We have previously reported (see here, here, and here) on the IRS's bizarre plans to outsource some of its tax collection responsibilities to private companies. The plan has received strong opposition from many in Congress and ended up costing taxpayers money the last time the IRS tried it. But they are pushing forward, nonetheless, and even talking about expanding the program before it has even been tried. Now the Government Accountability Office has weighed in on the program and they seem to think something doesn't smell quite right: IRS has not documented criteria that it will use to determine whether the limited implementation performance warrants program expansion. IRS officials indicated that they are considering criteria that could trigger a go/no go decision, such as the amount of taxes collected and indications of PCAs [private collection agencies] abusing taxpayers or misusing taxpayer data. IRS has not decided on whether these targets will include comparing the taxes collected to program costs, which was a key reason for canceling a 1996 PCA pilot program. GAO goes on to say that the IRS is planning a comparative study to determine whether the private collection agencies are doing a better job than the IRS could do if it invested its resources in collecting the taxes itself. But get this - under the design of the study, the IRS will exclude the fees paid to the private agencies (up to a quarter of the revenue collected) in their analysis comparing the two different approaches. Because of this strange exclusion, GAO concludes: [The study] will not compare the results of using PCAs with the results IRS could get if given the same amount of resources, including the fees to be paid to PCAs, to use in what IRS officials would judge to be the best way to meet tax collection goals. Adequately designing and implementing the study is important to ensure policymakers are aware of the true costs of contracting with PCAs and know whether PCAs offer the best use of federal funds. So not only has IRS rushed forward with plans to implement this program despite heavy criticism, it has also skimpted on designing the process to guage the program's success, and then in order to stack the deck, decided to exclude the actual costs from its lone study - thereby rendering the evaluation essentially useless? This definitely smells fishy.