Second interim report of the Work Rewards demonstration project follows up initial two year findings with longer term results on the effectiveness of three different types of work incentives.


“Policymakers have long sought to improve voucher holders’ labor market outcomes. Toward that end, HUD funds the Family Self-Sufficiency (FSS) program, providing public housing authorities with modest resources to hire case managers, who work with participants to develop plans for becoming self-sufficient and to connect them with services in their communities.” (p.ES-1).

”…the Work Rewards demonstration encompasses tests of three distinct strategies: (1) FSS alone, (2) FSS plus special work incentives, and (3) the special work incentives alone. The first two of these tests (’FSS-only’ and ’FSS+incentives’) are both part of the ’FSS study’ in this report, and they involve households with vouchers obtained through HPD. The third test (without FSS), referred to as the ’incentives-only’  study in this report, involves households with vouchers obtained through NYCHA. Using two parallel, randomized controlled trials (RCTs), the evaluation is determining the effects, or ’impacts,’ of the FSS program and the new special work incentives on voucher holders’ employment outcomes, housing subsidy receipt, receipt of other public assistance benefits, and various quality-of-life outcomes.” (p.ES-2).

“All three projects — Family Rewards, Spark, and Work Rewards — rely on a random assignment design to assess their effectiveness, which is considered the strongest design for determining program effects. In such studies, participants are randomly assigned to a program group, which is exposed to an intervention, or to a control group, which is not, and the differences in outcomes between the two groups indicate the program effects, or ’impacts.’ Similarly, two (or more) different program groups can be exposed to different interventions and their respective outcomes compared. MDRC is conducting the evaluation of Work Rewards, while Seedco, along with a small group of community-based organizations (CBOs), operated the Work Rewards demonstration” (p.2).

(Abstractor: Author)

Major Findings & Recommendations

“The study is testing the effectiveness of three programs: FSS alone, FSS plus special cash work incentives, and the special incentives alone. The incentives were offered as a way to test whether attaching more immediate cash rewards to work-related activities (compared with the more distant reward of escrow savings) produces positive labor market and other effects. An earlier report concluded that in its first two and a half years, FSS yielded benefits for some participants but not for others. Longer-term follow-up from Work Rewards for the first four years of the study, which is presented in this second, interim report, corroborates the findings from that first report. Overall, the study found no sustained effects for the full samples in any of the three programs. But it did identify positive effects for a certain subgroup: FSS combined with the special work incentives produced large increases in employment and earnings for participants who were not working at study entry. Those gains did not, however, translate into reduced poverty or reduced reliance on public benefits, suggesting, perhaps, that more must be done to help this population advance once they find jobs. These findings can serve as the foundation for building stronger self-sufficiency programs in New York and other cities” (p.xi). “The findings show that the FSS+incentives intervention produced large and statistically significant impacts on employment and earnings for voucher holders who were not already working when they entered the program. It also appears to be producing some later-occurring reductions in the receipt of TANF and food stamps for the core (that is, nonelderly and non-disabled) sample, though no reductions in housing subsidies. The FSS-only and incentives-only interventions did not produce a consistent pattern of statistically significant impacts on employment and earnings overall or for the employment subgroups, but FSS-only appears to have contributed to reductions in food stamp receipt for the group that was not working at the time of random assignment” (p.ES-3-4). (Abstractor: Author)