Race to the Bottom? Diffusion Processes in US Safety-net Programs

• In Event: Policymaking and Responsiveness in States and Localities

Abstract

Background: Why are some states more generous in their distribution of social welfare benefits than others? Diffusion scholars have long been interested in "competitive federalism,'' or the role that interstate competition plays in policy adoption. The 1996 welfare reform that block granted cash-assistance programs to the states is believed to have set-off a “race-to-the bottom” dynamic whereby states faced pressure to retrench the generosity of cash benefits to avoid becoming “welfare-magnets.” The empirical existence of this theoretical race-to-bottom dynamic has been debated, but little research has been updated since the early 2000s and research has largely been limited to the nations’ main cash assistance program- Temporary Assistance for Needy Families (TANF). This research updates past attempts at analyzing the inter- and intra-state spatial and temporal dynamics of welfare programs drawing on a composite index of four major safety-net programs (TANF, Supplemental Nutrition Assistance Program, Medicaid, and Unemployment Insurance) between 1996-2012.

Methods: Information on eligibility/enrollment rules and benefit levels for each program were compiled into a time-series, cross-section dataset across the 48 contiguous states (excluding Washington DC) for the period 1996-2012. Using spatial lag, spatial error, and mixed effects models, we examine the spatial dynamics of trends in state safety-net generosity across states in aggregate and across different programs. We conducted initial diagnostics to identify the existence and nature of any potential spatial process – lag, error, or mixed. We account for key political factors relevant to state welfare generosity, including unified state partisan control versus divided government, state citizen ideology, race/ethnic composition and state economic indicators. All diagnostics and our main regression models were implemented in R using the splm package. Replication materials are available from the authors.

Results: Overall, we find that each program appears to be governed by different political logics. First, only TANF has experienced consistent, across the board decreases in generosity across states over time. SNAP, Medicaid and UI have generally seen improvements in generosity in regards to eligibility, enrollment rule easing and benefit levels or have remained fairly flat across states. In regards to spatial diffusion processes, diagnostics reveal that while TANF and SNAP exhibit spatial diffusion processes, Medicaid and Unemployment Insurance do not. However, only TANF exhibits evidence of a race to the bottom dynamic in that there is evidence of a positive association between neighborhood generosity and focal state generosity in the context of a general downward trend. By contrast, SNAP exhibits a negative association whereby if a neighboring state increases its SNAP generosity, a focal state is likely to decrease its SNAP generosity.

Discussion: We find evidence of a race-to-bottom dynamic long hypothesized by theories of competitive federalism; however, we find that this is limited to the TANF program. While TANF generosity has been trending downward across states over time, other programs have been getting more generous or remaining constant. Moreover, temporal trends in state variation in Medicaid and UI do not appear to exhibit any detectable neighborhood effects suggesting non-spatial processes are more important determinants of generosity for these programs. The negative feedback effect detected in SNAP whereby a focal state’s reduction in generosity prompts a neighboring state to increase generosity requires further investigation. Rather than a race-to-the-bottom dynamic, SNAP-austerity in a state appears to prompt neighbors to adopt a more pro-poor stance. Future research can investigate the extent to which partisan dynamics in neighboring states help to explain this spatial dissimilarity or negative neighborhood effect.

Conclusions/implications: While the 1996 welfare reform does appear to have led to a negative feedback loop whereby states have competitively reduced their TANF generosity as scholars predicted, the impacts may have been somewhat buffered by increases in generosity in other less spurned safety-net programs. However, these results lend insight into the downward pressure that would likely be unleashed from block granting other safety-net programs.

Authors

• Ashley M Fox, SUNY at Albany

• Matthew C. Ingram, University at Albany, SUNY

• Wenhui Feng, University at Albany, SUNY