SECTION: CAPITOL HILL HEARING TESTIMONY

LENGTH: 10062 words

COMMITTEE: HOUSE WAYS AND MEANS

SUBCOMMITTEE: HUMAN RESOURCES

HEADLINE: TESTIMONY WELFARE OVERHAUL EFFECTS

TESTIMONY-BY: MARK GREENBERG , SENIOR STAFF ATTORNEY

AFFILIATION: CENTER FOR LAW AND SOCIAL POLICY

BODY:
Statement of Mark Greenberg, Senior Staff Attorney, Center for Law and Social Policy Testimony Before the Subcommittee on Human Resources of the House Committee on Ways and Means Hearing on Welfare Reform March 15, 2001 Mr. Chairman and Members of the Subcommittee: Thank you for inviting me to testify. I am a Senior Staff Attorney at the Center for Law and Social Policy (CLASP). CLASP is a nonprofit organization engaged in research, analysis, technical assistance and advocacy on a range of issues affecting low income families. Since 1996, we have closely followed research and data relating to implementation of Personal Responsibility and Work Opportunity Reconciliation Act. In addition, we often talk and visit with state officials, administrators, program providers, and individuals directly affected by the implementation of welfare reform efforts. Today's hearing focuses on the experience since 1996 in addressing the four goals of the Temporary Assistance for Needy Families block grant structure. These goals are to: (1) provide assistance to needy families so that the children may be cared for in their homes or in the homes of relatives; (2) end the dependency of needy parents on government benefits by promoting job preparation, work, and marriage; (3) prevent and reduce the incidence of out-of-wedlock pregnancies and establish annual numerical goals for preventing and reducing the incidence of these pregnancies; and (4) encourage the formation and maintenance of two-parent families. I want to begin with two general observations. Since 1996, there has been a large research effort funded both by governments and private organizations, and much is known as a result of this effort. This research makes it possible to discuss what has happened, i.e., changes in caseloads, employment, income, family structure, but it is more difficult to state what role the 1996 law or particular components of the law played in affecting these changes. Second, the fact that the law is scheduled for reauthorization next year makes this a logical time to consider the experience; at the same time, this is still a relatively early point in a major national policy shift, and to date, we've only seen the experience in the context of a strong national economy, and before families have reached the five-year time limit restricting federally funded assistance. So, while it is important to look at the story so far, it is also important to recognize that more time will be needed to know how the new structure will function over time. Since the law was enacted, both the TANF assistance caseload and the nation's child poverty rate have fallen significantly. However, participation in TANF assistance has fallen much more rapidly than child poverty has declined. Since 1994, there has been a historically unprecedented decline in the number of families receiving assistance. In early 1994, five million families were receiving AFDC. The number fell to 4.4 million by the time the 1996 law was enacted, and then dropped to 2.2 million by June 2000. Since enactment of the law, the caseload has fallen by at least 50% in twenty-nine states and by at least 20% in all states. Caseload decline is sometimes cited as evidence of success in itself, but a state's caseload can fall either because families no longer need assistance or because families who need assistance are not receiving it. Part of the caseload decline is clearly due to reduced need - during this period, child poverty fell, from 21.8% in 1994 to 16.9% in 1999. However, participation in welfare fell much more rapidly than child poverty. In 1994, 62% of poor children were receiving AFDC assistance; by 1998, only 43% of poor children were receiving TANF assistance. Why did the share of poor children receiving assistance fall? There are three main possibilities: a) parents got jobs and stopped receiving TANF, even though family income was still below poverty; b) families left assistance without finding work; and c) families who were poor and potentially in need of assistance chose not to apply or applied and were unable to get assistance. As discussed below, there is evidence that most families leaving welfare are working, but that many of these families remain poor; there is also evidence that a significant group of families have left welfare but are not working. It is less clear whether there has been a drop in applications or application approval rates. Under TANF, states are required to report to HHS on the total number of applications, denials, and approvals, but not the reasons for denials. In any case, none of this information about application processing under TANF has been published. Since 1996, there has been a significant increase in employment among female-headed families. There's broad agreement that TANF has played an important role, but is not the only factor, in these increases in employment. A set of data sources all point to a significant increase in employment among TANF recipients, and more generally among female- headed families. The principal information about why families have left welfare has come from studies looking at the circumstances of families who have left assistance. These studies consistently find that about 60% of leavers are working, and that employment is the most common reason given for why families left assistance. Typically, an even larger share report having worked at some point since leaving assistance. The share of adults working while receiving TANF assistance also increased - from 8% in 1994 to 28% in 1999. Part of the explanation for this increase in employment among recipients is that under TANF, most states have liberalized the rules for reducing assistance when a family has earnings; as a result, a family is more likely to retain eligibility when a parent enters a low-wage job. Census Bureau data also point to a large increase in employment among female-headed families in recent years. In 1994, married mothers were more likely to be employed than were single mothers. (64.7% vs. 57.1%). By 1999, the employment rate of single mothers (68.4%) was greater than the rate for married mothers (67.1%). The differences are more dramatic among low income families. For families with incomes below 200% of poverty, between 1994 and 1999, employment rates stayed essentially flat for married mothers but grew by thirteen percentage points for single mothers. As a result, in 1999, 59% of single mothers in families with incomes below 200% of poverty were working, versus 43.4% of married mothers. It is probably impossible to isolate the independent role of TANF in this increase in employment. The process started before 1996 - the nation's caseload decline began in 1994, and the growth in employment of low income single mothers with young children began between 1992-93. And, during the 1990s, a set of factors supported or contributed to employment growth for single mothers: the strong national economy, the large expansion of the Earned Income Tax Credit, increased availability of child care subsidies, expansion of health coverage for children, the minimum wage increase, improved child support enforcement. There seems to be a consensus among researchers that welfare reform efforts did play an important role, with the effects more pronounced in latter years. At the same time, a set of factors occurring at the same time all pushed in the same direction, and we don't know how the same policies would have worked in a different economy, or how one component would have worked without the others. When we talk about the effect of "TANF" or "welfare reform," it is important to appreciate that there are a number of components in what states have done under TANF. In formal rules, states generally expanded requirements to participate in work-related activities; increased the penalties for failure to comply with such requirements; restricted access to education and training; provided increased income support for families with earnings; liberalized program asset requirements; broadened eligibility for two-parent families; imposed time limits on assistance; and expanded the availability of "transition" benefits for families leaving assistance. In addition state agencies often emphasize that there has been a fundamental change in the basic orientation of their systems, as the principal focus has shifted from providing income support to an emphasis on requiring and supporting employment. TANF implementation also meant an infusion of additional funds for states. Since funding levels were generally set to reflect welfare caseloads from the early-mid 1990s, and caseloads began falling in 1994, the effect of TANF from the beginning was to provide increased federal funding to states. The General Accounting Office estimated that if all states had participated in TANF throughout 1997 and had met applicable maintenance of efforts, states would have had an additional $4.7 billion above the funding level that they would have had under prior law. And, since block grant funding remained constant as caseloads declined, the funds potentially available for services steadily increased over time. This additional funding made it possible for states to expand employment-related services, child care, and support services. In summary, TANF has contributed in important ways to the increase in employment among female-headed families. However, it is not possible to precisely state its independent impact; we do not know how the same policies would operate in different economic circumstances; and we need to recognize that TANF implementation has been far more than work requirements and time limits - it has also involved additional resources and an array of new and expanded supports to help families enter and maintain employment. Much of the employment for families receiving or leaving TANF assistance, at least initially, is in low-wage jobs. There is evidence of some earnings growth over time, but so far, earnings remain low for most of the affected families. Families still receiving assistance tend to have low earnings - according to administrative data, earnings for working adults receiving assistance averaged $597.97 per month in FY 99. Leavers studies also report generally low wages and earnings for those who have left welfare and entered employment. According to the Urban Institute's Nation Survey of America's Families, the median wages for leavers in 1997 were $6.61 an hour; individual state studies typically report wages at or in that range. In a set of leavers studies funded by HHS, the median quarterly earnings for the first full quarter after leaving were $2526, i.e., $842 a month. So, unless these families have additional sources of income, they are often likely to still be in poverty. In fact, studies in Missouri and Washington reported poverty rates of 58% for TANF leavers. The fact that many families are entering low wage employment was not unexpected, because a strong focus in TANF implementation was to encourage parents to enter employment as rapidly as possible, even at low wages, with the hope that earnings would grow over time. To find out if this is occurring, one would want to follow the experiences of families over time. Unfortunately, state reporting to the federal government provides only limited longitudinal data. To compete for the federal "high performance bonus," states report earnings data for families during a quarter and the second subsequent quarter for people employed in both quarters. From this data, we can see that in 1999, national average earnings grew from $2114 in a quarter to $2578 in the second subsequent quarter. However, since the first quarter includes both individuals continuously employed and those entering employment during the quarter, we cannot tell the extent of earnings growth from this data, and we cannot get a longer- term picture from this data. A limited number of states have reported longitudinal data for families leaving welfare and entering employment. From that data, it looks like earnings do increase after leaving assistance, but remain low. For example, in nine federally-funded leavers studies, median earnings grew from $2526 in the first to $2821 in the fourth quarter of employment. Probably the best longitudinal data comes from the Institute for Research on Poverty's tracking of families that left welfare in Wisconsin in 1995. In the IRP study, median earnings (in 1998 dollars) were $8608 in the first year after leaving, $9627 in the second year, and $10,924 in the third year. Some states have responded to evidence of low earnings for families leaving assistance by creating new initiatives to help working families who have left assistance advance to higher earning jobs. Most of these initiatives are still in the earliest stages, and there isn't yet clear evidence of their effectiveness. There are also indications that in the last several years, a number of states have softened some of their prior restrictions on access to education and training programs, though participation in such activities by TANF recipients remains low. The fact that many exiting families have low earnings has focused attention on the importance of access to Food Stamps, Medicaid, child care assistance and child support services for families leaving assistance. Studies consistently report sharp declines in participation in Food Stamps and Medicaid after families leave assistance. Probably not more than one-third of working leavers receive child care assistance. Child support enforcement has improved, though most leavers still do not receive child support. Because many leavers have low incomes, continued access to Food Stamps and Medicaid are important for families leaving TANF. However, a consistent finding has been a sharp drop in receipt of these benefits after exiting TANF. In a set of state and county leavers studies funded by HHS, the share of single parents receiving food stamps in the quarter after leaving TANF ranged from 33% to 57% (and was 9% in one county). The share of adults receiving Medicaid in the quarter after leaving was between 41% and 57% in seven of nine sites. There are indications that these drops in participation are partly due to eligibility rules, partly due to lack of awareness of continued eligibility, partly due to administrative practices that sometimes resulted in closing Food Stamp and Medicaid cases at the same time TANF cases were closed. There are also indications that families who could benefit from child care assistance after leaving welfare do not consistently receive such assistance. In state leavers studies with data on this issue, only about one third or less of working leavers were receiving child care subsidies. However, the Urban Institute has calculated that 50% of working low-income single parents (income below 200% of poverty) with children under age 13 have child care costs, averaging $230/month, and representing 19% of family income. Federal administrative data indicate that some child support is collected for 40% of welfare leavers. This probably reflects significant improvement in recent years, though a substantial share of leavers are still not receiving regular child support payments. (Moreover, about half of support arrears collected on behalf of leavers are not paid to the family, but instead are retained by government to recover prior assistance costs.) Families still receiving assistance are a heterogenous group, but generally have more serious barriers to employment than those who have left assistance. Most families now receiving assistance are either working or do not include an adult receiving assistance. In FY 99, 28% of adults receiving assistance were working, and 29% of families receiving assistance were "child-only" cases, i.e., cases in which no adult received assistance because the child is living with a grandparent or other non-parent relative; the parent is disabled and receiving Supplemental Security Income, the parent is an ineligible alien or under sanction. State and local agencies often note that many of the families still receiving assistance are likely to have serious employment barriers. It is difficult to provide a precise picture, because the types of problems often identified - health and mental health, domestic violence, substance abuse, limited or no English proficiency, severe basic skills deficits - are not the types of information routinely collected or reported in program administrative data. The Urban Institute reports that among adults receiving assistance in 1999, 44% had less than a high school education, 38% either had poor health, a health condition preventing work, or very poor mental health; and 27% last worked three or more years ago. Most (56%) of those with no identified barrier to work were working; in contrast, only 20% of those with two or more barriers to work were working. People sometimes ask if the incidence of these problems is more severe now than was the case in the past. It is difficult to know, because most states have not had consistent measures of these potential obstacles over time. The Urban Institute found that the incidence of barriers in 1999 did not look substantially different from 1997. It is clear, though, that families still receiving assistance are much more likely to have multiple barriers to employment than families who have left. State leavers studies also consistently report that those still receiving assistance are likely to have less education and less work history than those who have left. And, for many states, the key question isn't really whether the prevalence of these problems has increased, because five years ago, families with multiple barriers were typically exempt from program requirements. Now, these families are subject to work and time limit requirements, and states need to develop appropriate service strategies to work with them, so the issues of how to structure services for such families have become an increasingly significant challenge in TANF implementation. A group of families with serious barriers to employment is no longer receiving assistance. About 40% of families who have left assistance are not working. Generally, there is less information about these families, and it would be helpful to have a clearer picture of why they left and how they are managing. On one hand, families have always left assistance for reasons besides work -- for example, the family might move, or children might turn eighteen. And, nonworking leavers are sometimes living with spouses, partners, or other adults. However, there are indications that, as compared to working leavers, the nonworking leavers are likely to have more serious employment barriers -- e.g., less education and work history, greater likelihood of illness or disability. They are also likely to be poorer than working leavers. In a recent analysis, the Urban Institute focused on one group of nonworking leavers: those adults who had not worked since leaving, were not receiving disability benefits, and were not residing with a working spouse or partner. This group of "at risk" leavers was estimated to be 17% of all leavers, and they showed a greater likelihood of having multiple barriers to employment than current TANF recipients 50% of this group was in very poor health; 47% had not worked in at least three years; 38% had less than a high school education; 19% were caring for a disabled child. Why are families with such severe barriers leaving assistance despite not working? There is a clear need for additional research, but one part of the explanation is likely the increased use of "sanctions," i.e., grant reductions and terminations and other case closures for noncompliance under TANF. States vary in their extent of sanctioning; in some leavers studies, only a small share of case closures are attributable to sanctions and in other studies, 25% or more of case closures are for such reasons. Studies of sanctioned families consistently report that such families are likely to have less education and work history than other leavers, and that they are less likely to be working and more likely to have lower earnings when they do work. The concern, then, is that at the same time that many states articulate a goal of working with those families with the most severe employment barriers, state practices are sometimes having the effect of terminating assistance to those families. Some states have responded to these concerns by developing "second look" procedures before imposing sanctions or additional efforts to engage in outreach to sanctioned families. However, at this point, the TANF law does not actively encourage such efforts, and the federal government does not collect information on state practices to avoid terminations of assistance to families with multiple employment barriers. The large gains in employment have resulted in increased income for many female-headed families; at the same time, the average incomes of the bottom 20% have declined since 1994, because losses in public benefits have been as large or greater than gains in earnings. Researchers have examined Census data during the 1990s and examined the circumstances of female-headed families. Typically, this work has looked at post-tax, post-transfer income, taking into consideration both the Earned Income Tax Credit and food stamp benefits. The research, summarized in recent interchanges between Wendell Primus and Ron Haskins, essentially finds that when one divides female-headed families into five quintiles, there have been both earnings and income gains in the second- lowest quintile, though the rate of gain was stronger from 1993- 95 than from 1995-99, and most of the earnings gains from 1995-99 have been offset by declines in public benefits. However, for the bottom quintile, average income in 1999 is actually lower than it was in 1994 and 1995, because losses in means-tested income have been greater than increases in earnings. (While 1999 income is lower than 1994 income, it is higher than 1993 income; however, the key point here is that whatever base year is used, the story is that for the bottom quintile, gains in earnings have been more than offset by losses in public benefits.) Why have the earnings gains not translated to greater economic gains for families? While the Census data cannot directly answer this question, the findings seem consistent with those already described from the leavers' studies: many families entering low wage employment have lost public assistance despite still being in poverty; Food Stamp participation has sharply declined for families leaving TANF; and a group of families has left TANF without finding work. These findings are particularly striking in light of recent work by the Manpower Demonstration Research Corporation, looking at experimental studies to examine the impact of work-welfare initiatives on children. MDRC found that while many programs raised employment rates, only some raised income, because gains in employment were often offset by losses in benefits. In those programs where employment was associated with increased family income, MDRC found evidence of positive effects on childrens' school achievement. In addition, two of three programs that raised incomes found increases in positive behaviors by children, and one found a decline in problem behaviors. In contrast, programs in which increased employment did not translate to increased incomes had mixed effects, and no clear positive impact on children's' well-being. The authors conclude: " W e found that programs that provided earnings supplements had consistently positive impacts on children's achievement... Raising employment without increasing income may not be sufficient to boost the healthy development of children in low-income families." Morris, et. Al, How Welfare and Work Policies Affect Children: A Synthesis of Research (Manpower Demonstration Research Corp. 2001). The TANF caseload decline has made it possible for states to make a major redirection of resources to child care. The freed-up resources have also allowed for significant program expansions in other areas, though the existing federal reporting system does not provide a good picture of how funds are being used. In 1996, TANF funding was set to reflect historic federal funding levels, and to remain essentially constant through 2002. The decline in TANF caseloads meant that federal and state funds became available for redirection to other services and activities to further the goals of TANF. Initially, many states were hesitant to redirect TANF funds, because they understood that federal funding would remain fixed even if caseloads began rising, they were uncertain whether the caseload decline would continue, and they were often unclear about when and how TANF funds could be used for expenditures outside of the traditional welfare system. A key shift occurred when HHS issued final TANF regulations in April 1999. These regulations made clear the breadth of state flexibility in using TANF, explaining that states could use the funds to benefit "needy families" whether or not those families were receiving traditional welfare assistance and that states could use TANF to structure supports for working families outside the traditional welfare system. In exercising this new flexibility, the clearest response has been in committing resources to child care. States can transfer up to 30% of their TANF block grants to their programs under the Child Care and Development Block Grant, and can also spend TANF funds directly for child care. Each year, the amount of TANF funding redirected to child care has grown; in 2000, we calculate that $3.5 billion in TANF funding was redirected to child care, an amount as large as the entire federal child care block grant. This additional funding has allowed states to increase the numbers of families receiving subsidy assistance, raise eligibility levels, improve payment rates to child care providers, and expand spending for child care quality initiatives. At the same time, state administrators often express apprehension about the extent to which child care expansions have relied upon a source of funding that is not seen as stable or predictable. The expanded funding has made an important impact, there are still significant indicators of unmet needs. As noted above, most working leavers do not receive child care subsidies. There is no available national data on the share of TANF children receiving child care assistance; a study by the National Center for Children and Poverty and Abt Associates reported that in 1999, the share of TANF children receiving child care subsidies in ten states ranged from 7% to 26%. (An eleventh state reported a higher figure, but there appear to be data problems affecting the reliability of that figure.) And, HHS has estimated that nationally, 12% of potentially eligible children were receiving subsidy assistance under the Child Care and Development Block Grant in 1999. While the numbers receiving assistance would be higher if all funding sources were considered, the basic picture of unmet need would remain. Moreover, resource constraints in child care are expressed in a range of ways: in state eligibility levels, fee scales, payment rates to providers, limited outreach to eligible families. For example, state child care payment rates are important in determining whether families receiving subsidies have equal access to child care arrangements comparable to families above CCDBG income levels, as envisioned under federal law. The federal government has said that a payment rate high enough to encompass 75% of the local market would be considered sufficient to provide equal access. However, according to the Children's' Defense Fund, in 1999, most states were not making payments that met this standard based on a recent (within two years) market survey. Moreover, there continue to be concerns about high staff turnover, compensation and training issues for child care providers; shortages of care for particular populations and needs (e.g., infant care, special needs care, sick care, non-traditional hour care); and about the uneven quality of care in a range of settings. In other areas, TANF funding has made it possible for states to undertake initiatives such as creating or expanding refundable Earned Income Tax Credits, initiating programs of Individual Development Accounts to promote asset formation, expanding services for victims of domestic violence and for families in need of substance abuse and mental health treatment, increasing services for non-custodial fathers, expanding "diversion" assistance for families in need of short-term, emergency help, and increasing services for family preservation and reunification. However, there are also accounts of states that have used some of their TANF funds to "supplant" state spending, i.e., to use TANF funds to replace existing state expenditures for allowable activities under the TANF law. Unfortunately, the current federal reporting structure does not provide a clear picture of TANF spending; state plans often provide only cursory detail about state spending choices, and the broad categories in federal spending reports are not very informative. Accordingly, one key challenge for researchers and for the structuring of federal reporting involves efforts to better describe the choices states are making in their use of funds. During the 1990s, teen birth rates declined and the share of children born out of wedlock appeared to level off, though remaining at about one-third of all births. These changes began before states implemented TANF. In advancing the law's family formation goals, there is a need for both a stronger research agenda and a recognition of a set of difficult issues about the appropriate role of government. Between 1991 and 1999, the nation's teen birth rate dropped 20%. The teen birth rate in 1999 was 49.6 births per 1000 women aged 15-19, a record low for the nation, though still high in comparison to many other countries, and still involving nearly 500,000 births to teenagers. The decline in the rate of teen births is significant, both because of concerns about outcomes for children born to teenagers and because nearly 80% of teen births are out of wedlock births. As in other areas, it is difficult to determine the role of TANF in contributing to the decline in teen births. On one hand, the trend began well before implementation of the 1996 law. However, almost all states are funding teen pregnancy prevention initiatives using TANF funds, and TANF funding has made it possible for states to create or expand after-school programs, pregnancy prevention programs, stay-in-school programs, and an array of youth development initiatives aimed at reducing teen pregnancies and promoting stronger outcomes for low-income teens. During the 1990s, the birth rate for unmarried women declined slightly, and the share of births to unmarried women appears to have flattened. At the peak year, 1994, the birth rate for unmarried women aged 15-44 was 46.9 per 10000; preliminary data indicates that the rate declined to 43.9 by 1999. After many years of growth, the percentage of births to unmarried women has hovered around 32-33% since 1994. Again, the change was evident before enactment of the 1996 law. (The National Center for Health Statistics reports that because of underreporting in some states, the actual peak would have been in the early 1990s rather than 1994 had numbers been fully reported.) It is possible to characterize the out of wedlock birth numbers in one of two ways. On the one hand, the upward trajectory of recent decades clearly slowed and perhaps stopped during the 1990s. At the same time, it seems to be the case that the dramatic declines in welfare participation and increases in employment have, at least to date, not yet been associated with equally dramatic changes in out of wedlock birth patterns. In 1996, some observers argued that welfare itself was a principal factor accounting for the rise of out of wedlock births in the United States, and that to alter these trends, it was necessary to eliminate or curtail the availability of assistance for single parent families. It is possible that reduced welfare participation and increased employment will eventually result in significant changes in family formation. However, it is not yet apparent whether that is occurring, and the other possibility is that the role of welfare benefits in contributing to out of wedlock births may have been overstated. A TANF provision provides $100 million a year, to be divided among up to five states that have demonstrated the largest reductions in the share of out of wedlock births while also demonstrating a decline in the number of abortions in the state. Five states were awarded $20 million each under this provision in 1999 and 2000. It remains unclear whether these states had initiated distinctive programming, or whether the bonus was principally rewarding demographic changes. Thus, it is not clear that awarding these bonuses is contributing to our understanding of effective approaches for reducing out of wedlock births. Similarly, under HHS rules announced last year, states will be eligible for a high performance bonus if the share of children residing in married families in the state increases, although there are comparable concerns that this type of bonus may also result in rewarding demographic changes rather than state performance. In recent months, some observers have argued that states have failed to actively advance the TANF goals of promoting marriage, reducing out of wedlock births and encouraging the formation and maintenance of two parent families. On the one hand, there was only limited discussion in 1996 of what Congress expected states to do to advance these goals - for many, it was assumed that reducing welfare caseloads was the intended strategy toward accomplishing them. However, as discussions of the family formation goals of TANF have increased, it seems clear that states face at least three questions in deciding how to proceed: what would work; what's the appropriate role for government; and how can the goals be advanced in ways that do not risk unintended adverse effects for children or parents. In efforts to promote employment, states have often been able to rely on an extensive research base developed over many years; there is not a comparable research base to draw from in efforts to reduce out of wedlock births or promote family formation. There is a stronger research base for teen pregnancy prevention efforts, but less to point to for initiatives affecting adults. Recent evidence suggest that stronger child support enforcement is correlated with reduced out of wedlock births and reduced marital break-ups. And, the evaluation of the Minnesota Family Investment Program found that a welfare reform effort that expanded eligibility for two-parent families, combined a work requirement with a modest sanction (10% grant reduction for noncompliance) and included a substantial earnings supplement for employed families, was associated with an increased likelihood that participants would enter or stay in two-parent families. A set of programs have demonstrated success in increasing engagement of fathers in the lives of their children. And, it has been suggested that a strong focus on promoting employment for young men could have an important impact in increasing their likelihood of marriage. Overall, though, a federal research agenda could strengthen the knowledge base in these areas. Even with additional knowledge, states still face difficult issues in determining the appropriate role for government and how to advance these goals without risking unintended consequences. In their efforts to promote work, states drew on a strong public consensus that, with limited exceptions, adults of working age ought to be working. There is not a comparable public consensus that adults of marriageable age ought to be married. There are strong indications from research findings that, all else being equal, better outcomes for children in a number of areas are associated with having been raised in a two-parent family. Part of the effect is because two-parent families are likely to have more income, but the research does suggest an independent effect for family structure. However, all else is not always equal, and it is difficult for states to determine how to translate these broad research findings into advice for an adult contemplating marriage or divorce. While marriage is often the best arrangement for parents and children, there are some circumstances where marriage or continued marriage may not be the best arrangement, e.g., in high-conflict marriages or circumstances of domestic violence. Moreover, because marriage and divorce involve highly personal decisions in which people must weigh an array of factors, there are concerns that strong government pressures or creation of financial incentives to marry could distort decisions in ways that may not necessarily be best for parents or children. While it is possible that such policies could increase the number of marriages, they might also increase martial dissolution rates and high conflict marriages. Government can clearly communicate that parents have an obligation to support their children, and states have done so through strong emphases on work and child support. Government can also seek to ensure that its policies are not having the effect of creating barriers to marriage and family formation. At the same time, the focus on family formation in TANF draws from a concern about seeking to improve outcomes for children, and efforts to address these goals need to proceed in ways that are both sensitive to the appropriate role of government and that do not present risks of leading to worse outcomes for children. Conclusion Thank you for inviting me to testify today. We appreciate the Subcommittee's interest in seeking to develop a picture of the experience since 1996, and hope this information can be helpful.

LOAD-DATE: March 17, 2001, Saturday