IRS tax audits may show unintended racial, ethnic bias

The Internal Revenue Service doesn't collect data on taxpayers' race or ethnicity, but some of its methods could lead to disparities in its audit rate, according to a new report.

The report, released Tuesday by the Government Accountability Office, examined potential systemic biases in how the IRS chooses which returns to audit. A Stanford University study released last year found the IRS audits Black taxpayers at 2.9 to 4.7 times the rate of non-Black taxpayers. The disparity in audit rates seems to be due in part to automated algorithms that flag discrepancies in claims for tax credits.

The GAO report pointed out that the IRS uses an automated system to select returns that claim refundable credits, such as the Earned Income Tax Credit, for audit. The IRS regularly reviews this system, but doesn't consider potential unintended bias in the data and assumptions the system uses to make selections. The report recommends addressing this issue and others found by the GAO. 

Andrew Harrer/Bloomberg

The IRS's Wage & Investment Division audits returns claiming refundable tax credits, including the EITC and the Premium Tax Credit. The division uses a combination of manual and automatic processes to determine the number and types of audits it will conduct and relies on four objectives to develop its audit workplan. One of its objectives is the no-change rate, the percentage of returns that will yield no extra revenue after audit.

The IRS is trying to achieve a low no-change rate because that indicates it's auditing noncompliant taxpayers. However, the calculation of the no-change rate includes default audits, that is, audits closed as a "change" because taxpayers didn't respond to a notice or provided insufficient responses. 

IRS officials said their recent research found Black taxpayers are more likely not to respond to IRS correspondence than taxpayers of other races. The default audits also may be more common among low-income and EITC taxpayers, because of challenges that make communicating successfully with the IRS harder, such as being transitory or not having bank accounts. The IRS's W&I division uses past results to inform current audit planning, which could lead to W&I disproportionately selecting the types of returns that have historically resulted in filers' nonresponse rather than in confirmed noncompliance.

The main system that the IRS employs to pick specific returns for audit is its Dependent Database program, an automated system that flags returns for potential risk of noncompliance. The IRS regularly reviews this program, but it's review process doesn't comprehensively take into account all the data inputs and assumptions that could inform the agency about the demographic equity of the audit selection process, creating the potential for unintended bias in audit selection. For example, the GAO found that some risk scores contained in the DDB program vary by sex, which could skew selection, and have not been updated since 2001.

The GAO made six recommendations to the IRS, including to calculate the no-change rate without default audits, improve its reviews of audit selection processes, and use additional performance measures in assessing its selection systems. The IRS agreed with all of the GAO's recommendations.

"The IRS is committed to enforcing the tax laws in a manner that is fair and impartial," wrote IRS deputy commissioner Douglas O'Donnell in response to the report. "It is important to reiterate that the IRS does not and will not consider race in our case selection and audit processes. We are committed to improving fairness in tax administration."

He noted that's reflected in the IRS's Inflation Reduction Act Strategic Operation Plan for fiscal years 2023 to 2031, which the IRS recently updated to focus audits more on large partnerships and corporations and wealthy individuals.  O'Donnell noted that starting in fiscal year 2024, the IRS has substantially reduced the number of correspondence audits focused specifically on refundable tax credits like the EITC, American Opportunity Tax Credit, Health Insurance Premium Tax Credit and Additional Child Tax Credit. It has implemented a new scoring model for EITC audit selection and conducted an in-depth analysis of the EITC audit selection process, while also cracking down on unscrupulous tax preparers who convince their clients to underreport income or overclaim credits or deductions.