbia the difference in benefit receipt by current employment status is quite large. Figure IV.6 illustrates the differences in food stamp receipt between leavers who are currently employed and not currently employed.

Table IV.12:
Percent of Single-Parent Leavers Receiving Government Benefits at Time of the Survey by Employment Status: Survey Data1

Program

State/Study Exit Cohort Timing of Survey Post Exit Currently Employed Not Currently Employed Never Worked Since Exit

Food Stamps

Arizona

1Q98 12-18 months 30 38 40
 

District of Columbia2

4Q98 ~ 12 months 26 69 59
 

Iowa3,4

2Q99 8 - 12 months 36 41 49
 

Massachusetts2

Dec.1998-Mar.1999 ~10 months 33 36 55
 

South Carolina2

Oct.1998-Mar.1999 12 months 49 72 n.a.
 

Washington4

Oct-98 6-8 months 44 50 56

Medicaid

Arizona

1Q98 12-18 months 27 29 36
 

District of Columbia2

4Q98 ~ 12 months 41 72 75
 

Iowa2,3

2Q99 8 - 12 months 37 46 52
 

Massachusetts2

Dec.1998-Mar.1999 ~10 months 83 80 92
 

South Carolina2

Oct.1998-Mar.1999 12 months 82 83 n.a.
 

Washington

Oct-98 6-8 months 52 56 57

SSI

Arizona

1Q98 12-18 months 11 10 18
 

District of Columbia2

4Q98 ~ 12 months 4 2 16
 

Iowa2,3

2Q99 8 - 12 months 3 4 17
 

Massachusetts2,5

Dec.1998-Mar.1999 ~10 months 17 19 27
 

South Carolina2,6

Oct.1998-Mar.1999 12 months 6 15 n.a.
 

Washington

Oct-98 6-8 months 3 3 8

1All figures except South Carolina calculated from public use data files.
2Results are for all cases; not just single-parent cases.
3Month prior to survey.
4Figures are for time since exit from welfare.
5Also includes SSDI and Social Security.
6Includes adult or child in household receiving.
Source: See Appendix B for a complete listing of the leavers studies referenced.

Figure IV.6:
Food Stamp Receipt of Welfare Leavers by Employment Status

Figure IV.6: Food Stamp Receipt of Welfare Leavers by Employment Status

Notes: See table IV.12 for more information.

Table IV.12 also shows results for those who have never worked since exit. Leavers who never worked since exit are more likely to receive food stamps and Medicaid than currently employed leavers. In several studies these differences are quite large. In addition, in almost all of the studies, those who have never worked since exit are more likely to receive benefits than those who are not currently employed. This suggests that those who have never worked since exit are more dependent on government benefits than leavers who have worked at some point since exit. These survey results might differ from the administrative results because they measure receipt in a specific month, versus receipt at any time over a quarter. If there is a great deal of movement on and off benefit programs, as is likely more true for those working at any point since exit than those never working since exit, receipt at some point during a quarter will be higher than receipt in a given month27

Finally, examination of SSI receipt by employment status shows that a relatively high percentage of those who have never worked since exit are receiving this form of cash assistance. The percentages range from 8 percent in Washington to 27 percent in Massachusetts (which includes SSDI and Social Security benefits). That these numbers are relatively high for those who have never worked is not surprising since SSI requires that recipients have a disability that prevents work. However, the majority of former recipients that have not worked since exit are not receiving disability benefits.

Five studies report some information on the participation of leavers in TANF, food stamps, or Medicaid in the year after exit by race and ethnicity (Table IV.13). For the most part, the share of black leavers receiving these forms of government assistance in the year after exit is higher than the share of white and other racial/ethnic group leavers.

Table IV.13:
Program Participation of Single-Parent Welfare Leavers by Race/Ethnicity: Administrative Data Findings
  Race/Ethnicity Group
  Black White Hispanic Other1

Returns to TANF in Year After Exit (%)

Arizona2

33 21 30 40

Florida

32 22 22 16

Georgia

23 13 n.a. 10

Missouri3,4

32 27 n.a. n.a.

South Carolina5

8 6 n.a. n.a.

Receipt of Food Stamps in Year After Exit (%)

Arizona2

70 62 71 69

Florida

73 63 55 50

Missouri3,4

72 69 n.a. n.a.

Adult Covered by Medicaid in Year After Exit (%)

Arizona2

75 68 77 70

Florida

78 69 70 63

Missouri3,4

37 44 n.a. n.a.

1Other includes Asain/Pacific Islander, Native American, and other groups unless otherwise specified.
2"Other" category represents Native American. A small percent of caseload (less than 1%) is other race/ethnicity groups not included in this table.
3Category "black" refers to nonwhite and includes Hispanic, Native American and other leavers who comprise less than 2 percent of Missouri's leavers.
4Results are for all cases; not just single-parent cases.
5Results are from survey data.

In all five studies, the percentage of black leavers who return to TANF at some point in the year after exit is higher than the percentage of white leavers who return to TANF. The difference ranges from 2 percentage points in South Carolina (8 percent of black leavers compared with 6 percent of white leavers) to 12 percentage points in

Arizona (33 percent of black leavers compared with 21 percent of white leavers). The two studies reporting results for Hispanic leavers, Arizona and Florida, show mixed results. In Arizona, returns to welfare among Hispanic leavers (30 percent) are closer to those of black leavers than to those of white leavers. Returns for Native American leavers are even higher, at 40 percent. However, the percentage of Hispanic leavers returning to TANF in Florida (22 percent) is the same as for white leavers. The other category in Florida and Georgia show very low returns to TANF, 16 and 10 percent respectively, but these groups are very small percentages of the entire group of leavers in each study.

Leaver subgroups with higher returns to TANF tend to have higher participation in food stamps and Medicaid. The percentage of black leavers receiving food stamps and the percentage covered by Medicaid in the year after exit are for the most part higher than the percentage of white leavers receiving these benefits. The differences in receipt between these two groups ranges from 3 to 10 percentage points. The exception is in Missouri, where fewer nonwhite leavers have Medicaid coverage than white leavers. The pattern for Hispanic leavers follows the pattern of returns to TANF: in Florida, fewer Hispanic leavers receive these benefits than blacks, and in Arizona, more Hispanic leavers receive these benefits than blacks.

It is not surprising that receipt of food stamps and Medicaid is generally higher for racial and ethnic leaver groups that have higher rates of return to TANF. Receipt of these benefits is likely easier for those who are currently receiving TANF than for former recipients. However, it is somewhat surprising that black leavers have a higher rate of return to TANF than white leavers in light of results described in an earlier chapter showing black leavers tend to have higher employment and earnings than white leavers.28 This may reflect differences in the reason for leaving across groups. If black recipients are more likely to leave for employment than white recipients and employment leads to less permanent transitions than other reasons (such as marriage) then employment rates and returns to TANF could be higher for black leavers. We do not have evidence on the differences across race and ethnic groups in the reasons for leaving TANF.