Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
1 Introduction Robert A. Moffitt Whether the U.S. welfare system has an effect on marriage, childbearing, living arrangements, and other aspects of demographic and family structure is an issue that has a long history both in the public mind and in research circles. In public and media discussions, the notions that welfare provides an incentive for women to not marry or remarry, to have children out of wedlock, and to live independently rather than at home with parents, have been prominent for over 30 years. Indeed, public attention to these issues accelerated in the l990s as welfare reform debates in Washington and around the country became increasingly fo- cused on "values" and as specific reform measures began to be proposed to reduce undesirable incentives (e.g., limiting the amount of welfare benefit a mother could receive by having an additional child). At the same time, in re- search circles, these ideas have been treated instead as hypotheses that should be made subject to test, and the research community has produced a long string of research studies examining these issues in great detail. The research literature itself has also accelerated to some degree in recent decades, with more social scientists examining the issue in the 1980s than in the 1970s, and more in the l990s than in the 1980s. This research trend is undoubtedly a response to the shift in public attention to the issue. In May 1996, the Committee on Population and the Board on Children, Youth, and Families of the National Academy of Sciences-National Research Council and Institute of Medicine convened a workshop on the Effects of Wel- fare on the Family and Reproductive Behavior. Its purpose was to assess what the research community has learned from the studies that had been conducted to date, to identify gaps, and to suggest new areas of research that would be relevant 1
2 INTRODUCTION to the changing policy environment to assess, in the words of the organizers, "what we know and what we need to know" about welfare and the family. The workshop brought together approximately 60 experts to hear a series of presenta- tions by prominent researchers in the area and to discuss future directions. A summary of the discussion that took place at the conference is reported elsewhere (Haaga and Moffitt, 19981. This volume contains the revised and edited presen- tations from that conference. This introduction first briefly summarizes each of the chapters in this volume and then discusses their implications for the welfare environment, which has changed dramatically from that in place at the time of the workshop. Just 3 months after that time, the U.S. Congress passed what is generally regarded as the most significant piece of welfare reform legislation since the Social Security Act of 1935. The legislation, titled the Personal Responsibility and Work Opportu- nity Reconciliation Act (PRWORA), passed Congress and was signed by the President in August 1996. The act eliminated the well-known Aid to Families with Dependent Children (AFDC) program and replaced it with the Temporary Assistance to Needy Families (TANF) program, delegated most of the program- matic and budgetary responsibility for the new program to the states, mandated new work requirement and time limit provisions for the program, and modified in major dimensions the eligibility conditions and provisions of many other welfare programs. The act could have major consequences for the demographic behavior of the low-income population. The following section discusses the relevance of the chapters in this volume for the new welfare environment. CHAPTERS IN THIS VOLUME Chapters 2 and 3 provide overviews of recent trends in demographic behav- ior and the welfare system, respectively. In Chapter 2, Christine Bachrach sum- marizes the changing circumstances of fertility, marriage, and out-of-wedlock childbearing since 1970. Bachrach demonstrates what most researchers in this area know but the general public often does not, which is that childbearing trends for the whole female population as well as for younger women have not exhibited drastic swings over the last 30 years. However, a major shift has occurred in the proportion of births that occur outside marriage. Bachrach develops this point further by showing that there have been much larger declines in the rate of marriage than of childbearing and that the timing of marriage has drastically shifted toward later ages. She then addresses the increasing rate of nonmarital childbearing and shows that the trend is explained by, more than any other factor, a decline in age-specific marriage rates. This is particularly true for the black population but also for the white population, although the latter has experienced significant declines in childbearing rates among unmarried women as well. This finding has the very important research implication that it is the decline in marriage, rather than increase in desire for children, that should be the focus of
ROBERTA. MOFFITT further research, and the important policy 3 i1 Implication that it is marriage, rather than childbearing, that should be the focus of any policy measures intended to address nonmarital childbearing. Bachrach also provides a useful overview of trends in sexual behavior and contraceptive use, in the acceptability of having a child outside of marriage, and with respect to socioeconomic differentials, find- ing that upward trends in nonmarital childbearing have been concentrated among more disadvantaged groups. In Chapter 3, Rebecca Blank summarizes trends in the U.S. welfare system. Blank shows that the major turning point in benefits and caseloads in the system occurred in the late 1960s and early 1970s, when caseloads in the AFDC program grew dramatically and the Food Stamp and Medicaid programs were introduced. She shows that both benefits and caseloads for the programs were, however, quite stable over the rest of the 1970s and early 1980s. In the late 1980s and early 1990s, AFDC and Food Stamp costs and caseloads grew but were dwarfed by an enormous increase in Medicaid expenditures, in large part because of expansions of the eligibility pool for the program. Blank shows that this trend has put great pressure on state budgets. She then provides a detailed description of the PRWORA legislation, calling it "both less radical and more radical" than often claimed. The most radical provisions are those converting the AFDC program to a block grant, those requiring time limits on receipt of benefits in the new TANF program, and the new work requirement mandates. Blank concludes by emphasizing that the PRWORA legislation merely pushed further trends that had already been occur- ring for several years, including an increasing emphasis on behavioral require- ments as a condition of program eligibility (with particular emphasis on work behavior), an increasing trend toward decentralization in the design of programs, and a trend toward reductions in expenditures and entitlements to welfare. The remaining four chapters provide reviews and summaries of research findings in four specific areas: the effect of welfare on marriage and fertility; the connections between welfare and abortion; the effect of pre-PRWORA welfare reform interventions on demographic outcomes; and the effect of welfare on children. In Chapter 4, I review the large research literature on the issue of whether the welfare system, especially the AFDC program, has discouraged marriage and encouraged childbearing. The review concentrates on behavioral research using secondary datasets and household surveys, leaving a review of demonstration research to Chapter 6 by Maynard et al. I argue that the consensus in the research community shifted over time from the 1970s, when it was generally believed that the welfare system had very little effect on marriage and childbearing, to the 1980s and 1990s, when most analysts came to believe that there is an effect. But the magnitude of any effect that is present is highly uncertain and unresolved; some researchers argue that the effect is small and others argue that it is sizable. Research has not shown the welfare system to have been the major contributor to the recent trend in nonmarital childbearing (documented in Chapter 2 by Bachrach)
4 INTRODUCTION and, for this reason, the consensus in the research community is that other forces must have been at work in generating the trend. The chapter also goes into considerable detail on the methodologies used by different researchers to mea- sure the existence and magnitude of welfare effects and criticizes the research community for failing to reconcile differences in findings that are reported in different studies. In Chapter 5 on welfare and abortion, Jacob Klerman reviews several issues. One is whether reductions in the generosity of the welfare system, such as those that will result from PRWORA, are likely to affect the rate of abortions. Klerman argues that there is strong reason to believe that abortions will, if anything, increase from such a reform, although he discusses alternative perspectives that may not lead to that result. His review of the empirical studies that have been conducted leads him to conclude that there is little evidence that benefits in the AFDC program have, historically, had any significant effect on the abortion rate. However, he reviews both the data and the statistical difficulties in these studies and finds that the data have significant deficiencies and that statistical limitations reduce confidence in the results. However, Klerman does find that Medicaid restrictions on abortions have had an impact on the abortion rate, according to the research literature. In Chapter 6, Rebecca Maynard and her coauthors turn away from the nonexperimental behavioral research to the findings of demonstration research. Maynard et al. document that there were a large number of demonstration projects mounted in the states in the 1980s and 1990s prior to PRWORA, most of which were known as "waiver" projects, that were aimed at testing various reforms in the AFDC system. The authors also document the change in the goals of the waiver demonstrations, from an emphasis on work in the 1980s to an emphasis on family structure, parenting, and socially desirable behavior in the 1990s (echoing the trend noted by Blank). Maynard et al. show that there were only two types of waiver demonstrations, however, that directly addressed de- mographic issues. These were waivers testing a "family cap" a restriction on the increase in benefit payment to a welfare mother who has had an additional child and waivers relaxing the stringent eligibility requirements in the AFDC- UP (unemployed parent) program, the program for which two-parent families are eligible (a reform that might be thought to encourage marriage). The authors find that very few evaluations of these waivers have yet reported results. From those that have, such as the family cap demonstration in New Jersey, a decidedly mixed and complex picture has resulted, possibly because of flaws in the dem- onstration design. The authors do review the findings of a few prior demonstra- tions aimed at assisting single mothers but usually not directly aimed at their childbearing or marital behavior and show that the findings from those demon- strations are also quite mixed, some demonstrations increasing childbearing and others decreasing it, for example. Maynard et al. conclude by reviewing the importance of good designs when conducting demonstration research and make
ROBERTA. MOFFIIT s specific recommendations to states regarding how to evaluate their PRWORA programs. In Chapter 7, Janet Currie takes a broad look at several major welfare pro- grams (Food Stamps, Medicaid, AFDC, housing, and several others) to assess their effects on child outcomes such as birthweight, nutrition, health care, test scores, and the like. Currie provides considerable discussion of the methodologi- cal and statistical difficulties in assessing the true effects of the programs. Nev- ertheless, her review of the research yields one striking finding: unrestricted transfers such as AFDC and the Earned Income Tax Credit have relatively few discernible effects on children, but transfer programs that have specific targeting on children such as the school nutrition, WIC, and Head Start programs are much more likely to show positive effects. This finding has clear and significant policy implications. On the other hand, Currie finds that the research literature has neglected many important issues such as the long-run effects of the programs, leading to significant gaps in our knowledge of those types of effects. In her conclusion, she lays out a series of key research questions that should be ad- dressed in future research in this important area. IMPLICATIONS FOR POLICY AND RESEARCH IN THE PRWORA ERA Chapters 4 through 7 reach conclusions that have many similarities. Most find that the majority of studies show either no significant effects of AFDC and other welfare programs, effects that are statistically significant but small in mag- nitude, a set of mixed effects indicating some that are favorable and some unfa- vorable, or effects that occur only for some specific types of programs. Although the research reviewed in these chapters does not support a finding of no effect whatsoever of welfare programs on demographic behavior, it would be difficult to argue that the research often indicates very sizable or stable effects. Whether this is a result of problems with the studies themselves, as discussed at length by the chapter authors, or whether it is the true state of affairs cannot be decided with certainty at this time. However, it is also fair to note that if there were a sizable effect of welfare on demographic behavior, it would probably be more evident with the available statistical methods than appears to be the case in the research literature. The findings reported in the chapters are, on the contrary, consistent with the existence of a small, real effect but one that is difficult to detect and sensitive to the methodology used because it is small relative to other factors determining demographic outcomes. The wide dispersion of findings in many of the research literatures surveyed by the authors weakens the confidence one can have in this or any other conclu- sion. To some extent, a variance in research findings is common to all areas of investigation and is not particularly surprising. However, it does worsen the traditional conflict between the desires and needs of policy makers, who want
6 INTRODUCTION certainty before making policy recommendations, and the research community, which is willing to accept uncertainty and thus is more accustomed to reaching tentative conclusions. Still, it would be unquestionably preferable for the amount of uncertainty to be reduced below what it is in many of these research literatures, because the impact of research on policy is, in general, in strong inverse propor- tion to the degree of dispersion of its findings. In considering the implications of this research for the demographic effects of the PRWORA legislation, the clear prima facie implication is that those effects can be expected to be small. However, whether this will occur depends on whether the effects of historical programs like AFDC can be extrapolated to the findings of a much different set of state programs that will evolve in the next few years. Certainly PRWORA both requires and allows changes in the programs that are more significant than the types of variations in the AFDC program used to estimate demographic effects in past behavioral and demonstration research. The work requirements of PRWORA go considerably beyond those of the tradi- tional AFDC programs and beyond those of the Family Support Act of 1988, as clearly do the time limits, which have been tested in the past only in waiver form.) Thus it is absolutely necessary to conduct rigorous evaluations of PRWORA and to include demographic outcomes (marriage, childbearing, etc.) as part of those evaluations. Several of the chapters address the relative advantages of nonexperimental behavioral and demonstration research in studies of the past welfare system, an issue that must be addressed in evaluations of PRWORA as well. The discus- sions in the chapters suggest that neither type of research should be relied upon exclusively. The methodological discussions contained in various chapters make clear the difficulties of nonexperimental behavioral research, while the review of demonstration research provided by Maynard et al. shows how difficult conduct- ing a good demonstration is as well. The chapters also provide a basis for concluding that even good demonstration research provides answers to only a narrow set of questions and that context, as well as exploration of the mecha- nisms by which responses occur, is more easily obtained from the analysis of nonexperimental, secondary datasets. Contextual and ethnographic perspectives also bring added information to the response of families to welfare reform that is not provided by either nonexperimental behavioral modeling or demonstration research. A balanced strategy employing a mix of all approaches, not unlike that lit is worth noting, however, that the PRWORA legislation did not, in the end, have many provi- sions directly aimed at demographic outcomes. Neither family caps, significant changes in the AFDC-UP program, nor prohibitions on the provision of benefits to unmarried mothers were man- dated, for example, despite the extensive legislative discussion of such provisions in the debates preceding passage of the bill. In the end, the U.S. Congress left those decisions to the states. For this reason, the major impact of PRWORA on demographic outcomes will operate indirectly from the reduction of benefits and eligibility that will result from the main provisions of the bill.
ROBERTA. MOFFITT 7 outlined by Maynard et al. in the concluding sections of Chapter 6, would appear to be the safest overall strategy toward PRWORA evaluation. At this writing, many evaluations of PRWORA are under way, and no doubt many more will be initiated over the next few months and years. A series of demonstration projects in the individual states, many of which are funded in part or in whole by the U.S. Department of Health and Human Services, have been put in place to study state-level welfare reform and child care projects, as well as child outcomes. Some of these demonstrations are continuances of evaluations of pre-PRWORA waiver demonstrations, some of which were continued after PRWORA, and others are new evaluations. Some use experimental evaluation methods and others consist of longitudinal data collection designs (e.g., of the child welfare system). Another study has been initiated as well by the U.S. General Accounting Office to monitor welfare reform in the states as it proceeds. Many more local studies have also been begun by state agencies. At the same time, several survey research projects have been set in motion to evaluate the effects of PRWORA by methods coming more from the tradition of nonexperimental behavioral research. The U.S. Bureau of the Census has initi- ated a Survey of Program Dynamics intended to follow for several more years a group of families previously interviewed under the auspices of the Survey of Income and Program Participation. The families will be interviewed on a peri- odic basis to ascertain responses to PRWORA. The Urban Institute has con- ducted one wave of a cross-sectional telephone survey of families in 13 states to ascertain welfare responses and will follow this survey up with a second wave in 2 or 3 years. Surveys of welfare recipients, former but recent welfare recipients, and nonrecipients in a set of specific cities or counties sometimes supplemented by administrative data have been begun by research teams centered at the Man- power Demonstration Research Corporation, Johns Hopkins University, Princeton University, and other locations. Ethnographic and contextual research is a part of several of these efforts. A project to study the administrative response to welfare reform has also been initiated by the Rockefeller Institute at the State University of New York at Albany. In addition, the research community can expect to see more traditional research studies conducted using well-known national surveys like the Current Population Survey, the Panel Study on Income Dynamics, the National Longitudinal Survey, and others, as well as using aggregate caseload data from the new welfare system. Although not all of these studies will have demographic outcomes as a major focus, many if not most will at least have them as a minor focus. While there is no guarantee that such a decentralized and uncoordinated set of activities will produce consensus findings on the effects of PRWORA and one of the lessons of the chapters reviewed in this volume is that the investigators conducting these studies will have to work to meet the challenge of reconciling differences in findings across studies there are reasonable prospects that much useful information on the effects of PRWORA will be gathered.
8 INTRODUCTION The research surveyed in these chapters, covering studies conducted on the old system right up to the point of its demise, should provide in the years to come a statement of where we were, in terms of research knowledge, on the eve of the beginning of a new welfare era. If a similar volume comes to be produced in the future, say 10 years hence, it will be interesting to compare its reviews with those reported here in terms of the fundamental question of whether welfare affects demographic outcomes, as well as whether the research community has been any more successful in reaching consensus on what those outcomes are. REFERENCE John Haaga and Robert A. Moffitt, eds. 1988 Welfare, The Family, and Reproductive Behavior: Report of a Meeting. Committee on Population and Board on Children, Youth, and Families, National Research Council and Institute of Medicine. Washington, D.C.: National Academy Press.