Summary:

Abstract: Poverty policy in the United States is not what it used to be. Over the past twenty-five years, the American style of governing poverty has grown more dispersed in its organization and more muscular in its efforts to enforce behavioral norms. Policies that once prioritized governmental control and federal involvement have shifted decisively toward devolution, privatization, and competition. At the same time, policies that once emphasized rights, supports, and opportunities now adopt a stance that is more directive, supervisory, and punitive. Public officials have proved remarkably willing to hand policy control over to lower-level jurisdictions and private providers. They have also become more willing to use state institutions and resources in ways that overtly promote values, monitor behavior, enforce obligations, and punish deviance.
Numerous scholars have taken note of these changes, particularly in the areas of welfare provision and criminal justice (Mead 1998, 2005; Wacquant 2001; Lowi 1998). Today, for example, one can find sophisticated empirical literatures on the consequences of mass incarceration (Western 2006) and the effects of work-promotion and sanctioning in welfare programs (Grogger and Karoly 2005). One can also find a number of ambitious efforts to theorize recent changes in poverty governance and their relation to social control (e.g., Lowi 1998; Garland 2002; Wacquant 2001). What both streams of literature lack, however, is an empirically-grounded analysis of how discipline actually gets practiced in a decentralized policy system: how it is organized and managed, why it operates as it does, and what difference this makes for the who, when, and where of disciplinary action.
In this paper, we take some preliminary steps toward such an account. To do so, we analyze field observations, interview transcripts, and administrative data regarding the use of sanctions in Florida?s version of the Temporary Assistance for Needy Families (TANF) program, Welfare Transitions (WT). In 1996, federal lawmakers abolished the Aid to Families with Dependent Children (AFDC) entitlement, replacing it with the TANF block grant to the states. Today, state TANF programs offer conditional, time-limited aid. In addition to having children and meeting a means-test, adult participants must satisfy a variety of behavioral requirements to remain eligible for income support. ?Sanctions? are penalties that reduce or eliminate aid when a client fails to comply with program requirements. It is the primary disciplinary action available to TANF case managers as a tool for influencing client behavior and, if deemed necessary, imposing consequences for non-compliant behavior. Like other aspects of the TANF program, it is administered in many states in a highly decentralized fashion.
Decentralized administrative structures tend to pose a variety of significant management challenges (Kettl 2002), and one might expect devolved systems of discipline to be no different. Yet leading studies of TANF sanctions, which focus mostly on individual characteristics and economic contexts, have paid virtually no attention to how sanctioning may be influenced by management tools and organizational practices. This omission is striking, in the first instance, because sanctioning is obviously carried out in the context of organizational routines, by actors immersed in organizational cultures. Indeed, large amounts of time and resources are invested by administrators on the assumption that changes to organizational routines and cultures do, in fact, affect this sort of program outcome.
Our goal in this paper is to analyze how performance management influences the practice of disciplinary action in a decentralized welfare system. After a discussion of issues related to case selection, we begin our empirical analysis close to the ground. Drawing on in-depth interviews and field observations, we first explore the salience of performance management at the frontlines of welfare-to-work and its potential to shape administrative behavior. We then turn to the question of how performance feedback might affect sanctioning practices, advancing arguments against one possible connection and outlining an alternative set of causal linkages that strikes us as more empirically defensible. Drawing this analysis of field research into dialogue with existing scholarship, we then outline a set of hypotheses more suitable for testing with a statistical analysis of administrative data. Using panel data models of local sanctioning rates, we present findings that are consistent with our on-site observations but that also go beyond what it is possible to demonstrate through field methods. Combining the two halves of our analysis, we offer an account of how performance management affects sanctioning practices on the ground, how performance pressures influence the rate of sanctioning, and why the effects of performance pressures tend to be concentrated in some locales and social groups rather than others.