A new report documents systemic discrimination in how the I.R.S. selects taxpayers to be audited, with implications for a debate on the agency’s funding.

Black taxpayers are at least three times as likely to be audited by the Internal Revenue Service as other taxpayers, even after accounting for the differences in the types of returns each group is most likely to file, a team of economists has concluded in one of the most detailed studies yet on race and the nation’s tax system.

The findings do not suggest bias from individual tax enforcement agents, who do not know the race of the people they are auditing. They also do not suggest any valid reason for the I.R.S. to target Black Americans at such high rates; there is no evidence that the group engages in more tax evasion than others.

Instead, the findings document discrimination in the computer algorithms the agency uses to determine who is selected for an audit, according to the study by economists from Stanford University [Hadi Elzayn, Daniel Ho], the University of Michigan [Evelyn Smith], the University of Chicago [Jacob Goldin, Arun Ramesh] and the Treasury Department [Robin Fisher, Thomas Hertz] [Measuring and Mitigating Racial Disparities in Tax Audits]:

Abstract
Government agencies around the world use data-driven algorithms to allocate enforcement resources. Even when such algorithms are formally neutral with respect to protected characteristics like race, there is widespread concern that they can disproportionately burden vulnerable groups. We study differences in Internal Revenue Service (IRS) audit rates between Black and non-Black taxpayers. Because neither we nor the IRS observes taxpayer race, we propose and employ a novel partial identification strategy to estimate these differences. Despite race-blind audit selection, we find that Black taxpayers are audited at 2.9 to 4.7 times the rate of non-Black taxpayers. The main source of the disparity is differing audit rates by race among taxpayers claiming the Earned Income Tax Credit (EITC). Using counterfactual audit selection models for EITC claimants, we find that maximizing the detection of underreported taxes would not lead to Black taxpayers being audited at higher rates. In contrast, in these models, certain policies tend to increase the audit rate of Black taxpayers: (1) designing audit selection algorithms to minimize the “no-change rate”; (2) targeting erroneously claimed refundable credits rather than total under-reporting; and (3) limiting the share of more complex EITC returns that can be selected for audit. Our results highlight how seemingly technocratic choices about algorithmic design can embed important policy values and trade-offs.]

Some of that discrimination appears to be rooted in decisions that I.R.S. officials made over the past decade as they sought to maintain tax enforcement in the face of budget cuts, by relying on automated systems to select returns for audit.

Those decisions have produced an approach that disproportionately flags tax returns with potential errors in the claiming of certain tax credits, like the earned-income tax credit, which supplements low-income workers’ incomes in an effort to alleviate poverty. Those tax returns are more often selected for audits, regardless of how much in owed taxes the agency might recover.

The result is audit rates of Black Americans that are between three and five times the rate of other taxpayers, even when comparing that group to other taxpayers who also claim the E.I.T.C.

In effect, the researchers suggest that the I.R.S. has focused on audits that are easier to conduct and as a result, finds itself disproportionately auditing a historically disadvantaged group rather than other taxpayers, including high net-worth individuals.