In the months following the release of joint research finding Black taxpayers are more frequently audited than others, the tax community has engaged in an ongoing conversation on the agency’s race-neutral stance, biases in enforcement algorithms, and what can be done to promote equity amid systemic issues.
The Stanford paper.
In January 2023, a mixed team of academics and IRS researchers collaborated at the Stanford University Institute for Economic Policy Research to release a working paper titled “Measuring and Mitigating Racial Disparities in Tax Audits.” The paper’s core finding that made its way through mainstream media was that Black taxpayers have their tax returns audited between 2.9 to 4.7 times the rate of non-Black taxpayers.
Earned income tax credit (EITC) claims were identified as the main driver of this disparity. That is, Black EITC claimants were audited far more than others claiming the EITC. The researchers point to the IRS’ internal audit selection algorithms for this, as they are designed to identify erroneous refundable credit claims instead of focusing on the highest amounts of underpaid tax. Minimizing the so-called no-change rate, or the rate in which returns picked for examination do not result in a change, is the priority for compliance algorithms. In other words, the IRS favors auditing the highest amount taxpayers they think, intentionally or unintentionally, overstated tax benefit claims rather than recouping the most uncollected tax possible.
The IRS does not ask for racial identifying information on tax returns, and this is by design to maintain an air of colorblindness in which the agency’s activities are not influenced by demographics, but sheerly on a letter-of-the-law basis for compliance purposes. The Stanford team needed to develop its own predictive models for estimating races of taxpayers who were audited for the study, in this case tax year 2014.
Jake Goldin of the Treasury Department’s Office of Tax Policy and co-author of the Stanford paper said at a May conference in Washington, D.C., hosted by the American Bar Association that outside data needed to be used to impute race absent direct information from the IRS. For example, outside data sets where race was self-reported was relied upon, such as the North Carolina voter registry.
Goldin in recalling the team’s process in conducting its research explained that they were clued in that something “might be going on” when they were confident in which taxpayers were Black or not. “We know that Black and non-Black taxpayers in the U.S. have very different income distributions,” he began. “We know that the IRS audits people differently based on their income. And what you see is that at every income rate, no matter where you look, Black taxpayers are audited at higher rates than non-Black taxpayers. So it’s not just income differences that are driving this finding is what this tells us.”
Because the EITC is a low-income tax benefit the IRS believes is frequently overstated on returns (arguably because of complex eligibility requirements that require supporting documentation from taxpayers—and because of temporary changes made to the program during the COVID-19 pandemic), one might assume the disparity emerges simply from the focus on EITC claims combined with the racial wealth gap.
“It turns out that’s not mostly what’s going on,” Goldin said, “and so this was surprising to us.” He reiterated that “something else is going on” in light of the result that Black taxpayers are audited more even within the EITC claimant subgroup. According to Goldin, “if the IRS were to select taxpayers based on the predicted amount of underreporting on the returns, then they will select Black EITC claimants at lower rates than non-Blacks … that’s the opposite of what we see today.”
“These algorithms, even when they are ‘neutral’ … can end up exacerbating existing sources of disadvantage in society.”
Another co-author on the Stanford team, Tom Hertz of the IRS Research, Applied Analytics and Statistics (RAAS) Division, presented subsequent, follow-up work at a joint conference held in June by the IRS and the Urban-Brookings Tax Policy Center. Hertz and a team of three other agency officials, “basically replicated the headline finding and confirmed that it’s relatively stable over time.”
Hertz echoed Goldin’s note that Black EITC claimants would be audited less if the IRS’ models pursued a different objective in capturing revenue and not reducing the no-change rate. “But it’s important to be clear that the Stanford revenue estimates only hold if all audits are [National Research Program] audits, which are full scope audits and they take about 18 hours apiece,” Hertz said,” whereas the vast majority of EITC audits are correspondence audits with a limited scope. They take about an hour and a half, so that gap is big enough to be able to conclude that the trajectory in the Stanford paper and this sort of approach to reducing disparity is not a feasible trajectory, just simply cut and pasted to operational reality.”
Government response.
Senate Finance Committee Chair Ron Wyden, Democrat of Oregon, tasked IRS Commissioner Danny Werfel, to immediately look into disparate audit selection upon being confirmed by the Senate to the position. Werfel essentially corroborated the Stanford paper’s conclusion.
“While there is a need for further research, our initial findings support the conclusion that Black taxpayers may be audited at higher rates than would be expected given their share of the population,” Werfel wrote to Wyden in a letter May 15. “We are dedicating significant resources to quickly evaluating the extent to which IRS’s exam priorities and automated processes, and the data available to the IRS for use in exam selection, contribute to this disparity. As part of this work, we are evaluating the potential impact of methodological changes to case selection (e.g., optimizing on broader tax issues rather than focusing on EITC overclaims).”
He went on to pledge to deploy resources from the 10-year, $80 billion appropriation under the Inflation Reduction Act (PL 117-169) to address “any bias that exists” in the audit program, across age, gender, geography, race, and ethnicity. He added that evaluating EITC audit selection algorithms is ongoing, but “the topmost priority within this larger body of work.”
Wyden seemed satisfied with Werfel’s acknowledgement and willingness to act. “The IRS is currently in the process of overhauling its approach to tax enforcement with the funding it received in the Inflation Reduction Act,” Wyden said in a statement. “I’m going to ensure that as part of that overhaul, the IRS retools these algorithms to eliminate any racial bias. Commissioner Werfel has committed to implement changes to fix this issue before the next tax filing season begins, but it’s clear there’s more to learn and a lot of work to be done. As Chairman of the Senate Finance Committee I expect the IRS to provide regular, public updates on their progress.”
Tax community weighs in.
At the same ABA panel where Goldin walked through the Stanford paper, Howard University School of Law Professor Alice Thomas stated: “Racial neutrality is not what we need.”
“We need race-conscience targeted policies to get us out of the hole. There’s no surprise at what we’re finding because behind the algorithm is a person. That person might have racial bias unless they untrained themselves not to be racially biased.”
Given that the IRS’ neutrality led to disparities anyway, and given Thomas’ personal experience, she said she does not need data or tables to put know what inequity looks like. “When I looked at the question of what might be going on here, what we’re discovering in the Tax Code, it took me all the way back through history, because I want to let people know I’m a descendant of people who where formerly enslaved in this country,” she told an audience of mostly practitioners and government officials. “I am not an immigrant. I am not a migrant. I am a descendant of captives … and we are still searching for our racial equity in America.”
The question of whether the IRS should collect race data to help with transparency and accountability in audit selection has been often raised in the fallout of the Stanford study. One audience member asked Thomas if she thinks this information should be asked on returns. She responded that for many Black taxpayers, there would be an assumption that self-reporting their race would “run downhill” and result in a “negative consequence.”
“If we put on that tax return: ‘check and say you’re Black,’ would you check it? We don’t trust you. I’m not going to check ‘Black.'”
Amara Enyia, Policy and Research Manager for the Movement of Black Lives, believes it is worth pursuing such demographic information on returns in light of the “myth” of algorithmic objectivity. “There’s a saying: ‘Measure what you treasure,'” she told Checkpoint in an interview. “When they don’t collect racial demographic data, it can have the effect of obscuring what’s actually happening with the tax system.” She added that the Stanford researchers would not have needed to find “roundabout ways of getting to the data that shows these outcomes” if it were more quickly accessible.
“When it comes to data collection, sometimes the unwillingness is not because of lack of resources or that it’s too difficult,” Enyia continued. Not collecting the data “allows for things to fly under the radar and for issues to not necessarily get addressed—[for] it to be front and center. We’re going to have to continue pushing because there’s always going to be some resistance … [The] findings are not easy to take because it’s showing that there are clear disparities in the agency, whether intentional or unintentional,” and the IRS should take responsibility and “own up to” and “face those outcomes that are being illuminated and to set about adjusting them.”
Enyia said she thinks Werfel taking over as commissioner sets conditions that “are perhaps more favorable than in the past” but “we can’t let our guard down. Why? Because at the end of the day, the tax system is one of the few places where you can actually create more equitable outcomes. But you have to be intentional about it. We can’t get to those outcomes if we have a system that creates disparities … which is what we have.”
On the issue of potential resistance, Indiana University Kelley School of Business Professor Goldburn Maynard Jr. said at an American Tax Policy Institute event in late February that when race-conscience policies are implemented, a side effect is “you actually increase bias and harassment against said groups.”
“In some sense, perversely, race-conscience efforts actually help oligarchs because they provide a distraction and also a powerful way for the rich to unite with the lower classes.” He claimed that the wealthy seek to sow division and maintain the status quo when race-conscience policies are put forth.
David Rubenstein Fellow Keon Gilbert at a June 8 virtual forum hosted by Brookings offered that these more powerful figures, such as corporations, can play in a role in positively shaping policy “because they are at the center of so much of what is happening. They benefit from policies that allow them to not pay taxes, but communities do not benefit from them not paying taxes. We really have to think about what is economic growth or challenge what economic growth in a particular community looks like.”
As part of its Strategic Operating Plan for spending the money authorized under last year’s inflation bill, the IRS hopes to expand community outreach in underserved areas to bridge gaps in access to tax assistance. Leslie Book, Professor at the Villanova University Charles Widger School of Law, at the same panel alongside Maynard accused the IRS of not truly embracing its role as a benefits administrator, however. He called it a “conscience decision” to omit this aspect of the IRS’ function from its mission statement.
Regarding the IRS’ steadfast position on what George Washington University School of Law Professor Jeremy Bearer-Friend described as “colorblind tax enforcement” in a paper published last year, Book claimed “there is no such thing as a nonracist, race-neutral policy in light of the unequal footing that different racial and ethnic groups experience.”
He argued that the IRS “becoming a force for antiracist policies” and taking part in progressing towards racial justice would have an indirect benefit of increasing trust in government. This would “facilitate greater tax compliance and reduce the tax gap,” according to Book, who added that “policymakers who commit to emphasizing social justice should directly acknowledge that these issues are not part of a zero-sum game where helping some means hurting others.”
Book in closing said he would start with looking at the tax return filing process. He noted the IRS Free File Alliance, where the IRS partners with third-party tax preparation software providers to let taxpayers under a certain income threshold complete and submit returns at no cost, is severely underutilized. A report from the Government Accountability Office estimated that over 70% of individuals are eligible to file federal income tax returns using Free File yet only 3% actually use it.
Tax filing and race equity.
At the end of this tax filing season, Werfel announced the IRS is developing tax filing tool to allow taxpayers to complete their returns without the need of a middleman. Next tax season the IRS will launch a pilot program for what it is calling Direct File, an optional trial run certain taxpayers can opt into. More specifics on who will be eligible will be made available before then, but Werfel indicated some simpler returns can be pre-populated, likely meaning some taxpayers with straightforward income streams such as through a single Form W-2 can be told what their refund or tax liability is without much of the legwork—and sign off on it free of charge.
After the announcement, 200 national, state, and local organizations came together to join a letter supporting the IRS development of a direct file tool. The letter was sent by the Coalition for Free and Fair Filing, a coalition spearheaded by the Center for the Study of Social Policy, Code for America, the Economic Security Project, Groundwork Action, NETWORK Lobby for Catholic Social Justice, Public Citizen, and RESULTS.
In the July 19 letter to Werfel, the group expressed its staunch support of Direct File, which it hopes will undercut predatory practices by major players in the tax prep industry, especially Intuit, which made settlement payments to customers in a legal battle over misrepresentations of the “free” version of TurboTax.
“Today, it is far too difficult for American families to file tax returns,” read the letter. “The IRS estimates that an average taxpayer spends 13 hours and $250 annually fulfilling their filing obligations. These families miss out on critical benefits administered through the Tax Code, like the [EITC] and the Child Tax Credit [CTC]. Moreover, the burden of these high costs and missed benefits fall disproportionately on people of color and other communities who are marginalized and excluded from support …”
The Economic Security Project released a supplemental fact sheet expanding more on how the tax filing system disfavors families of color. Pulling from IRS data on the agency’s website, the document states about one in five eligible families don’t receive their EITC due to filing barriers. “Nearly every part of the tax code interacts with race and ethnicity in ways that aren’t written into the code but which have very real racially disparate consequences,” it continues. “Some policies, like the EITC and CTC carry the promise of outsized impacts for households of color, but the lion’s share of tax policies actually create significant disadvantages to people of color compared to white people—from marriage bonuses to preferential tax treatment for capital gains, homeownership, and retirement savings.”
Economic Security Project’s Vice President of Campaigns and Political Strategy Adam Ruben told Checkpoint in an interview that his organization’s involvement on the issue of government-automated tax filing and later participation in the Coalition began with work stemming back to 2017. Then, they were looking at refundable tax credits like the EITC and the CTC, including at the state level.
“We were trying to figure out how do you put more money back in working, middle-class people’s pockets and we realized … ‘Okay, these tax credits are one of the kind of tools that we have to do that'” and “had been very successful, but we could make a much bigger and more inclusive way of reaching more people and also boosting the racial equity of those credits,” Ruben said.
He added that this led to the conclusion that a simpler tax filing system where federal and state governments are more proactive and pay “out the credits to you automatically” is necessary for cutting filing costs and time for low-income taxpayers and taxpayers of color. Ruben recalled before the pandemic, staffers on the House Ways and Means and Senate Finance Committees were “literally laughing out loud at the idea” of the IRS just sending checks to eligible taxpayers. The pandemic proved the IRS’ decades-old systems could be updated to issue stimulus checks and monthly CTC checks, which made the prospect of Direct File more feasible and potentially a reality, according to Ruben.
After the announcement of the pilot, the Economic Security Project reached out to other organizations that had been working on refundable credits, as there was already a sort of pre-installed network of different groups at various levels, he explained.
“I think even though the Tax Code is race-neutral on its face, its impacts aren’t and a lot of families of color are locked out of valuable tax benefits, like the [CTC] and the [EITC],” said Ruben. “It’s stressful, particularly so for low-income families and people with language barriers.”
Ruben and the Coalition estimate 87 million taxpayers “lose as much as $13 billion of their refunds to tax prep companies for something that ought to be free” and that 76% of surveyed taxpayers support Direct File despite claims from the industry that it is unpopular.
“Filing taxes is complicated now, but it doesn’t have to be.”
Get all the latest tax, accounting, audit, and corporate finance news with Checkpoint Edge. Sign up for a free 7-day trial today.
