The IRS's new system to detect tax return fraud still has a few bugs to be worked out.

The Return Review Program, the agency's new system to detect returns that show signs of identity theft, was tested in pilot programs in 2014 and 2015. The Treasury Inspector General for Tax Administration found that while the RRP did find fraudulent returns the previous systems didn't, it missed others with refunds totaling $313 million.

Related:Read the Report

Identity theft, and the refunds claimed from it, have become an increasing problem the IRS is battling to address. The agency said it rejected 1.8 million fraudulent returns filed in 2014, worth $10.8 billion in refunds.

The RRP, which has been in development since 2009, is meant to eventually replace the Electronic Fraud Defection System, which the IRS uses to catch electronic identity theft returns. The IRS also uses the Dependent Database to cross reference IRS files with Department of Health and Human Services and Social Security Administration files to also detect fraudulent returns.

A 2014 pilot test of the RRP system to detect a limited number of returns found that it did find 25 percent more bogus filings than the systems currently in place. Based on those results, IRS ramped up the number of returns RRP would process in 2015. But a TIGTA analysis of the 2015 results found that the RRP system did not detect 54,175 confirmed fraudulent returns that the DDb system did discover.

IRS officials told TIGTA that the gap was because not all of the RRP capabilities were put in place in 2014 and that the system was being tested to find newer schemes rather than older ones that would be easily detected by older systems.

"Because the RRP selection models were not fully implemented in [Processing Year] 2014, the IRS explained that analysis of identity theft selections in PY 2014 would not provide useful information to identify the gaps between the identity theft detection systems," the report said.

The IRS processed more than 96 million electronically filed tax refund returns as of May 27. The agency analyzed tax returns selected by the three systems for identity theft screening, showing that each of the systems flagged that the other systems missed. The analysis also accounted for the percentage of false positives flagged by the systems, where the case has been flagged for potential identity theft, but verification has confirmed taxpayer identity.

  • The RRP system flagged returns with refunds worth $2.5 billion that the DDb system didn’t, with a 28.8 percent rate of false positives.
  • The DDb system flagged returns with refunds worth $1.15 billion that the RRP system missed, with a 63.4 percent rate of false positives.
  • The two systems combined flagged returns with refunds worth $1.94 billion, with an 8.1 percent rate of false positives.
  • The EFDS system flagged returns with refunds worth $86 million that neither the DDb or RRP systems found, with a 33.1 percent rate of false positives.

TIGTA offered four recommendations regarding the RRP and refund checks, including ensuring that RRP will find the same fraudulent returns that the previous systems found before implementing it.

TIGTA also recommended that IRS reissue refund checks after finding 113 cases where checks were returned from undeliverable addresses, but were not re-sent after recipients contacted the agency with the correct address.

Other recommendations focused on correcting data entries when an undelivered check hasn't been resolved after 30 days and developing policies to place an identity theft indicator on a paper refund check that hasn't been cashed after 14 months.

IRS officials agreed with three of the recommendations and partially agreed with the paper refund check recommendation, saying that it would have a further analysis of those checks in December 2016.

There is no scheduled date for the full implementation of the RRP system.

Share:
In Other News
Load More