The Role of Nonfinancial Factors in Exit and Entry in the TANF Program

Abstract

The dramatic decline in the AFDC-TANF caseload in the 1990s has refocused attention on the process of exit from and entry into welfare, a long-standing topic of interest in the research literature on the U.S. welfare system. This paper focuses on the role of non-financial factors in exit and entry in the post-1996 TANF program. The non-financial factors are work and other requirements, sanctions, and diversion. Using data from a study of welfare and nonwelfare families in Boston, Chicago, and San Antonio in the period 1999-2001, both descriptive evidence and evidence from an econometric model suggest that these factors played a large role in exit and entry over the period. The dramatic decline in the AFDC-TANF caseload in the 1990s has refocused attention on the process of exit from and entry onto welfare, a topic upon which a considerable research literature has been built, starting from Boskin and Nold (1975) and continuing through the present (Bane and Ellwood, 1994; Blank and Ruggles, 1994, 1996; Moffitt, 2001; see Moffitt, 1992,2003, for reviews). Most recent attention has been focused on the determinants and consequences of the decline in the AFDC-TANF in the last decade, with one strand of literature focusing on consequences of leaving welfare for employment and income (see Acs and Loprest, 2001; Brauner and Loprest, 1999; Moffitt, 2002 for reviews) and another strand focusing on the estimation of aggregate caseload models attempting to parcel out the relative contributions of the economy and welfare reform (see Blank, 2002, for a review). The first strand of literature shows that welfare “leavers” have substantial increases in employment upon exit but only modest increases in total income, while the second strand shows similarly large increases in employment but also large increases in income in the welfare-eligible low-income population as a whole. The larger income increases in the latter literature suggest that income may have risen among women who did not enter welfare or among those who remained on welfare. This paper focuses on the contribution of non-financial factors to exit and entry in the TANF program. The non-financial factors examined are work and other requirements, sanctions, and diversion. While each of these factors has some financial implications--earnings should rise as a result of work requirements, benefits should be reduced from sanctions, and so on--they also impose a non-financial utility cost on welfare participation, in the case of work requirements and sanctions, and on welfare entry, in the case of diversion. These utility costs should be expected 1 In the traditional leisure-income model, the four determinants of welfare participation are the wage rate, nonwelfare nonwage income, the welfare guarantee, and the welfare tax rate. Studies of the AFDC program have shown all four to be significantly related to static welfare participation rates as well as welfare entry and exit. See Moffitt (1992, 2003) for reviews of this literature. 2 to increase exit and decrease entry, and are conceptually separate from the traditional financial factors such as benefits, wages, and the EITC which have been shown to play a strong role in welfare participation prior to 1996. Another important financial reform occurring since 1996 has been a lowering of earnings disregards, but they are unlikely to be an explanator for declining caseloads because they decrease exit, increase entry, and increase static welfare participation rates. There has been a considerable amount of analysis attempting to estimate the effects of cross-state variation in work requirements, sanction, diversion, and other rules on the aggregate caseload, usually in the pre-1996, waiver period when not all states had implemented reforms. Unfortunately, that literature has produced generally insignificant and/or uninterpretable results (Bell, 2001; Blank, 2002), probably because the formal rules are not easily measured, because they may be only weakly correlated with actual implementation, or because there are unobserved state-specific factors that are correlated with, and hence confound, the estimated effects of the rules. This paper uses instead a data set from three cities (Boston, Chicago, and San Antonio) containing information from a survey which gathered respondent-supplied information on the actual experiences of a set of low-income single mothers with these rules. For TANF recipients, questions were asked about experiences with work and other requirements and with sanctions, and, for TANF applicants, questions were asked about experiences with diversion. The data set is a two-wave panel and hence allows the estimation of the determinants of both exit and entry,