Published in the Fall 1998 Edition of theGeorgetown Public Policy Review
Note: This report is also available in PDF format, which many find useful when printing.
The authors thank Linda Mellgren and Reuben Snipper of the U.S. Department of Health and Human Services for their comments on earlier drafts of this paper and Kyna Rubin for editorial assistance. Support for this research is from the U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation. The opinions expressed in this paper are those of the authors and do not necessarily reflect the opinions of the Urban Institute or its sponsors.
The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders.
Abstract
To what degree has the nation been successful in reducing welfare costs through child support, and how much more welfare savings might we expect if more fathers were to pay child support? Using the Urban Institute's unique microsimulation instrument (TRIM2), we find that child support collections reduced the combined costs of AFDC, the Food Stamp Program, and Medicaid by 2 percent in 1989 and that if all custodial mothers had child support orders that were fully paid, child support collections could reduce costs by another 8 percent. In order to assist policymakers in judging the likely impact of incremental reforms, we also present estimates of "cost avoidance"—government savings per dollar of child support collected. We estimate that, in 1989, each dollar of child support produced an average of $0.14 in program savings and that incremental expansions in child support enforcement could yield roughly $0.23 in savings for each additional dollar collected.
Introduction
Child support reform was one of few areas of consensus among federal legislators as they enacted the Personal Responsibility and Work Opportunity Reconciliation Act of 1996. The nearly universal support for child support reform stems, in part, from the expectation that it can reduce welfare costs, welfare dependency, and poverty. Elaine Sorensen's estimate that nonresident fathers would pay an additional $34 billion in child support if each had a child support order and paid the full amount due has been cited extensively by President Clinton and other policy makers, and would seem to suggest a potential for large savings in welfare programs (Sorensen 1997).
But the potential for child support payments to reduce spending in government transfer programs is determined by a complex array of factors including the income and welfare characteristics of potential recipients, the amount of additional child support each family would actually receive, and the treatment of child support income in determining eligibility and benefits for various programs. In this paper we use the Urban Institute's unique microsimulation instrument, the Transfer Income Model version 2 (TRIM2), to estimate savings generated by child support payments in 1989, and to estimate the potential for additional savings under two hypothetical expansions to the child support system.
Our analysis may be both more narrow and more broad than some readers would expect: narrow, in that we do not capture the administrative savings that occur when child support removes a family from the welfare roles, nor the costs associated with establishing and enforcing child support orders; broad, in that we analyze all child support payments, regardless of whether they are made privately or through the government's Child Support Enforcement Program. Government savings presented here are savings to all levels of government, and are not disaggregated into federal and state or local shares. We do not consider the potential savings from payment of child support "arrearages"—child support owed from prior years. We seek simply to show the effect of child support dollars on benefits and participation in Aid to Families with Dependent Children (AFDC), the Food Stamp Program, and Medicaid.(1)
We begin with an overview of the current child support and welfare status of custodial mothers and their families. We then summarize prior research estimates of the additional amount that noncustodial fathers can afford to pay and of the potential for child support to reduce welfare costs. We describe the method used here to estimate savings from current and increased levels of child support, then present our simulation results, both in terms of aggregate savings in transfer programs and in terms of savings per dollar of child support paid. We close with a summary of findings and research caveats.
Child Support Payments, Poverty Status, and Welfare Use Among Custodial Mothers(2)
The percentage of custodial mothers who receive child support is not high. In 1989, only 58 percent of custodial mothers had child support awards (U.S. Bureau of the Census 1991).(3) Of those who were due payments, only half received the full amount due. Even receiving the full amount of child support due did not necessarily guarantee custodial mothers large sums of child support. In 1989, the average amount of child support due was only $3,055; the average amount received was $2,544.(4)
Many custodial families are poor. In 1989, 30 percent of custodial mothers and their families lived in poverty, comprising three-fifths of the nation's poor children. With insufficient earnings and little or no child support, many poor custodial mothers turn to public assistance for help. In 1989, 28 percent of custodial families received AFDC at some point during the year. The average AFDC benefit was $3,735. About the same percentage of custodial families (28 percent) received food stamps, with an average food stamp benefit of $1,727. About 36 percent of custodial families received Medicaid in 1989; their average benefit was $2,600.
Research Suggests Noncustodial Fathers Can Afford to Pay More
As policymakers look for ways to reduce government welfare costs and to improve the well-being of custodial families, they have turned their attention toward requiring noncustodial parents to fulfill their share of the financial burden. On average, noncustodial fathers are better off financially than custodial mothers are. Whereas 30 percent of custodial families are poor and one quarter of custodial mothers do not work, only 15 percent of noncustodial fathers are poor and 90 percent work (Sorensen 1997). But getting fathers to pay more child support is only a first step. Even full payment by noncustodial fathers would not eliminate welfare and poverty entirely. The extent to which welfare and poverty can be reduced is determined by the amount of additional child support that the noncustodial fathers of poor children have the ability to pay.
Sorensen (1997) estimates that if each custodial mother had a child support award that was calculated by applying the Wisconsin child support guidelines to fathers' current income, and each father paid the full amount due, an additional $34 billion in child support would be paid.(5) This suggests a potential for sizable welfare savings, but the estimate includes potential payments to nonpoor as well as to poor custodial families. Sorensen arrived at the estimate by comparing the current amounts of child support reported by noncustodial fathers on the 1990 Survey of Income and Program Participation (SIPP) to the amounts calculated by applying the Wisconsin child support guidelines to fathers' current income. Sorensen finds that, on average, noncustodial fathers (including those who did not pay child support) paid only 7 percent of their personal income on child support in 1990. Full payment under the Wisconsin guidelines would raise this figure to 21 percent and would nearly triple the total amount of child support paid to custodial families. However, to the extent that the 15 percent of fathers who are poor are the fathers of children on welfare, the ability to reduce welfare costs through increased child support will be limited.
Earlier Research into the Potential Program Savings from Increased Child Support
Previous estimates of the extent to which increased child support could reduce welfare costs have focused on savings to AFDC. Until it was replaced by Temporary Assistance to Needy Families (TANF) in 1997, AFDC was the primary cash assistance program to poor families with children.(6) Using data from the April 1982 Current Population Survey, Robins (1986) estimated the potential effect on AFDC costs and participation of different levels of improvement in child support collections in 1981. He found that under a system in which each custodial mother has an award and the full amount is collected, AFDC participation among custodial families would fall by 6 percent and AFDC costs for custodial families would fall by nearly 30 percent.
Oellerich, Garfinkel, and Robins (1991) tested the theoretical limits of the ability of the child support system to reduce AFDC costs and poverty by examining a scenario in which each noncustodial father pays the full amount of an award that is set in accordance with his ability to pay. The authors tested the effects of using two different uniform normative standards of the ability to pay. The Wisconsin standard sets the child support obligation as a percentage of the noncustodial parent's gross income, while the Colorado standard determines the needs or cost of the child[ren] based on the combined gross incomes of the custodial and noncustodial parents. The authors estimate that if all awards were set at a percentage of the father's income in accordance with the Wisconsin child support guidelines and the full amount were collected, AFDC participation among custodial families would fall by 16 percent and AFDC costs for custodial families would fall by 33 percent. The Colorado standard is estimated to have a similar effect.
Brien and Willis (1997) estimate that absent fathers are able to provide child support equal to as much as 40 to 50 percent of AFDC benefits. This estimate is derived by using the Wisconsin guidelines to calculate an average monthly child support payment based on the discounted present value of fathers' earnings until the child reaches 18. Brien and Willis use the discounted present value of future earnings in order to draw attention to fathers' wage growth potential. The 40 to 50 percent savings estimate is derived by comparing the maximum AFDC benefit in the state of Wisconsin to the median child support award calculated for various groups of custodial mothers. A weakness of this approach is that it does not reflect the fact that child support awards are based on current earnings, not discounted future earnings. Although it might be argued that child support based on the discounted value of future earnings reflects an average savings to AFDC over the lifetime of the child, such savings will be overstated for mothers who leave AFDC before the child turns 18.
The estimates generated by these studies provide an indication of the maximum potential savings to the AFDC program from increased child support, but do not address the potential for savings in other programs. Lewis (1988) estimates the potential for savings in other programs that benefit poor families, including the Food Stamp Program and Medicaid. In addition, he describes the interactions between these programs—the ways in which a change in a family's benefit from one program can affect the benefit it receives from another program. Taking these interactions into consideration, Lewis estimates the effect on costs to the federal government of increased child support to a prototypical family at different levels of earnings. However, he does not provide an estimate of the aggregate effects of increased child support for all custodial families.
The advantage of using a microsimulation model such as TRIM2 is that the type of program interactions described by Lewis are captured not just for a prototypical family, but for each custodial family in a nationally representative survey of households. Using TRIM2, we can calculate the effect of child support on each family's eligibility and benefits and then aggregate the results to determine the total effect of increased child support on program costs and participation. Our methodology is conceptually similar to that of Oellerich, Garfinkel, and Robins (1991) in that we also use microsimulation techniques to simulate the effects on the AFDC program of an expanded child support system. However, while Oellerich, Garfinkel, and Robins consider only the impact on AFDC, we also consider the effect on the Food Stamp Program and Medicaid.
Modeling Program Savings Generated by Child Support
Below, we use the Urban Institute's TRIM2 model to estimate AFDC, Food Stamp Program, and Medicaid savings generated by child support paid in 1989 and the potential for additional savings under two hypothetical expansions to the child support system. We use 1989 data because more recent data are not generally available.(7) TRIM2 is a comprehensive microsimulation model of government tax and transfer programs that is used to estimate the costs and effects of changes in government programs.(8) The database underlying the model is the March Current Population Survey (CPS). For this particular analysis, we used the March 1990 CPS, which provides us with income and demographic data for calendar year 1989 and is referred to as the "1989 input file."
We conducted four simulations for this analysis. The first simulation created a baseline to serve as a reference for the remaining simulations. This baseline simulation applied 1989 program rules to each family on the 1989 input file. Child support payments and award amounts were imputed onto the 1989 input file, based on equations estimated using the March/April 1990 CPS matched file, which includes detailed child support information. Eligibility and benefits for AFDC, food stamps, and Medicaid were then calculated for each household.(9)
In the second simulation, we set all child support payments to zero and recalculated eligibility and benefits under the AFDC, Food Stamp, and Medicaid programs. This simulation provides an estimate of what program costs would be if no child support were paid. Comparing the second simulation to the baseline simulation provides us with an estimate of program savings generated by child support paid in 1989.
The remaining two simulations explore the potential for additional program savings under two hypothetical expansions in the child support system: "full payment" and "full establishment and full payment." In the "full payment" simulation, we address the problem of underpayment of child support orders. As stated previously, half of all women with a child support order in 1989 did not receive the full amount due. In order to estimate the effect of perfect enforcement of all existing child support awards, we (hypothetically) assign each woman with a child support award the full amount due, and recalculate eligibility and benefits for AFDC, food stamps, and Medicaid.
Because 42 percent of custodial mothers did not have a child support order in 1989, a simulation of full payment of existing child support orders only begins to capture child support's potential to reduce program costs. Under the "full establishment and full payment" scenario, we simulate the maximum potential of child support to reduce program costs by assigning each woman without a child support award an award that is typical for a woman with her income and demographic characteristics.(10) We then simulate program eligibility and benefits assuming full payment of child support awards.
Simulated Program Savings Generated by Child Support Payments in 1989
Table 1 shows child support payments and program participation and costs in the baseline simulation and under the scenario in which no child support is paid. We estimate that nationwide there were 11.4 million custodial families in 1989, of whom 4.2 million received a total of $10.8 billion in child support. A total of 9.6 million persons in custodial families received AFDC, 11.3 million received food stamps, and 12.1 million received Medicaid. Persons in custodial families accounted for 72 percent of all AFDC recipients, 44 percent of all food stamp recipients, and 52 percent of all Medicaid recipients (not shown). Benefits paid to custodial families totaled $28 billion, of which $11.7 was for AFDC benefits (less child support retained by the government to help offset AFDC costs), $5.6 billion for food stamps, and $10.7 billion for the insurance value of Medicaid coverage.(11) Benefits to all households under these three programs totaled almost $60 billion that year (not shown). Thus, custodial mothers and other members of their households received nearly 50 percent of these program benefits.
If no child support had been paid in 1989, the number of custodial family members receiving AFDC and food stamps would have been 3 percent higher, and 1 percent more would have enrolled in Medicaid. Program costs for custodial families would have totaled $29.4 billion, 5 percent higher than in the baseline simulation. Program costs for all households would have been 2 percent higher. Thus, we estimate that child support payments made in 1989 caused benefits paid under these three programs to be 2 percent lower than they would have been otherwise. Savings were greatest for AFDC and lowest for Medicaid. Below we provide a detailed explanation of savings in AFDC, the Food Stamp Program, and Medicaid.
AFDC Savings. Of the three programs, child support payments generate the greatest savings to AFDC. In the absence of child support, program costs for custodial families would have been $1 billion higher in 1989. Thus, child support produces savings of 8 percent in AFDC costs for custodial families and 6 percent in costs for all families. Custodial mothers and their families represent about 72 percent of the AFDC caseload, so an 8 percent reduction in costs to custodial families is equivalent to a 6 percent reduction in overall program costs.
The reason that savings are greatest for the AFDC program is that any child support payments in excess of $50 per month to a custodial mother receiving AFDC are retained by the government in order to reduce AFDC costs. AFDC recipients are required to assign their rights to child support to the government as a condition of receiving AFDC. However the first $50 of child support received, or any lesser amount, is given to the custodial mother without affecting her AFDC benefit—a practice better known as the $50 pass-through rule. The government keeps anything beyond the first $50 to reduce program costs.
Food Stamp Program Savings. Food Stamp Program costs for custodial families are $300 million lower than they would be in the absence of child support. This represents a savings of 6 percent in program costs for custodial families, and 3 percent in overall program costs.
Food Stamp Program savings are smaller than AFDC savings for two reasons. First, since AFDC retains much of the child support paid on behalf of custodial families participating in AFDC, this money is unavailable for reducing Food Stamp Program costs. Second, under the rules of the Food Stamp Program, a unit's benefits are reduced by only thirty cents for each additional dollar of income.(12)
Medicaid Savings. Medicaid program costs for custodial families are $100 million less than they would have been in the absence of child support. This represents a 1 percent savings in costs for custodial families and a negligible savings in overall program costs.
TRIM2 measures Medicaid costs as the insurance value of Medicaid coverage, taking into account the demographic characteristics of each enrollee, but does not capture savings to Medicaid from child support orders that require the noncustodial parent to provide health insurance coverage. Therefore, we miss a source of savings associated with child support.
Of the three programs—AFDC, Food Stamp, and Medicaid—Medicaid experiences the smallest savings from child support. AFDC and the Food Stamp Program experience savings from child support paid to current recipients as well as to families whose child support income renders them ineligible for assistance, but savings to Medicaid are accomplished only through reduced program participation. In other words, child support payments do not affect Medicaid program costs unless they are large enough, in combination with other income, to render a family ineligible for Medicaid. Had we been able to simulate savings to Medicaid from child support orders that include private health insurance coverage, we would have found additional savings.
Potential Program Savings from Two Hypothetical Child Support Expansions
Table 2 displays potential child support payments and program savings for two hypothetical expansions to the child support system. Under the "full payment" expansion, all custodial mothers with a child support award would receive full payment. This would increase the amount of child support paid by $6 billion relative to the baseline simulation. The number of AFDC recipients in custodial families would drop by 3 percent, the number of food stamp recipients by 4 percent, and the number of Medicaid enrollees by 2 percent. Program costs for custodial families would decline by 7 percent for AFDC, 5 percent for the Food Stamp Program, and 4 percent for Medicaid, for a total reduction of 5 percent from 1989 baseline levels. Overall program costs would decline by 5 percent for AFDC, 2 percent for the Food Stamp Program, and 1 percent for Medicaid, for a total reduction of 2 percent from 1989 baseline levels.
Under the "full establishment and full payment" expansion, each custodial mother without an award would be assigned an award that is representative of awards for custodial mothers with similar income and demographic characteristics, and all custodial mothers would receive the full amount due.(13) This would increase the amount of child support paid in 1989 by $20.8 billion. The number of AFDC recipients in custodial families would drop by 9 percent, the number of food stamp recipients by 11 percent, and the number of Medicaid enrollees by 5 percent. Program costs for custodial families would decline by 26 percent for AFDC, 19 percent for the Food Stamp Program, and 5 percent for Medicaid, for a total reduction of 17 percent from 1989 baseline levels. Overall program costs would decline by 20 percent for AFDC, 9 percent for the Food Stamp Program, and 2 percent for Medicaid, for a total reduction of 8 percent from 1989 baseline levels.
Our estimate that full establishment and full payment of child support awards would result in a 9 percent reduction in the number of members of custodial families receiving AFDC and a 26 percent reduction in AFDC program costs for custodial families is quite similar to that of Robins (1986). Robins estimated that under these conditions, AFDC participation among custodial families would fall by 6 percent and AFDC costs would fall by nearly 30 percent. Our results are not directly comparable to those of Oellerich, Garfinkel, and Robins (1991) who simulated full establishment and full payment of child support under scenarios in which awards are set according to the Wisconsin and Colorado child support guidelines, but they are reassuringly close. Oellerich, Garfinkel, and Robins estimate that AFDC participation among custodial families would fall by 16 percent and AFDC costs for custodial families would fall by 33 percent. Since Oellerich, Garfinkel, and Robins incorporate the effects of child support guidelines into their simulation, it is not surprising that their savings estimates are higher than ours.
Our study builds upon these earlier works by showing the potential for savings in the Food Stamp Program and in Medicaid and by showing savings as a percentage of overall program costs, not just costs of custodial families. Since custodial mothers and members of their families accounted for only an estimated 72 percent of all AFDC recipients; 44 percent of all food stamp recipients; and 52 percent of all Medicaid recipients in 1989, we believe that the potential impact of child support on these programs should be assessed in the context of overall program costs, rather than just in the context of custodial family costs.
We selected the full establishment and full payment simulation in order to estimate the maximum theoretical potential for program savings from increased levels of child support. However, we believe this simulation overstates the actual potential for additional child support collections and program savings. It is unrealistic to expect a world in which each custodial parent has a child support award and receives the full amount due. For example, since an estimated 450,000 noncustodial fathers were institutionalized in prisons or mental institutions in 1990 (Sorensen 1997), it is unlikely that all noncustodial fathers have income and are able to pay child support. Our methodology for assigning child support awards may introduce additional error. By assigning women without awards an award that is representative of the awards of women with similar income and demographic characteristics, we may overstate potential awards, since the absence of an award may reflect a lower potential award amount.
Cost Avoidance: Program Savings Per Dollar of Child Support
It is interesting to know the theoretical limits to child support's ability to reduce government transfer program costs, but such information is of limited relevance to those attempting to judge the potential impact of a particular incremental reform. In judging such a reform, policymakers would want to know how much additional child support would be collected as a result of the reform, and how much savings those dollars could produce. To help with such analyses, we use a "cost avoidance" measure that is equal to the average savings in government transfer programs for each additional dollar of child support collected.(14) By applying the cost avoidance measure to the expected increase in child support collections, government analysts can roughly estimate the likely savings from the reform.
Table 3 shows cost avoidance for child support paid in 1989 and for each of the two hypothetical expansions. The cost avoidance measure of .14 for total program costs in 1989 indicates that each dollar of child support paid in 1989 saved, on average, $0.14 in total program costs. The cost avoidance measure of 0.23 for the two expansion simulations indicates that each dollar of additional child support payments under these expansions would save an average of $0.23 in total program costs. For the simulations presented in Table 3, cost avoidance ranges from .09 to .15 for AFDC, from .03 to .05 for the Food Stamp Program, and from .01 to .05 for Medicaid.
Cost avoidance varies among these simulations due to differences in the way in which child support payments are distributed. In general, the higher the percentage of child support payments going to poor custodial families, the higher the cost avoidance. For example, a reform that rescinded doctors' licenses of noncustodial parents who failed to pay child support would probably yield little cost avoidance, since payments would primarily benefit upper-income custodial families. But a reform that increased paternity establishment in out-of-wedlock births would probably yield greater cost avoidance, since a higher proportion of low-income custodial families would benefit.
Cost avoidance is lower for child support payments made in 1989 than it is under either of the hypothetical expansions because poor custodial mothers receive the least child support under the current system and have the most to gain from the expansions. For example, only 13 percent of child support payments made in 1989 were made on behalf of custodial families who received AFDC at some point during the year, but 28 percent of the additional child support payments generated under the full establishment and full payment simulation would go to these families.
In Table 4, we take a closer look at the different levels of cost avoidance associated with different groups of women for child support paid in 1989. In the first column we see that, in 1989, subtracting $10.8 billion in child support from 4.2 million families would have resulted in $1.5 billion in new program costs. Overall, cost avoidance of child support payments made in 1989 was 0.14 ($1.5 billion divided by $10.8 billion). However, if we focus only on poor custodial mothers who receive AFDC in each month of the year (second column) cost avoidance is much higher. Each dollar collected for these mothers generates an average of $0.75 in program savings—primarily to the AFDC program, but also to the Food Stamp Program. Cost avoidance for poor women who do not receive AFDC in each month of the year is also high. Each dollar collected for these mothers (third column) generates an average of $0.43 in program savings—about half of which go to AFDC, with the remainder split between food stamps and Medicaid. Cost avoidance is lower for child support paid to women with incomes between 100 and 200 percent of the poverty line (fourth column). Each dollar collected for these women generates an average of $0.17 in program savings. Finally, cost avoidance is negligible for women with incomes higher than 200 percent of the poverty line. The vast majority of these families is not eligible for AFDC, food stamps, or Medicaid, and would not be eligible even in the absence of child support.
As suggested above, the overall cost avoidance figure will be determined largely by the characteristics of the recipients. In 1989, 46 percent of custodial mothers had incomes above 200 percent of the poverty line. Of custodial families receiving child support, 57 percent had incomes above 200 percent of the poverty line and received 62 percent of the total amount paid. In contrast, 17 percent of custodial mothers were poor and received AFDC in each month of the year. However, of custodial families receiving child support, 10 percent were poor AFDC recipients and received only 7 percent of the total amount paid. Since only 7 percent of child support dollars paid in 1989 were paid on behalf of the group where cost avoidance is highest, and 57 percent were paid to the group where cost avoidance is lowest, the overall cost avoidance measure is fairly low (0.14). Cost avoidance would be higher in the expansion scenarios presented in Table 3 because the poor would enjoy a larger share of the newly generated child support payments than they did in 1989.
Conclusions
Clearly, child support has not fully realized its potential for reducing government transfers to poor families. But child support is only part of the solution. Even if each custodial family had an award and received payment in full, the combined costs of AFDC, the Food Stamp Program, and Medicaid would be reduced by only 17 percent for custodial families and by 8 percent overall.(15) Since it is unrealistic to expect a world in which every custodial mother receives child support, our estimates should be considered upper bounds to child support's potential to reduce program costs. However, our cost avoidance estimates suggest that incremental reforms that move us closer to this "perfect world" would reduce program costs by an average of $0.23 for each dollar of additional child support collected.
There are several caveats to our analysis. First, it was conducted using 1989 data and reflects 1989 demographics, child support levels, and program rules. Since then there have been many improvements in child support enforcement—suggesting that we may already have made some progress toward our "perfect world" scenario. On the other hand, an increasing percentage of custodial mothers were never married to the fathers of their children, making improvements in child support enforcement more difficult to achieve.
A second caveat is that behavioral responses to increased child support enforcement are not included in our model and might produce additional savings to government programs. For example, if additional income from child support encourages a custodial mother to start working, the long-run savings from child support could be greater than we have simulated. However, behavioral incentives could also serve to discourage work on the part of the custodial mother or noncustodial father. In the end, the potential net effect is ambiguous.(16)
A third caveat is that award levels in our simulation may not adequately represent fathers' ability to pay. For mothers who had child support in 1989, award levels may underrepresent fathers' ability to pay, since awards often do not keep pace with growth in fathers' income. On the other hand, awards assigned to women in our second simulated expansion may be unrealistically high. Assigning a woman without a child support award an award that is typical for a woman with her income and demographic characteristics will overstate the amount of the award if women are more likely to seek and obtain a child support award from a father who is better able to pay.
Finally, how will cost avoidance be affected by the most recent welfare and child support reforms enacted in 1996? Given the numerous changes in welfare and child support that are being implemented as a result of this legislation, it is difficult to predict. As we see it, some of the changes enacted in 1996 will increase the ability of child support to offset welfare costs. For example, under TANF, states are no longer required to pass through $50 of child support per month to custodial families receiving assistance. To the extent that states cease the $50 pass-through, additional welfare costs will be avoided from child support collections. On the other hand, to the extent that welfare reform reduces the number of families receiving cash assistance, a smaller percentage of total child support dollars will be directed toward reducing TANF costs. If custodial families no longer receiving cash assistance continue to rely on food stamps and Medicaid, then child support dollars that would otherwise have gone toward reducing the costs of TANF will help to reduce costs in these programs (although by not as much as would have been the case with TANF). At the moment, we can only speculate about the overall effect of these reforms. As data become available, it will be important to revisit the question of cost avoidance.
References
Barnow, Burt and Timothy M. Dall, Barbara E. Dannhausen, and Scott B. Scrivner. Forthcoming. The Impact of the Child Support Program on Avoiding Costs to Public Programs. Project Report for the U.S. Department of Health and Human Services. The Lewin Group.
Brien, Michael J. and Robert J. Willis. 1997. The Partners of Welfare Mothers: Potential Earnings and Child Support. The Future of Children. 7(1): 65-73.
Clark, Sandra and Linda Giannarelli. 1994. Child Support Simulation Capabilities of the TRIM2 Model. Technical Documentation. The Urban Institute
Giannarelli, Linda. 1992. An Analyst's Guide to TRIM2. Washington, D.C.: Urban Institute Press.
Lewis, Gordon H. 1988. Winners and Losers in Child Support Enforcement. Carnegie Mellon University.
Oellerich, Donald T., Irwin Garfinkel, and Philip K. Robins. 1991. Private Child Support: Current and Potential Impacts. Journal of Sociology and Social Welfare 18(1): 3-23.
Robins Philip K. 1986. Child Support, Welfare Dependency, and Poverty. American Economic Review 76(4): 768-788.
Sorensen, Elaine. 1997. A National Profile of Noncustodial Fathers and Their Ability to Pay Child Support. Journal of Marriage and the Family 59(November): 785-797.
U.S. Bureau of the Census. 1991. Child Support and Alimony: 1989. Current Population Reports, Series P-60, No. 173. Washington, D.C.: U.S. Government Printing Office.
Tables
|
Table 1
Change in Child Support, Program Participation, and Program Costs,
Assuming No Child Support Paid in 1989
|
|
|
Baseline
|
Alternative:
No
Child Support
Paid
|
Percent Change from
Baseline, Calculated Over:
|
Custodial Families
|
All Households
|
Child Support
|
| Custodial Families (millions)
|
11.4
|
na
|
na
|
na
|
| Recipient Families (millions)
|
4.2
|
0
|
-100%
|
na
|
| Payments (billions)
|
$10.8
|
0
|
-100%
|
na
|
Program Participation by Members of Custodial Families (millions)
|
| AFDC
|
9.6
|
9.8
|
3%
|
2%
|
| Food Stamp Program
|
11.3
|
11.7
|
3%
|
1%
|
| Medicaid
|
12.1
|
12.2
|
1%
|
1%
|
Benefits Paid to Custodial Families (billions)
|
| Total
|
$28.0
|
$29.4
|
5%
|
2%
|
| AFDC
|
$11.7
|
$12.7
|
8%
|
6%
|
| Food Stamp Program
|
$5.6
|
$5.9
|
6%
|
3%
|
| Medicaid
|
$10.7
|
$10.8
|
1%
|
0%
|
| Source: Urban Institute's TRIM2 model, based on March 1990 Current Population Survey.
|
|
Table 2
Potential Program Savings Under Two Scenarios of
Increased Child Support in 1989
|
|
|
Full Payment of
Existing Awards
|
Full Establishment and
Full Payment
|
Child Support
|
| Recipient Families (millions)
|
5.5
|
11.4
|
| Payments (billions)
|
$16.8
|
$31.6
|
| Change (billions)
|
$6.0
|
$20.8
|
Percentage Change in Program Participation by Members of Custodial Families
|
| AFDC
|
-3%
|
-9%
|
| Food Stamp Program
|
-4%
|
-11%
|
| Medicaid
|
-2%
|
-5%
|
Percentage Change in Program Costs (for Custodial Families)
|
| Total
|
-5%
|
-17%
|
| AFDC
|
-7%
|
-26%
|
| Food Stamp Program
|
-5%
|
-19%
|
| Medicaid
|
-4%
|
-5%
|
Percentage Change in Overall Program Costs (all households)
|
| Total
|
-2%
|
-8%
|
| AFDC
|
-5%
|
-20%
|
| Food Stamp Program
|
-2%
|
-9%
|
| Medicaid
|
-1%
|
-2%
|
| Source: Urban Institute's TRIM2 model, based on March 1990 Current Population Survey.
|
Table 3
Cost Avoidance: Program Savings Per Additional Dollar of Child Support Paid,
Under the Current System, and Under Two Scenarios of Increased Child Support in 1989
|
|
|
Current
System
|
Full Payment of
Existing Awards
|
Full Establishment and
Full Payment
|
| Total
|
.14
|
.23
|
.23
|
| AFDC
|
.09
|
.13
|
.15
|
| Food Stamp Program
|
.03
|
.05
|
.05
|
| Medicaid
|
.01
|
.05
|
.03
|
| Source: Urban Institute's TRIM2 model, based on March 1990 Current Population Survey.
|
Table 4
Change in Child Support and Program Costs, Assuming No Child Support Paid in 1989
|
|
|
All
Custodial
Families
|
Poor
|
Nonpoor
|
Full Year
AFDC
|
Other
|
100-200% of
Poverty
|
>200% of Poverty
|
Change in Child Support
|
Custodial Families
(millions)
|
11.4
|
1.9
|
1.5
|
2.7
|
5.3
|
| Recipient Families (millions)
|
-4.2
|
-0.4
|
-0.4
|
-1.0
|
-2.4
|
| Payments (billions)
|
-$10.8
|
-$0.8
|
-$1.0
|
-$2.3
|
-$6.7
|
Change in Program Costs (Billions)
|
| Total
|
$1.5
|
$0.6
|
$0.4
|
$0.4
|
$0.1
|
| AFDC
|
$1.0
|
$0.5
|
$0.2
|
$0.2
|
$0.1
|
| Food Stamp Program
|
$0.3
|
$0.1
|
$0.1
|
$0.1
|
0
|
| Medicaid
|
$0.2
|
0
|
$0.1
|
$0.1
|
0
|
Cost Avoidance of Child Support Payments in 1989
|
|
|
.14
|
.75
|
.43
|
.17
|
.01
|
| Source: Urban Institute's TRIM2 model, based on March 1990 Current Population Survey.
|
Notes
1. Although TRIM2 simulates federal income taxes, payment of child support does not affect the taxes of either the noncustodial parent or the custodial family. Therefore, tax estimates are not presented here.
2. In about 12 percent of custodial families, fathers, rather than mothers, have custody of the children. Data limitations prevent us from including these families in our analysis.
3. We examine 1989 data here because our microsimulation results are from 1989.
4. Statistics that are not cited are based on TRIM2 estimates.
5. Under the Wisconsin guidelines, a noncustodial father is required to pay 17 percent of his gross income for one child and 25, 29, 31, and 33 percent for 2, 3, 4, or 5 or more children, respectively.
6. Under the Personal Responsibility and Work Reconciliation Act of 1996 (PRWORA), AFDC was replaced by the Temporary Assistance to Needy Families block grant (TANF).
7. The Census Bureau released 1991 data on child support but have not yet released 1993 or 1995 data.
8. See Giannarelli (1992) for detailed information about the TRIM2 model and its methodology, and Clark and Giannarelli (1994) for detailed information about the TRIM2 child support methodology.
9. An increase in child support income may affect a custodial mother's decision about whether or how much to work. This "behavioral response" and the resulting change in program eligibility and benefits is not captured in our simulations.
10. The "full establishment and full payment" scenario falls somewhere between Sorensen's (1997) lower-bound and upper-bound scenarios for estimating fathers' ability to pay. Sorensen's lower-bound scenario assumes that all fathers would pay support at the same levels as those currently paying support, and her upper-bound scenario assumes that all fathers would pay the full amount of child support that would be due under the Wisconsin child support guidelines.
11. These benefits represent costs to both federal and state governments. Administrative costs for these programs are not included in our analysis.
12. A family's food stamp benefit is calculated by subtracting 30 percent of the family's net income (income after certain deductions) from the maximum food stamp allotment. An additional dollar of child support income will reduce the family's food stamp benefit by $0.30, assuming the family has positive net income (i.e., income exceeding deductions). If deductions exceed income, the family receives the maximum food stamp allotment, and additional dollars of income will not affect the family's food stamp benefit until the family's income begins to exceed its deductions.
13. An alternative to this approach would be to calculate awards using the formulae set forth in state child support guidelines, using imputations of noncustodial fathers' income (since no actual data about the father is available). We experimented with this approach, using equations similar to those developed by Oellerich, Garfinkel, and Robins (1991). The imputations produced much different estimates of fathers' abilities to pay than Sorensen's (1997) estimates, leading us to question their reliability and to decide upon the approach used here.
14. Several approaches to measuring cost avoidance have recently been developed. See Barnow et. al. (forthcoming) for a review of this and other approaches.
15. Child support has a similarly modest effect on the number of custodial families living in poverty. We estimate that about 3.4 million custodial families were poor in 1989, and that if no child support had been paid, the number would have been about 7.2 percent higher. "Full establishment and full payment" of child support would reduce the number of poor custodial families by 11.4 percent relative to the 1989 level. Child support has a limited impact on poverty for two primary reasons. The amount of money necessary to lift a typical poor family from poverty is quite high relative to the actual or potential child support orders for most poor families, and much of the child support paid on behalf of poor families is retained by the government to help offset AFDC costs.
16. Additional income from child support can encourage a custodial mother to work by helping to cover child care and other costs associated with work. But for a mother who is already working, additional child support may reduce hours of work, since the mother can afford to work less and still enjoy the same standard of living. Child support may discourage work on the part of the noncustodial father because it is essentially a tax on his income, and in many cases may be taken from his paycheck through wage withholding.
Browse by Topic:
PDF