Among older workers, the majority of self-employment income missing from tax records is due to unreported earnings to the Internal Revenue Service rather than differences in how income is classified, a new University of Michigan study finds.
The analysis shows that the amounts of self-employment income reported in the Health and Retirement Study, but absent from tax data, are consistent with IRS estimates of unreported earnings. The HRS is a long-term research project at U-M that surveys Americans over age 50.
The IRS has determined through random audits that the net tax gap-the shortfall between taxes owed and taxes collected by the government-for the 2022 tax year (the most recent year for which data is available) amounts to an estimated $606 billion, for example.
For the U-M study, researchers analyzed the 2004-2016 Health and Retirement Study linked to administrative data, and identified significantly more self-employment income, aligning with IRS compliance estimates.
"While the IRS estimates are based on random audits, this research uses a different source, the HRS survey," said Joelle Abramowitz, associate research scientist at U-M's Institute for Social Research. "This research provides additional insight into the characteristics of people who underreport self-employment income to tax authorities as compared to the survey, the amounts they are underreporting and the types of work they are doing.
"The results were very surprising to me because a comparison of other surveys to tax data suggests that people are less likely to report self-employment income in those surveys than in tax data," Abramowitz said. "This survey might be better able to identify this information than others, and we can explore how to use this survey's methodology to improve other surveys to identify this information better."
Previous research shows that self-employment income is often underreported in administrative records and household surveys, affecting tax policy analysis.
"I find that the majority of the missing self-employment income reports do not reflect misreporting of income, but rather reflect income that is not being reported to the IRS," Abramowitz said. "I also find evidence that compared to people doing more formal self-employment, people doing more informal self-employment are less likely to report any self-employment income to tax authorities but report it on the survey."
Abramowitz says the missing self-employment income is concentrated among people reporting very little income to tax authorities and among the highest earners.
"For example, overall, people reporting work earnings to the IRS that are between the 5th and 75th percentiles underreported self-employment income of $4,000 on average," she said. "In contrast, those below the 5th percentile underreported self-employment income averaging $21,000 while those above the 95th percentile underreported self-employment income averaging $27,000, and this is a conservative estimate, as earnings above $250,000 are masked in the data."
The lost revenue limits the government's ability to lower its budget deficit or fund programs to improve people's lives, she says.
"These findings could inform how to target audits better," Abramowitz said. "They are also important because policies and regulations are based on survey and administrative data. This study finds that tax data and other surveys do not accurately identify self-employment income.
"Missing this income component affects estimates of poverty and inequality, which are used to understand well-being, as well as efforts to understand the effects of policies on these outcomes, and these in turn influence policymaking and the development of programs."
In ongoing work, researchers examine how discrepancies in reporting vary by race and sex, considering how different measures of income based on these reports affect the estimation of the poverty rate. They also examine how doing other kinds of self-employment affects well-being.
The study is published in the National Tax Journal.
