Setting the Terms of Relief: Explaining State Policy Choices in the Devolution Revolution

The landmark welfare legislation of 1996 offers students of politics a unique opportunity to pinpoint the determinants of state-level policy choices-a case in which the fifty states responded virtually simultaneously to a single policy mandate. Taking advantage of this opportunity, we investigate the factors that led states to make restrictive policy choices after 1996 and use this analysis to evaluate general -theories of welfare politics. Specifically, we test six types of explanations for why some states responded by adopting 'get-tough program rules: theories that identify welfare policy as a site of ideological conflict, as an outcome of electoral politics, as a domain of policy innovation, as an instrument of social control, as an outlet for racial resentments, and as an expression of moral values. The results of our ordered and binary logit models suggest that state policies have been shaped by a variety of social and political forces, but especially by the racial composition of families who rely on program benefits. n 1996, the federal government passed legislation that transformed public assistance provision in the United States. The Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA) abolished the federal entitlement to aid that grew out of the Social Security Act of 1935 and reached fruition through legal victories in the 1960s (Mink 1998; Lurie 1997). In its place, the federal government created Temporary Assistance for Needy Families (TANF), a system of block grants that gives states more freedom to select among policy tools but also imposes a forceful mandate to promote work, reduce welfare usage, and change poor people's behaviors (Albelda and Tilly 1997). In this article, we present a political analysis of the ways states responded to this new policy environment. Specifically, we investigate the factors that shaped state-level policy choices after 1996 and use this analysis as a basis for evaluating general explanations for welfare policy outcomes. Our study builds on a long tradition of quantitative research that has attempted to illuminate state-level politics by asking why states adopt different welfare policies (Howard 1999; Rom 1999; Brace and Jewitt 1995; Peterson and Rom 1990; Plotnick and Winters 1985). The analysis presented here, however, departs from prior work in two important respects. First, most state-level research has sought to explain interstate differences in benefit levels and spending patterns (Howard 1999, 424-425;