From Promoting Opportunity Demonstration: Final Evaluation Report, submitted to Social Security by Mathematica, a contractor:
... POD [Promoting Opportunity Demonstration] was a randomized controlled trial that included two treatments of a benefit offset. The two treatment groups had the same benefit offset but different termination rules. Treatment group 1 (T1) did not face termination, but treatment group 2 (T2) faced termination after 12 consecutive months of earnings above the full offset amount (the point at which benefits were reduced to zero). ...
The key features of POD implementation included benefits counseling services and support for processing earnings adjustments, led by the implementation team, and recruitment, led by the evaluation team. ...
Approximately 30 percent of treatment group members used the POD benefit offset, with a median monthly offset amount of $351. More than 80 percent of offset users experienced a work-related overpayment or underpayment, requiring a retroactive adjustment to reconcile the difference. ...
We did not observe any statistically significant differences in outcomes between the two treatment groups for overall offset usage or the impact estimates for the primary outcomes. ...
There were limited statistically significant differences in observed outcomes for the POD treatment and control groups. There were impacts on one primary outcome (annualized SGA) and several other employment-related measures. For example, we found positive impacts on job search and use of Vocational Rehabilitation services, which might contribute to longer-term outcomes. These impacts were notable because they indicate that impacts could still emerge beyond the two-year evaluation window. ...
POD had positive net benefits for beneficiaries and net costs to SSA. The net benefits for beneficiaries were driven by increases in earnings and fringe benefits, and SSDI benefit amounts. The new costs were driven primarily by the increased benefit payments and costs for counseling services. ...
Even with counseling, 80% of those using the offset ended up with an overpayment or underpayment! The experiment had only a limited effect on outcomes and ended up costing more money than it saved! Other than that, how was the play Mrs. Lincoln?
Why can't policymakers admit the obvious? Social Security disability recipients are, for the most part, really, really sick. Everything under the sun has been tried to get them back to jobs. Nothing has worked. Nothing. The work incentive schemes just get more difficult and expensive to administer. They end up with messy results for the disability recipients who do attempt to return to work because the offsets are too complicated The schemes always end up costing more money than they save. There's no possible work incentives that will get any significant number of disability recipients back to work because they're too sick.
Crappy experiments like this are likely to go on forever because policymakers are blinded by their own preconceptions that it's easy to get on Social Security disability benefits and that a lot of disability recipients could work if given the right incentives. They don't bother to study the pathetic history of work incentive failure. They get sold on new schemes by contractors like Mathematica who end up getting paid even though their schemes never work. Even after this disaster this 406 page report ends with ideas for new schemes that could be tried!