Using Analytics to Close the Nonfiler Employment Tax Gap
Oct 21, 2019
One of the six strategic goals of the Internal Revenue Service (IRS) is to encourage compliance and close the tax gap, which is the amount of true tax liability that is not paid voluntarily and timely. The IRS estimates that the nonfiler tax gap accounts for $32 billion of the $458 billion gross tax gap each year. However, the IRS does not include employment tax in that estimate, and in a recently released report, the Treasury Inspector General for Tax Administration (TIGTA) estimates an additional $14.7 billion worth of assessments for employment tax nonfilers.
Under Internal Revenue Code Section 6020(b), the IRS is authorized to determine and assess a tax liability for a taxpayer that has a filing requirement but fails to file a tax return. The BMF Case Creation Nonfiler Identification Process identifies taxpayers that are nonfilers and creates case inventory that can be worked by nonfiler programs. One of these programs is the Automated 6020(b) program, which allows the IRS to automatically prepare a return to send to a taxpayer, who can respond by accepting the completed return, file their own return, or prove they do not have a filing requirement. The A6020(b) program assessed $2.2 billion in employment taxes from FY 2013 to FY 2017 and collected $719 million in employment tax revenue, closing 667,161 cases. However, significant reductions in resources, allocated to both BMF case creation and the A6020(b) program, have resulted in a reduction in case inventory to work from and a 92% reduction in A6020(b) case closures from 261,582 in FY 2013 to just 21,746 in FY 2017. For almost three years, since FY 2017, no new cases have been worked by the A6020(b) program.
Facing these resource constraints, analytics can be used in BMF case creation to identify potential nonfilers, then optimize existing resources in the nonfiler programs by allocating cases to the most effective nonfiler programs to work them. The majority of cases closed in the A6020(b) program result in the assessment determined by the IRS, so the program has the ability to reach hundreds of thousands of taxpayers and bring them into compliance using less resource-intensive automated processes. Using machine learning and predictive analytics to assign cases to the A6020(b) program that are most likely to come into compliance, more complex cases can be worked by other more resource intensive programs. This would enable to IRS to bring more taxpayers into compliance and help close the nonfiling tax gap.