Methods We draw from welfare and Unemployment Insurance (UI) administrative data from Wisconsin to examine outcomes among a cohort of participants who entered Wisconsin's TANF program, Wisconsin Works (W-2), in September or October of 2006 (N=682) at one of the four largest (caseload) agencies. We begin with descriptive chi square analysis to compare individual characteristics, welfare use patterns, and annual earnings pre- and post program entry between participant types. We then utilize time-discrete multinomial regression models (survival analysis) to examine the likelihood of leavers exiting welfare for a good job, bad job, or no employment. A time-discrete logistic regression model is employed to test the probability of returning to welfare among leavers. We then describe economic outcomes in the year following participation.
Results We find significant differences between participants by work-exempt status. Participants who are disabled and with infant children are older, have higher levels of education, and are disproportionately less likely to be African American. They also have higher and more consistent employment and earnings patterns pre- and post- welfare participation relative to those with assigned work requirements. Mothers with infant children earn and work the most, are on welfare for the shortest periods of time, and are less likely to re-enter. Overall findings suggest new mother participants and to a lesser extent, those who are disabled participants, enter welfare with higher levels of human capital skills and have different post-participation outcomes relative to welfare recipients assigned to work activities.
Conclusions & Implications Disentangling welfare outcomes by work-exemption status may assist policy makers in understanding the differential impact of welfare's work requirements, reconsider the role of the welfare program as a work-support, and open the door to innovative programs that might better meet the needs of work-exempt participants.