| This research project was designed to examine the effectiveness of welfare devolution since the mid-1990s. Specifically, it contains three essays investigating the three underlying assumptions of the devolutionary approach to welfare reform, respectively---namely, the accountability argument, the experiment argument, and the better policy argument.;The first essay focuses on whether or not state governments are accountable to the general public in their welfare policymaking under devolution. The results show that at the state level, one-to-one opinion-policy congruence regarding any of the important welfare reform provisions does not exist. However, if we define public opinion as a general outline to guide the implementation of state welfare reform, we can observe a robust linkage between state-level public opinion and policy output toward welfare.;The second essay focuses on how states serve as laboratories to experiment and disseminate policy innovations. When two broad dimensions of state welfare reform (Stick vs. Carrot) are identified, this study finds two distinguished paths of policy diffusion, respectively: first, states with successful "stick" policies are more likely to be emulated than are those with failing policies. Second, "carrot" policies are more likely to flow from states with unified Democratic governments to other states with the same partisan configuration.;The third essay focuses on whether or not welfare devolution promotes better reform policies in dealing with the major welfare problem since the 1970s---that is, the growing number of welfare caseloads. By using the two policy dimensions---"stick" policies and "carrot" policies---to summarize various welfare policy provisions at the state level in the post-reform era, this study shows that while "stick" polices may help reduce caseloads, "carrot'' policies, on the other hand, may help increase caseloads; however, during the time period between 2000 and 2003, the overall impact of welfare reform on caseload changes appears to be minimal. |
