In the following article, 52 empirical studies of welfare recipients are surveyed. The object of the survey is to compare empirical evidence with the assumptions of current welfare reform proposals. Discovering a large gap between policy assumptions and research findings, the survey concludes with suggestions for both. Following the lead of their fellow taxpayers, social scientists have devoted increasing amounts of time and effort to the study of poverty and dependency; discussions of marginal tax rates, motivation, work incentives, and general welfare reform are now common in the journals of sociology and economics. Despite this extensive and growing interest in these problems, relatively little in the way of hard data has appeared in professional journals; most discussions continue to reflect the casual empiricism and bias of the authors. While one might conclude from this state of affairs that little empirical work has been done, this is not the case. On the contrary, federal agencies, especially the Manpower Administration of the U.S. Department of Labor, have funded a large number of research projects which bear directly on these issues. The results of these projects and other studies provide considerable insight into the dynamics of impoverishment and the potential of public assistance programs. Unfortunately, the limited exposure of these studies has made their results inaccessible to all but a handfull of social scientists. This brief survey attempts to help close the communications gap. In the following section, the results of 52 empirical studies are surThe author is Assistant Professor of Economics, University of Maryland. * Support for this study was received from the U.S. Department of Labor; all opinions are those of the author. An earlier version of this paper, entitled Facts and Fictions of Welfare Reform, was released by the Department of Labor. The Journal of Human Resources * VIII * Supplement This content downloaded from 157.55.39.60 on Mon, 18 Jul 2016 04:11:05 UTC All use subject to http://about.jstor.org/terms 20 1 THE JOURNAL OF HUMAN RESOURCES veyed. For organizational and analytical purposes, these results are discussed in the context of welfare reform. In particular, I first attempt to identify the major assumptions of current and pending welfare reform (for example, the Work Incentive Program [WIN], the Talmadge Amendments, President Nixon's reform bill [H.R. 1], and Senator Long's proposed Guaranteed Job Opportunity Program) and then subject these assumptions to empirical scrutiny.' THE ASSUMPTIONS OF REFORM A. The Distinction Between Working and Nonworking Poor At the core of much welfare reform is the persistent distinction between poor and nonworking poor. President Nixon's proposal (H.R. 1) is the most explicit in this regard, actually creating separate programs (OFP and FAP) to serve each target population. Senator Long's bill includes the same categorizations, while the present system of Aid to Families with Dependent Children (AFDC) explicitly excludes poor families headed by working fathers [40].2 The distinction drawn is based on the assumption that a poor person's employment status is stable over a considerable period of time. Unfortunately, there is little empirical justification for such a categorization. On the contrary, numerous studies and even official program statistics indicate that (1) there is a high degree of mobility between employment and dependency status, and (2) simultaneous work and welfare status is common. With regard to status mobility, AFDC program statistics [52] indicate that the median time on welfare is only 20 months. Longitudinal and retrospective studies of these recipients show that both before and after receiving public assistance, families are in the labor force [11, 22, 30, 32, 37, 46]. Similarly, studies of the working poor have demonstrated that substantial numbers (estimates run over 25 percent) of poor persons now working re1 It should be noted at the outset that the studies synthesized here are remarkably consistent in their findings and of generally high research quality. Hence, the usual caveats about the tentativeness of conclusions are omitted; only in a few places have I found it necessary to point out that available evidence is scarce. 2 The AFDC program is the existing mechanism for public assistance to families with children (currently about 11 million recipients), while the AFDC-U program serves families with unemployed fathers (approximately 660,000 recipients). The WIN program is the adjunct manpower training program for adult AFDC and AFDC-U recipients. The Talmadge Amendments (effective July 1, 1972) expand the WIN program and increase related work requirments. H.R. 1 is the Nixon Administration's (revised) welfare reform package, which would extend aid to the working poor. For more detail see [38, 48]. This content downloaded from 157.55.39.60 on Mon, 18 Jul 2016 04:11:05 UTC All use subject to http://about.jstor.org/terms