Child support can represent an important income source for many low-income families, and the receipt of support may be most critical for women as they transition off welfare. In fact, in the post-welfare reform era emphasizing work and self-sufficiency, child support is seen by many policymakers as a key income support.

Administrators at all levels of government are increasingly recognizing the importance of collaboration between various social service agencies, and the agencies administering welfare (Temporary Assistance for Needy Families, or TANF) and child support are no exception. In the effort to support self-sufficiency, for example, it is important to make sure that women receive all child support to which they may be entitled while receiving benefits but also after leaving welfare, and that they understand the rules of child support and how much they should receive while on and off welfare.

This report examines the interaction of child support and welfare receipt by addressing several questions concerning child support receipt for samples of current and former welfare recipients. We use several data sources, covering different samples, time periods, and geographic areas, to address these questions in an effort to understand current and former recipients’ status in the child support system, such as their rates of receipt before and after leaving welfare, and the effects of child support on self-sufficiency. This report is the second and final in a series for this project. The first report provided an extensive literature review addressing each of the key research questions.1

Table S1 presents data on child support receipt by welfare status at the time of the follow-up surveys. In general, those who left welfare reported higher levels of child support receipt than those who were still receiving benefits.2 In the SIPP, for example, among women who were still on welfare, 36 percent had orders and 20 percent reported receiving payments. Among those who had left welfare, 49 percent had orders and 28 percent received payments. Among those with orders, women who had left welfare were also more likely to receive payments (50 percent versus 43 percent), although average order amounts were similar. This pattern also holds for the Urban Change sample, but is less pronounced for the PFS sample. Child support also represents a higher fraction of income for women who had left welfare.

The bottom panel presents yearly data for the CSDE sample, based on whether they did or did not receive welfare in 1999. Leavers are more likely than stayers to receive child support (50.5 percent versus 44.7 percent), and they received higher amounts on average. Because these data are from administrative records, these differences in amounts received are real, rather than a result of women on welfare reporting less than what is being paid on their behalf. Among all women, child support represents about 7 percent to 9 percent of family income, although the fraction is higher for women actually receiving payments.

Former welfare recipients may receive more in child support simply through the “mechanical” effect of receiving more of the collections paid on their behalf. Mothers on welfare, in contrast, will receive only the pass-through amount or, if they live in a state that eliminated the pass-through, no child support. In addition to this mechanical effect, leaving welfare may induce certain behavioral changes as well. Mothers may be more proactive in seeking child support once they no longer have welfare as a source of income. In addition, fathers may be more likely to pay or pay more in child support once mothers leave welfare, knowing that the mothers will now receive all of their payments. Further, it is possible that a programmatic decision to prioritize TANF cases or certain TANF cases would impact child support receipt.

• Child support outcomes have improved over time for the samples as a whole. More women are receiving payments; more women are receiving high payments; and child support has become a more important income source. The largest changes occurred for the CSDE and Urban Change samples, with modest changes for the SIPP sample. These improvements in child support outcomes were due in part to more women having orders and in part to higher payment rates among women with orders.

Data from the Current Population Survey Child Support Supplement (CPS-CSS) for all eligible women show large increases in receipt rates over the past decade for never-married women and more modest increases for previously married women. For the samples used in this report, there is also a trend toward improved outcomes over time, with the exception of the SIPP sample, which showed receipt rates of about 22 percent in both month 20 and 44 of the panel (see Table S2). In contrast, receipt rates increased for women in the Urban Change and CSDE samples, from 9 percent to 16 percent and 41 percent to 53 percent, respectively. Average amounts received (among women receiving payments) increased over time for all samples including the SIPP, with the result that child support became a greater fraction of family income.

• In general, child support outcomes improved more over time for welfare recipients than for those who had left welfare by the first follow-up survey. The increase in receipts for welfare recipients represented a real increase in payments made on their behalf and was not simply an increase in the amount of support received due to the fact that some of them left welfare subsequent to the first survey.

Figure S1 presents trends over time for the SIPP and Urban Change samples by welfare status. With one exception, all groups experienced an increase in award rates, and those with awards had an increase in payments rates. The changes were largest for the Urban Change sample, perhaps because there was more room for improvement. Average receipts (among those receiving payments) also increased for most groups, with relatively larger increases for those still on welfare as of the first survey. Further analysis from the CSDE indicates that this increase in receipts for those on welfare is due in part to an increase in payments made by noncustodial fathers, and not simply to the fact that mothers began to receive more of the payments on their behalf after they left welfare.

• Payments are somewhat unreliable from month to month. Among women who were receiving child support early in the PFS and SIPP panels, for example, from 35 percent to 39 percent did not receive payments for more than 5 consecutive months.

For the SIPP sample, women who received payments at some point during the first year received them for an average of 18.5 months out of the 48-month period. Women in PFS who received payments at some point during the panel received payments for an average of 7.8 months out of the 14-month period. Figure S2a presents one measure of reliability — the length of the first spell of payments among women who received payments in the first several months of the panel.3 The figure shows that a significant share stopped receiving payments within 5 months (39 percent for the SIPP sample and 35 percent for the PFS sample). In general, spell lengths are longer for the PFS sample, which partially reflects the fact that the PFS treatment increased child support payments and the reliability of payments.

• The yearly data from the CSDE show some long-term reliability and an increase in payments over time. Most women who did receive support received similar or higher amounts over time, although a significant number of women did not receive payments in year 1 or year 5.

Figure S2b presents a measure of reliability over a 5-year period. The figure shows child support status in year 5 by status in year 1. For example, among women who did not receive payments in year 1 (59.2 percent of the sample), over 60 percent also did not receive payments in year 5. The most stable situation is among women who received amounts over $4800 in year 1, given that more than 50 percent of them also received this high amount in year 5. However, this group represents less than 3 percent of the full sample. Looking at the intermediate groups shows that the majority in each group received an equal or higher amount in year 5 compared with year 1. Among those receiving $1 to $2400 in year 1, for example, more than 70 percent were receiving that amount or more in year 5.

• Several demographic characteristics of custodial mothers are associated with the likelihood of receiving child support, such as education level, race/ethnicity, and marital status. Some factors, such as education and marital status, increase the likelihood of receiving support partly through their effect on the likelihood that women have orders in place.

The results from the full report (not shown here) suggest that current welfare receipt is associated with a reduced likelihood of having an order in place and of receiving payments, although women with a longer history of welfare receipt are more likely to receive support than those who are relatively new to the welfare system. Black and Hispanic women are less likely than their white counterparts to have child support orders, are less likely to receive child support, and receive lower amounts. Less educated women are also less likely to receive support, in part because they are also less likely to have awards in place. Marital status also has a strong relationship with child support outcomes — previously married (divorced) women are more likely than married women to receive child support, although some of this effect may operate through their greater likelihood of having orders in place. Finally, the PFS data show that black women are less likely than white women to receive formal child support but are more likely to receive informal support.

Patterns of Child Support Receipt During the Transition from Welfare to Self-Sufficiency

• The SIPP data show that overall rates of child support receipt remain relatively steady as women transition off welfare, although there is a drop in receipt rates after the welfare exit among women receiving child support in the months prior to exiting welfare. This drop is offset by an increase in receipt rates among women who did not receive child support prior to exiting welfare.

For the full sample of women who left welfare during the panel, there is a small increase in receipt rates after exit, from 29 percent in the few months surrounding exit to 33 percent in the 5 to 10 months after exit (not shown). This overall rate reflects quite different patterns for women based on their receipt status prior to exit. Figure S3 presents receipt rates for women who did and did not receive child support in the wave before exit.4 For women receiving child support before exit, there is a substantial drop in receipt rates during the wave of exit, to 72 percent, and receipt rates decline slightly in the subsequent two waves. The opposite pattern holds for women who did not receive child support in the wave prior to exit — about 10 percent are receiving child support in the wave of exit and 17 percent by two waves after exit. These changes in receipt rates for both groups may simply reflect the instability of payments over time. However, some women may also experience a decrease in child support because of the reason they left welfare, for example if they left because of marriage to or cohabitation with the father or if their children aged out of eligibility for welfare and child support. Part of the increase for women not previously receiving child support may be due to the fact that under PRWORA many states stopped passing child support to TANF recipients. In this case, women would experience an increase in receipt rates once they left welfare, since they would begin receiving payments made on their behalf.

 The Effects of Child Support Receipt on Self-Sufficiency

• Across most samples, child support receipt has little effect on employment status. The effects on welfare status are more mixed and suggest that if any effects do exist, they may be short lived. These results may be due to the fact that the changes we observe in our data in child support amounts are not large enough to generate effects on these measures of self-sufficiency. It might be the case that child support can act as a support for leaving welfare, if the amount received is large enough.

Table S3 presents the effects of child support receipt on work and welfare status for the SIPP, Urban Change, and CSDE samples. Each number under the columns labeled “coefficient” represents a separate regression model and is the effect of child support receipt on the outcome given in the leftmost column. The first column of coefficients shows the effect of child support when it is entered directly into the regression model. The second column presents the results from an instrumental variables approach. In this case, child support is instrumented, or predicted, using state child support enforcement variables.5

For the SIPP sample, the noninstrumented models show that child support receipt increases the likelihood of leaving welfare, reduces the likelihood of returning, and has no significant effect on the likelihood of working. In the instrumented models, child support continues to increase the likelihood of leaving welfare and now has a statistically significant effect on work.

The Urban Change data show that the receipt of child support in wave 1 increases the likelihood of welfare receipt at wave 2 and has no effect on work. In this case, only the noninstrumented effects are available, since there is not enough variation in state policies to predict child support (recall that the Urban Change sample covers four states). The final panel presents results from the CSDE data. Since the CSDE treatment was found to increase child support payments, child support receipt is instrumented in this case using the variable indicating experimental, or treatment, status. The first set of columns show that child support receipt in year 1 reduces the amount of cash benefits received in year 2. However, the instrumented coefficient is substantially smaller and statistically insignificant.

Experimental Findings

CSDE and PFS were two experiments designed to increase child support payments and receipts, and results from both evaluations show that they achieved this goal. Both evaluations also examined each program’s effects on other outcomes, including mothers’ welfare receipt and employment. If child support has a true effect on work and welfare outcomes, then the program itself, through its effect on child support, should affect the work and welfare status of women in the evaluation.

In general, both evaluations found few lasting effects on these secondary outcomes (not shown). In CSDE, for example, mothers in the treatment group received fewer welfare benefits than those in the control group, although this effect did not persist beyond the first year. There were few effects on mothers’ employment and earnings. Results for PFS were similar, showing little effect on mother’s employment rates or welfare receipt rates, even for subgroups that showed relatively larger increases in child support receipt.

Parents’ Knowledge of Child Support Rules

• National data from the SIPP (not shown) suggest that most women are not aware of the amount of child support collected on their behalf and are likely to understate that amount.



• The majority of women in PFS who were receiving welfare reported that they received an amount of child support that was equal to or less than the pass-through amount, even while payments being made on their behalf were higher. However, nearly one-third of welfare recipients reported receiving amounts higher than the disregard, suggesting that at least some of them were reporting amounts paid on their behalf.

The top panel of Table S4 presents data from PFS on reported receipt of child support from the survey, compared with amounts paid according to CSE administrative records data. The survey and records data are likely to differ for women receiving welfare, since, at the time of the demonstration, most states, and all but one PFS site, only passed through the first $50 in payments made on their behalf.6 For these women, a difference in amounts does not necessarily indicate a lack of understanding of how much support was being paid on their behalf.

As expected, reported amounts received from the survey generally match records data reporting of what was paid more closely for women not on welfare at the time of the survey. Among women on welfare, 50 percent had a difference between the two sources of $100 or more. Among women on welfare, 69 percent reported receiving the disregard amount or less. Thus, even in sites with a $50 pass-through, one-third of women on welfare reported more than the pass-through amount. It is possible that these women were including informal payments in reported amounts or that there was considerable reporting error. On the other hand, it is also possible that these women were more aware of payments being made on their behalf and reported that amount on the survey.

• There is a fairly low level of knowledge of distribution rules in Wisconsin, among both fathers and mothers, but particularly among fathers. Fewer than half of the mothers and only about a quarter to a third of fathers responded correctly to questions about the pass-through rules. Status in the welfare system and actual child support experience seem more important predictors of knowledge than demographic characteristics.

The bottom panel of Table S4 presents data from the CSDE on parents’ knowledge of specific child support rules. In particular, two questions were asked of both mothers and fathers about the treatment of payments while the custodial mothers were on or off welfare. The correct answer to the first question, about whether the child’s mother would receive all support if she were receiving benefits, depends on the mother’s random assignment. The analysis is conducted for the full sample, by treatment group and by AFDC history. Those with a long history of AFDC receipt were more likely to have been exposed to the prior policy regime (the AFDC $50 pass-through and disregard, or partial pass-through) and thus may not have absorbed the new rules of the W-2 full pass-through and disregard (that is, all current support paid in a month).

The answers to the two questions indicate that mothers did not have a very clear understanding of the rules. At most, a little more than half knew the right answer to either question, and almost a third indicated they did not know (not shown). For the treatment group, less than one-quarter understood how their receipt of child support would be affected by the new policy being applied to them. Fathers’ knowledge levels are lower than mothers’ levels. In general, there are no large differences between the long-term AFDC group and other fathers.

The report also examined the correlates of parents’ knowledge of the rules (not shown here). Few demographic characteristics were found to predict correct responses to these questions, among both mothers and fathers. More accurate knowledge seems most consistently related to whether child support has been paid on a mother’s behalf since she entered W-2. This suggests that policy knowledge may follow from direct experience — mothers know more when they see how the system treats child support paid on their behalf. Fewer factors help explain fathers’ knowledge.

The Effects of Pass-Through and Distribution Policies on Child Support Outcomes

• More generous pass-through and disregard policies can lead to an increase in the number of fathers making payments, an increase in the number of mothers receiving payments, and an increase in the average payments mothers receive. These policies also have some short-term effects on paternity establishment, through a speeding up of the process.

Table S5 presents a summary of findings from the experimental CSDE evaluation.7 The CSDE evaluated the effects of a full pass-through and disregard of child support payments, compared with a policy that passed through $50 or 41 percent of payments, whichever was larger. Thus, the CSDE evaluates the effects of moving from a somewhat generous pass-through policy to a full, or even more generous, pass-through policy. The first row shows that the pass-through policy led to an increase in payments by noncustodial fathers and an increase in the average amount paid, although the latter effect was statistically significant only in the second year.

The pass-through policy also increased the receipt of child support and the average amount received by mothers. Note that the latter effect is to be expected in part due to a “mechanical” effect of simply passing through more support to mothers who are receiving cash welfare, rather than to a “behavioral” effect of fathers paying more. The increase in payments by fathers, however, indicates that the increase in receipts by mothers was not due entirely to this “mechanical” effect. Finally, the pass-through policy increased the rate of paternity establishment, but only in the first year. The effects on paternity establishment are also fairly modest in size, compared with the effect on payment and receipt rates.

The report also includes information from two nonexperimental studies of the effects of more generous pass-through policies. The findings tend to confirm the findings from the CSDE, indicating that more generous state distribution policies increase rates of collections, receipts, and paternity establishment.

The Effects of Welfare Reforms

• In two programs with welfare time limits, welfare receipt was reduced and child support receipt increased, while in one program with financial incentives and work mandates, welfare receipt was increased and child support receipt reduced.  The impacts on child support receipt tend to be linked with the impacts on welfare receipt.  There were no discernable effects on child support from the changes in pass-through policies that were part of two of the demonstrations. 

The report examines the effects of four waiver demonstrations on welfare and child support receipt. The four programs include Connecticut’s Jobs First program, Florida’s Family Transition Program (FTP), Minnesota’s Family Investment Program (MFIP), and Vermont’s Welfare Restructuring Project (WRP). All programs included financial incentives, and almost all included some type of work or participation mandate. Two programs (Jobs First and FTP) included time limits on the receipt of benefits, while two (Jobs First and WRP) included changes to the child support distribution rules that attempted to make the payments more visible and to give both parents a greater stake in how much is paid.

Figure S4a presents impacts (or differences between the treatment and control groups) in rates of welfare receipt and child support receipt, measured at the time of the follow-up surveys, or three to four years after program entry. FTP and Jobs First led to statistically significant reductions in welfare receipt, which is expected, given that they include time limits. MFIP, on the other hand, increased the rate of welfare receipt. This result is also expected, given the program’s more generous incentives that allowed families to work and still remain eligible for some benefits. With the exception of WRP, programs that reduced welfare receipt (FTP and Jobs First) led to increases in child support receipt, while those that increased welfare receipt (MFIP) led to reductions in the rate of child support receipt. WRP reduced welfare receipt, although the effect is not statistically significant and had little effect on child support receipt. The fact that child support receipt rates were already relatively high in Vermont (41 percent of the control group received payments) may help explain the lack of impacts in this program.

In general, there were no consistent effects for particular subgroups across programs. Jobs First, for example, had little effect on child support receipt for black women, while FTP led to a relatively large increase in child support. Both programs, in contrast, reduced welfare receipt for this group.

• Women who left welfare because of a time limit in FTP were more likely to receive child support, possibly a result of more attention paid by caseworkers at the point of welfare exit.

Figure S4b presents rates of child support receipt for three groups of women who left welfare during the follow-up period — those in the treatment group who left because of reaching their time limit, those in the program group who left before reaching their time limit, and those in the control group. In FTP, women who left because of a time limit were more likely than the other two groups to receive child support. The group with the lowest rate was control group leavers. A similar pattern was shown in the final report for the FTP evaluation, in which time-limit leavers were more likely to receive other transfers, such as food stamps.8 The authors attributed the difference in part to the exit interviews given to women reaching their time limits, in which eligibility for nonwelfare benefits was assessed. In contrast, receipt rates are very similar for the three groups in Jobs First.

Although exit interviews were also given to women approaching time limits in Jobs First, a key difference between Jobs First and FTP was the way in which the time limit was implemented. Under Jobs First, many recipients who reached their time limit without jobs or with very low earnings were given six-month extensions. Thus, the time-limit leaver group in Jobs First comprised women with higher average earnings than the other two groups, which may have reduced the need to pursue child support.