In 2021, the difference between what U.S. taxpayers owed the federal government and what they paid hit nearly $700 billion. This “tax gap” has been growing for years while the Internal Revenue Service’s budget has shrunk, declining by 18% between 2010 and 2021. The audit rate for millionaires fell by more than 70% in roughly that same period.

To reverse these trends, the Biden administration’s Inflation Reduction Act allocated $80 billion in additional funding to the IRS over the next decade. If it’s used wisely, that money could help the agency track down some of those missing billions, says Rebecca Lester, an associate professor of accounting at Stanford Graduate School of Business and an expert on taxation. “Prior research shows high returns to IRS business enforcement: If you allocate money to the IRS for enforcement, then for every dollar you put in, you get more than a dollar back,” she says.

In collaboration with colleagues from the Stanford RegLab, Lester has posted a new paper that, for the first time, maps out the complex “spiderweb” of partnership structures some taxpayers set up to reduce their tax bills. The research also examines what tools the IRS could use to most efficiently track and evaluate these tax-planning schemes.

Lester notes that this paper marks the start of a longer-term project and builds on prior work by her collaborators Daniel Ho of Stanford Law School and Jacob Goldin of the University of Chicago. Their coauthors include Emily Black of New York University, Jonathan Hennessy of Stanford, Ryan Hess of the University of Georgia, Zaynah Javed of Stanford, and Annette Portz from the IRS. Lester and Ho are fellows at the Stanford Institute for Economic Policy Research (SIEPR).

Hidden figures

There are three fundamental business types in the U.S., Lester explains. At one end of the spectrum are sole proprietorships – individuals who run small businesses. On the other end are large publicly traded companies. In the middle are millions of pass-through (or flow-through) businesses, like limited liability corporations, whose income is generally not subject to taxation at the business level but passes directly to the owners and is taxed as personal income.

Partnerships are one type of pass-through business; they vastly outnumber public firms and control more than $30 trillion in assets. They are increasingly common, growing from 29% of U.S. business entities in 2003 to 40% in 2020. They are estimated to be responsible for three times as much U.S. tax non-compliance as corporations.

They are also very poorly understood. “Most academic research focuses on public companies for the simple reason of data availability,” Lester says. “But we think these flow-through businesses are where a very large portion of tax planning happens and that they contribute disproportionately to the tax gap.”

Lester and her co-authors shed new light on these partnerships by gathering their tax forms and linking them to their owners in the tax data. Through this process, they created one of the first comprehensive maps of partnership-related businesses in the United States.

Most immediately, they found that roughly 80% of the nearly 7 million partnerships they examined are directly owned by individual taxpayers. These “simple” partnerships appear to largely be used for business operating and investing purposes.

The other 20% constitute what the researchers call “complex” partnerships, intricate networks of relationships between individuals and business entities whose ownership may not be readily apparent. They may have multiple layers of ownership involving clusters of overlapping partnerships. Their income increases with complexity: Their operational income was almost twice that of simple partnerships, and their investment income was seven times that of simple partnerships.

These more byzantine structures, Lester suggests, could be used by some taxpayers for tax planning. “When you look at these spiderwebs of relationships, it’s very hard to imagine reasons that they would exist for purely legal ends,” she says.

Follow the money

Lester and her colleagues studied the relationship between a partnership’s complexity and whether the IRS will uncover unpaid taxes. In short, the more complex an organization is, the less likely the agency is to identify additional amounts due when it conducts an audit. One possible explanation is that complex cases are simply more difficult to review. The researchers find that complex audits are more likely to be closed quickly with no tax assessed, possibly because a speedy “no change” decision allows examiners to focus on other cases where assessments are more likely.

The researchers also found that when IRS examiners do find and assess additional tax to a complex organization, the amount assessed is much larger than in other partnership audits. “This is consistent with the idea that complexity is somehow indicative of the total amount of potential tax avoidance,” Lester says. “There is something about these partnerships that is not simply for the legal protection of a business.”

These large assessments contribute to an overall high audit return on investment, the amount the IRS recovers from a taxpayer as compared to the cost of the audit. The researchers found that the ROI of complex audits is very high – 20 times the audit cost. This rate is twice that of simple audits and eight times that previously found for corporate audits.

The infusion of funding from the Inflation Reduction Act could help the IRS bring more of these complex partnerships into compliance, though Lester emphasizes that the agency must be judicious in how it spends its resources. “If the government wants to more effectively enforce the tax system, it will be challenging for the IRS to hire enough partnership auditors, as many with the expertise work for big accounting or law firms,” she says. “We need to think of other strategies the IRS might deploy as they approach the problem.”

This research into pass-throughs could be a start. “Building from this paper, we hope to shed new light on the structures taxpayers use and bring new data tools about these partnerships to help the IRS to even more efficiently enforce the tax code,” Lester says.

For more information

This story was originally published by the Stanford Graduate School of Business.