Table of Contents | Previous | Next |
CHAPTER V
POLICY OPTIONS TO ENCOURAGE MARRIAGE AND FAMILY FORMATION
One of the central motivations for welfare reform was the growing perception that welfare policies promoted undesirable behavior. Specifically, there was concern that welfare policies created disincentives to work, marry, pay child support, and delay childbearing (Maynard et al. 1998). State waiver requests and PRWORA responded to these concerns. Most of all, they emphasized removing disincentives to work. It is generally accepted that welfare reform successfully promoted work and contributed significantly to the historic fall in TANF caseloads that occurred after 1994.
Some are now calling on policymakers to broaden the focus of policy changes to include reductions in disincentives to marry (Rector 2002a). The welfare system has long been criticized for discouraging marriage and weakening family structure. As long as assistance programs provide benefits on the basis of the income of a family unit, there will inevitably be some disincentive to add persons to the family who have earnings or future prospects for earnings. The income of an additional working adult in the eligibility unit increases the likelihood that the family will be found ineligible and decreases benefits if the family is determined eligible. Although estimates of the magnitude of these disincentives vary, most empirical studies do show a significant negative correlation between the level of welfare benefits and marriage (Moffitt 1998). In addition to this marriage disincentive inherent in means-tested programs, some programs have rules that further discourage marriage.
This chapter describes some policy options that could encourage family formation. Some of these policies reduce disincentives to form two-parent families; others directly encourage marriage. The focus is mainly on programs overseen by ACF and on changes that states could implement without the need for federal legislation. The chapter begins by discussing changes to two large ACF programs—TANF (Section A) and Child Support Enforcement (Section B). It then describes changes in other policies and programs that could potentially affect the structure of low-income families, including health care, housing assistance, child care, and tax policies (Section C). The chapter concludes with some thoughts about implementing and testing these policy changes (Section D).
A. TANF
About two million families now receive TANF benefits. Hence, any disincentives to family formation in TANF could potentially have large effects. Changes in TANF to mitigate these disincentives and encourage family formation are described below.
1. Remove Categorical Eligibility Requirements
Aid to Families with Dependent Children (AFDC) was designed to serve needy children in one-parent households. At the program’s inception, a two-parent family was categorically ineligible for welfare even if the father was unemployed and the family needy. The law was changed in 1961 so families with jobless fathers could be eligible for the AFDC-Unemployed Parents (AFDC-UP) program. States could choose how to define “unemployed,” but federal regulations required that a parent must work less than 100 hours a month to be classified as unemployed. To be eligible, the parent also must have a significant history of employment, and the family must meet income and asset requirements.
PRWORA allows states to remove these restrictions on TANF eligibility for two-parent families, treating one- and two-parent families the same when determining eligibility, thus reducing the disincentive to form two-parent families. As of July 2000, 36 states have done so.1 However, 10 states still have at least one of the AFDC restrictions for two-parent eligibility, and 3 retain all of the AFDC requirements. North Dakota denies eligibility to two-parent families unless one parent is incapacitated.
2. Disregard Some or All of the Spouse’s or Cohabiting Partner’s Income
The main disincentive in TANF to form a two-parent household is that the income of a second adult may count against the family in determining TANF eligibility and benefits. The states have a wide degree of flexibility in determining who is in the assistance unit.
In many states, the income of both biological (or adoptive) parents living with their child is counted in TANF eligibility and benefit decisions regardless of their marital status. As parents need to live with the child for their income to be counted, there is a disincentive for a TANF mother and child to live with the father of the child if the father has income or prospects for income.
TANF also contains structural disincentives to marriage for couples that do not have a child in common. Many states do not count the income of a cohabiting partner who is not the biological parent of a child in the family. However, some TANF agencies “deem” income from a married stepparent to the mother and child when determining eligibility and benefits.
The rules also may create a disincentive for cohabiting couples to marry even if the eligibility rules do not depend on the couple’s marital status. This is because an unreported cohabiting partner is unlikely to be detected by the TANF caseworker, but it is harder to “hide” a spouse. So some couples may decide not to marry and not to report the existence of the cohabiting partner so that his income will not be counted.
Just as disregarding some earnings encourages work, disregarding some or all of the income of the second adult in the family in determining eligibility and benefits could encourage the formation of two-parent families. To encourage marriage over cohabitation, states could count income in households with married parents differently from income in unmarried two-parent households. Not deeming income of the stepparents to the biological mother and child could encourage marriage between the mother and a potential stepparent.
Several states are experimenting with ways to create marriage incentives through policies that disregard the income of spouses. For purposes of calculating benefit levels, four states (Alabama, Mississippi, North Dakota, and Oklahoma) disregard all the income of a new spouse for three to six months. Two other states (Tennessee and New Jersey) disregard a stepparent’s income if the household meets certain income criteria. Maine offers the option to include or exclude stepparents in the TANF assistance unit (Gardiner et al. 2002).
3. Provide Financial Bonuses for Marriage
Some TANF agencies have explored providing financial bonuses as an incentive to marry or remain married. West Virginia adds $100 to the monthly TANF benefit payments to married couples. Legislatures in Mississippi and Washington attempted unsuccessfully to establish programs to pay parents on cash assistance a lump sum payment if they remain married for at least a year (Gardiner et al. 2002). To promote marriage, the Torres Martinez Desert Indian Consortium provides Native American TANF recipients in California’s Riverside County and Los Angeles County a lump sum of $2,000 if they participate in a marriage promotion program. In addition, $1,500 is available to offset the costs of the wedding as long as the participant has a traditional Native American wedding.
Providing financial bonuses for newly married TANF couples and disregarding the income of a new spouse in determining TANF benefits may reduce the financial disincentives to marry by the same amount. If the two changes are perceived differently, however, they may have different effects on family formation. It is easier to communicate to TANF recipients how a financial bonus for marriage alleviates the disincentive to marry. It may be more difficult to communicate to recipients how a change in the treatment of a new spouse’s income will reduce this financial disincentive.
Policies that provide financial bonuses for marriage might have the unintended consequence of leading to more unhealthy marriages or “paper” marriages, which occur only so couples can receive financial incentives. However, most financial bonuses are small and are designed only to lessen the financial disincentive to marry rather than to provide direct financial incentives to marry. The unintended consequence may occur, however, if couples perceive in these bonuses a substantial financial reward for marrying.
4. Ease Work Requirements on Two-Parent Families
TANF policies have different work or work-related requirements for one- and two-parent families. Single parents with a child younger than six must work at least 20 hours, and other single parents must work at least 30 hours a week. Parents in two-parent families must work 35 hours, or 55 hours if they receive federally funded child care, but the parents can share the work hours. States are required to meet a minimum work participation rate of 50 percent for all families and a 90 percent rate for two-parent families. Some have argued that the separate work requirement for two-parent families should be eliminated (Fremstad and Primus 2002). In their welfare reauthorization proposals, the Bush administration and the U.S. House of Representatives would require 40 hours per week of work participation for all families, ending differential requirements for single- and two-parent families (CLASP 2002).
5. Provide Financial Security As Welfare Recipients Move Into Work
Some have argued that increasing financial security as welfare recipients move into work may encourage parents to marry and help them stay married (Knox et al. 2000). The main way to increase financial security is to increase earned income disregards so that welfare recipients’ income (earnings plus cash assistance) increases more as they begin to work. This argument is supported by the findings from an experimental evaluation of Minnesota’s welfare reform program—the Minnesota Family Investment Program (MFIP)–in which positive results were found on marriage within a group of long term welfare recipients, though not applicants. The evaluation found that 11 percent of MFIP recipients were married at the end of the third follow-up year compared with 7 percent of AFDC recipients (Knox et al. 2000). And among families that reported a spouse or cohabiting partner when they entered the study, 67 percent of MFIP families reported being married at the end of the third year compared with 48 percent of their AFDC counterparts. Although MFIP and AFDC differed in many ways, one major difference is that MFIP had a higher earnings disregard for calculating eligibility and benefits. MFIP increased family income by increasing both earnings and, because of the higher earnings disregard, benefits.
B. CHILD SUPPORT ENFORCEMENT
Child support enforcement may have unintended negative impacts on family formation. Several policy changes have been suggested to strengthen the incentives for family formation and remove unintended disincentives (McLanahan et al. 2001): (1) strictly enforce child support, (2) inform unwed fathers of their potential child support obligations, (3) align child support obligations with the father’s ability to pay, and (4) reduce the amount of child support retained by the government. The first two policy changes directly affect the incentive to marry. The third and fourth policy changes may encourage family formation by reducing the burden on noncustodial fathers.
1. Strictly Enforce Child Support
Theoretically, stricter enforcement of child support has an ambiguous predicted effect on family formation. By imposing some of the cost of bearing and raising children on fathers who do not live with their children, stricter child support enforcement may discourage fathers from leaving the family. On the other hand, by providing the custodial parent with another source of income (assuming she receives the additional income) and greater financial independence if unmarried, stricter child support enforcement can weaken the incentives for the mother to stay married or cohabit with the father of her biological children.
Empirically speaking, stricter child support enforcement is generally found to promote family formation (Carlson et al. 2002). Overall, findings from research on the impact of strong child support enforcement on family formation suggest it generally encourages families to form or stay together or to avoid having a child out of wedlock. States with stricter enforcement have lower rates of divorce and out-of-wedlock births than do states with looser enforcement (Nixon 1997; Case 1998; Garfinkel et al. 2002). Mincy and Huang (2001) found that in states with more effective child support collection for children on TANF, it is more likely that the mother marries her children’s father.
2. Inform Unwed Fathers of Their Potential Child Support Obligations
Another policy suggestion is to require paternity establishment for all unwed fathers even if the fathers live with the mothers and children, and to inform them of their potential child support obligations (McLanahan et al. 2001). Even though cohabiting fathers would not be required to pay child support, an awareness of their potential child support obligations may provide an incentive to stay with their families.
3. Align Child Support Obligations With the Father’s Ability to Pay, and Forgive Some Arrearages
Low-income fathers often have to pay a much higher proportion of their income in child support than do middle- or upper-income fathers (Carlson et al. 2002). This is because child support orders typically are based on the fathers’ presumed earnings based on what they earned in the past or would earn in full-time jobs at the minimum wage rather than on what they actually earn. In some states, fathers are required to reimburse Medicaid for their children’s birth immediately after delivery. If fathers are unemployed, underemployed, and/or incarcerated, as many fathers of low-income children are, they are unlikely to be able to pay the support and can incur huge arrearages (Sorenson et al. 2000). Fathers may be incarcerated for nonpayment of support, putting them further into debt when they are released. In addition, they may have children with more than one partner and face multiple child support orders.
Large child support obligations and arrearages may both contribute to the tension between parents and push fathers away from their families. The fathers often are overwhelmed by the amount they are required to pay and may resent the mothers and children. The mothers may perceive the fathers as being neglectful, uncooperative, and contributing too little to their families (Carlson et al. 2002, Sorenson et al. 2000; McLanahan and Garfinkel 2002).
One suggested policy change is to set child support obligations at a flat percentage of the father’s income (McLanahan et al. 2001). The advantage of this change is that obligations would automatically decline when the father is not working. But as actual support payments have been found to increase when support orders are expressed as a percentage of income (Bartfeld and Garfinkel 1996), the child may not receive substantially less financial support. This would, however, create somewhat of a disincentive for the father to work and would require a system in which changes in the father’s income are reported and verified.
Some states have forgiven arrearages under certain circumstances. Tennessee forgives a father’s arrears if he marries the mother of his children and lives with the family. Vermont forgives arrears if the biological parents reunite, although it does not specify that they marry (Gardiner et al. 2002). Other states have programs to lessen or forgive child support arrears if noncustodial parents participate in employment, fatherhood, or other programs designed to improve their earning potential and involvement as parents and if they make efforts to begin paying their debt. Maryland’s Child Support Arrears Leveraging Program is one example of this type of program.
4. Reduce the Amount of Child Support Retained by the Government
A large proportion of support payments to TANF recipients typically does not go to the children and mothers; it is retained by the state and federal governments to compensate for the costs of cash assistance. More than half the states retain all child support that is collected, and 15 pass through a maximum of $50 to the children and mothers (Center for Law and Social Policy 2002). Even after families leave welfare, about half of child support debt collected for former TANF recipients is kept by the government to cover arrearages incurred while the family was on assistance (Turetsky 2002). The low share of child support payments passed to the mother may decrease the father’s incentive to make payments and stay involved with the family. Increasing the amount of child support passed through to the family would allow the father to see improvements in his children’s resources and may encourage him not only to cooperate with the system but also to become more involved with his family.
Reducing the amount of child support retained by the government may, however, have a negative effect on the incentives for a couple to marry. If the government retains some child support, more of the father’s contribution to the family would be available to the child if the couple were married. This financial incentive may partly offset the reduction in TANF benefits that occurs if the couple marries. Several states have moved to pass through significant amounts of child support to the families and to disregard some or all of the child support payments as income for purposes of determining eligibility for cash assistance and benefit payments. For example, Connecticut passes through all collected child support to the families and disregards up to $100 per month for benefit calculations. In Wisconsin, members of a large experimental group receive all child support, and the full amount is disregarded for benefit calculation, while control group members receive only a $50 pass-through that is disregarded for eligibility and benefits.
Research shows that reducing the amount of child support retained by the government has some positive results. Findings from the Wisconsin study suggest that families in the experimental group who could receive the full amount of support have higher rates of paternity establishment and are more likely to receive child support (Meyer and Cancian 2001). Most measures of the nonresident father’s relationship with the mother and child revealed few differences between the experimental and control groups, although there was some evidence of higher informal transfers made by fathers in the experimental group. Analysis of data on paternity establishment across states also suggests that a higher pass-through is associated with higher paternity establishment rates (Meyer and Cancian 2002).
C. OTHER PROGRAMS AND POLICIES
The requirements of a range of other programs and policies have implications for family formation and suggest possible interventions to be tested.
1. Expand Health Care Coverage for Two-Parent Families
After Medicaid eligibility was delinked from eligibility for cash assistance in PRWORA, federal policy required states to create a “family coverage” category, which applies the old AFDC eligibility rules for two-parent families to the determination of Medicaid eligibility. But states may take steps to expand eligibility, including eliminating the 100-hour and recent work history rules and providing Medicaid coverage to new categories of people through Section 1115 waivers. States also may disregard certain income and assets, thereby expanding eligibility. As of 2000, 36 states based Medicaid eligibility entirely on a two-parent family’s financial circumstances, and 8 used waivers to cover two-parent families (Gardiner et al. 2002).
Despite these changes, some two-parent families become ineligible for coverage if their combined income pushes them past the threshold for eligibility. To address this situation, three states provide insurance for parents, including those in two-parent families, through state-funded programs, and six have received waivers from the State Children’s Health Insurance Program (SCHIP) to cover parents (Gardiner et al. 2002).
2. Disregard Spouse’s Earnings in Determining Housing Assistance Eligibility and Benefits
Housing assistance is available for low-income families from federal, state, and local programs. Rent subsidy programs generally reduce tenants’ rent payments to a fixed percentage of their income after deductions, with the government paying the rest. The federal rental subsidy limits the rent payments to 30 percent of family income.
As with other means-tested programs, housing benefits can fall substantially if a working spouse joins the family because all of the additional income is counted. Rector (2002b) reports that a typical single mother receives a housing subsidy worth $5,000. But if she marries a man with earnings of $18,000 or more, she will lose the subsidy. Fear of losing housing assistance has been found to be an important disincentive to increasing earnings through work (Miller and Riccio 2002) and may also be an important disincentive to marry.
One suggested policy is to ignore the first $1,000 of a spouse’s earnings in determining the married couple’s eligibility and rental subsidy (Rector 2002b). A public housing agency could, at its own cost, adopt a policy disregarding some or all of the income of a spouse who joins a family in public housing, although it currently is not permitted to do so for families receiving Section 8 vouchers to help pay for private housing (Sard and Waller 2002).
3. Reduce Any Disincentives to Family Formation Inherent in Child Care Policies
Eligibility rules for child care subsidies may unintentionally discourage two-parent family formation. Families are only eligible for child care subsidies under the Child Care and Development Block Grant if their income does not exceed 85 percent of the median income for families of their size in the state. Many states set income eligibility levels below this. As income eligibility cutoffs are low, working mothers may easily lose their child care subsidy if another working adult joins the family.
4. Reduce the Marriage Penalty in the Tax System
Income taxes are based on the individual’s income, or, if married, on the couple’s income. Depending on the distribution of income between the spouses, taxes could contain marriage penalties or bonuses. According to the U.S. General Accounting Office (1996), there are 59 provisions in the U.S. income tax code that either penalize or reward marriage.
The earned income tax credit (EITC) is of particular importance to low-income populations. The EITC is a refundable tax credit that increases with a filer’s earnings until it reaches a maximum. Over a range of income, taxpayers receive the maximum credit, and then it is phased out with additional earnings over a specific amount. The EITC can discourage marriage if the additional earnings of the spouse would reduce the amount of EITC benefits or make the couple ineligible for benefits. On the other hand, the EITC may encourage marriage if a woman with no earnings marries a man with low earnings and the couple becomes eligible for the EITC. Ellwood (2000) found that the EITC declines an average of $1,505 for 16 percent of couples after marriage and increases by an average of $1,367 for 11 percent of couples. However, he found no clear effects of the EITC on marriage patterns, and Dickert-Conlin (1999) found only modest and somewhat conflicting effects on divorce.
Changes in the tax system are clearly out of the purview of ACF. Nonetheless, states could change the marriage penalties/rewards in their own tax systems. Fifteen states and the District of Columbia had state EITC programs in place in 2000. Ellwood and Sawhill (2000) suggest several options for reducing the marriage penalty in the EITC, including making tax credits more universal or extending them up the income scale, allowing married couples to file separately and to split their income for EITC purposes, creating an EITC earnings deduction, and reducing the phase-out rate for married couples.
D. IMPLEMENTING AND TESTING POLICY CHANGES
The advantages of encouraging marriage through policy changes are that the changes can potentially affect a large population and, unlike the other interventions discussed in this report, will not require couples to agree to receive services.
The major limitation of this form of intervention, however, is that the policy changes do not affect relationship skills or change the attractiveness of a partner as a spouse. On their own, these policy changes may not affect the likelihood of “healthy” marriages nor have lasting effects. In addition, the magnitude of the effect of policy changes is uncertain. The effects may be small if unwed parents do not perceive the change in incentives or if the change is not large enough to change behavior.
In implementing the types of changes suggested in this chapter, policymakers will face three major challenges. First, the changes may be very costly. Removing disincentives to form families will increase the number of two-parent families that are eligible for assistance and the amount of benefits they receive. However, if the policy changes are successful and more two-parent families form and become self-sufficient, these costs will be offset, at least to some extent, by a reduction in the number of families needing assistance. Second, removing disincentives to form two-parent families will involve increasing the proportion of benefits paid to two-parent families. But since two-parent families generally have higher incomes and benefit from the economies of cohabiting, these policy changes will mean that the additional resources will not be targeted to the most needy families. Third, while disincentives inherent in the entire tax and public assistance system may deter couples from marrying, the disincentives in any one program may be small. Therefore, changes to multiple programs may be needed to achieve a significant impact on marriage, and this approach would require the cooperation of many different state agencies.
1The two-parent family eligibility rules were obtained from the Urban Institute's The Welfare Rules Database. (back)
| Table of Contents | Previous | Next |