Data: Stanford Institute for Economic Policy Research; Diagram: Axios Visuals
If you are a single black person with dependents claiming Earned Income Tax Credit (EITC), you have a 7.73% chance of being audited by the IRS in any given year . For all Americans, the corresponding figure is only 0.54%.
- Why is it important: Single men with dependents claiming the EITC are also screened at a very high rate if they are not black – 3.46% of them receive the dreaded letter in the mail – but that number still compares to that of their black counterparts.
The big picture: Black Americans are being screened at significantly higher rates at all income levels, according to a critically important new study by Stanford researchers in collaboration with the IRS.
- The study used extensive data from 2014 – “the most recent year for which the vast majority of audits were completed and made available to us,” the researchers write.
What they found: By far the greatest disparity is among Americans who claim the EITC. When it comes to EITC applicants, black applicants are much more likely to be selected than their non-black counterparts.
- “The racial audit disparity in EITC results contributes 78% of the overall disparity,” the researchers conclude.
Between the lines: As Emily Peck of Axios reported last year, EITC exams are much easier to pass than exams for high earners, although the wealthy tend to do a lot more tax evasion.
- Even under EITC audits, it is easier for the IRS to process applicants who have no business income. Black filers account for 21% of non-commercial EITC revenue, but only 11% of business EITC revenue.
- The IRS appears to effectively ignore business revenue: non-business audits account for 93% of EITC audits, although business dollar EITC revenue shows much more under-reporting.
How it works: The IRS does not observe race directly, and the selection for the EITC audit is algorithmic. But the design of the algorithm seems to have different effects on black Americans.
- In particular, the algorithm appears to be much more likely to select tax returns where the probability of even minimal underpayment is extremely high than those where underpayment is likely to be the highest.
- These latter returns are typically filed by high-income taxpayers and require significantly more work — and more experienced reviewers — than the relatively automated checks of non-corporate EITC applicants.
The bottom line: It looks very good that the problem of different algorithmic implications could be solved quite easily – by modifying the IRS filtering algorithm from a binary question “Is this tax return likely to be in error?” to one based on the expected dollar amount of the underpayment.
- However, this would mean an increase in the number of expensive accountants needed by the IRS. And Congress hated increase the IRS budget, even if it pays off many times over.
My thought bubble: The even better solution would be to not force everyone to file a tax return.
Source: www.axios.com
#tested
