Evaluating Performance of Performance Contracting in the Wisconsin Works (W-2) Program: A Difference-in-Difference Approach

Purpose

The Wisconsin Works (W-2) program, the TANF program in Wisconsin, has introduced performance contracting to align W-2 service providers' objectives with the State's goals by establishing financial and administrative incentives. This study will examine whether the introduction and revision of performance contracting has contributed to improving the performance of the W-2 participants.

Data

I constructed W-2 participant level panel datasets from three major data sources. First, the Client Assistance for Re-employment and Economic Support (CARES) system provided longitudinal information on the characteristics and program participation history of participants in the W-2 program. Second, I used employment and earning information from Wisconsin Unemployment Insurance (UI) dataset. Third, I utilized state policy documents on W-2 contracts to identify changes in W-2 performance contracting arrangements.

Analytic Strategy

There have been four rounds of biannual performance contracting (1997-99, 2000-01, 2002-03, and 2004-05) in the W-2 program. In the 2nd contract (2000-01), earning gain standard was first introduced and applied to all W-2 participants as a requirement standard to obtain a) the right of contract renewal, and b) performance bonuses.

However, in the 3rd contract (2002-03), the standard was applied only to specific participant groups (e.g. participants in unsubsidized employment placements). It might increase W-2 agencies' incentives a) to select participants who might be more likely to produce earning gain, and b) to utilize more resources to serve the groups, lowering their incentives to provide intensive services to other groups of participants (e.g. participants in subsidized employment placements). Thus, by comparing change in earning gain of the two groups from the 2nd contract to the 3rd contract, it is possible to identify the effects of earning gain performance standard on W-2 participants' outcomes.

The earning gain standard was changed from a required standard in the 3rd contract to an optional standard in the 4th contract (2004-05). It provides an additional opportunity to confirm the effect of earning gain standard, by observing what happened when the standard was no longer used as a required standard.

Results

Comparison of earning gain between the 2nd and the 3rd contract across the two groups reveals that the earning gain of the unsubsidized employment groups (treatment group) maintained the same level of earning gain, whereas subsidized employment groups (comparison group) suffered $160 decrease in quarterly earning gain. In addition, participants with higher educated and having better earning history were selected to the unsubsidized groups in the 3rd contract period. The elimination of earning gain standard in the 4th contract shows decline in the quarterly earning gain for the unsubsidized group by $74. It confirms the positive effect of adopting earning gain standard.

Conclusion

Under exacerbating economic condition and weakening financial incentives, adoption of earning gain standard contributed to maintaining substantial level of economic performance of relevant W-2 participants. This research provides insights to inform a better contract design to improve the effectiveness of the TANF service administration. Performance contracting plays an important role by inducing welfare service providers to achieve higher performance in the TANF program.