The Internal Revenue Service (IRS) estimates that taxpayers failed to pay 15% of their legal tax liability on time from 2014 to 2016. Late payments and enforcement actions recoup an additional 2% of the public's true tax obligation. Unpaid taxes contribute to the United States' federal deficit, and their accumulation over time may require policymakers to raise more revenue from those who comply with the tax code. This report describes the sources of unpaid federal taxes—also known as the tax gap—and lends context to the measure's implications by presenting information on how the IRS constructs it. This report also describes popular proposals to reduce the tax gap.

Key Elements of the Tax Gap

The tax gap is commonly measured in two ways:

• The gross tax gap is the amount of tax that individuals and businesses owe but do not pay on time, absent additional collections from enforcement action. From 2014 to 2016, the average annual gross tax gap was $496 billion, or 15% of true tax liability.¹
• The net tax gap is the amount of tax that taxpayers owe but do not pay, after including late payments and collections through enforcement actions. From 2014 to 2016, the average annual net tax gap was $428 billion, or 13% of true tax liability.²

The tax gap reflects the estimated amount of unpaid federal taxes. The estimate does not account for the budgetary costs to the federal government of administering the tax code or compliance costs borne by taxpayers, such as the cost of their time preparing and filing returns and their efforts to record their taxable income and qualifications for deductions and credits.

There are three ways individuals and businesses can fail to pay the taxes they owe:³

• 1. Underreporting of Tax: A taxpayer (e.g., an individual or business) files their tax return, but underreports their true tax liability. Underreporters may either report that they received less income subject to taxation than they actually did, or claim more in deductions, credits, or other tax benefits than the law allows, or some combination of these two. According to the IRS, underreporting comprised $398 billion of the average annual gross tax gap between 2014 and 2016, as illustrated in Figure 1. This represented 80% of the $496 billion gross tax gap. The majority of the underreporting tax gap, $278 billion, is attributable to underreporting of the individual income tax. Underreporting by pass-through businesses (which pay the individual income tax) was responsible for an estimated $130 billion or 26% of the gross tax gap.
• 2. Underpayment of Tax: A taxpayer files their tax return and correctly reports their tax liability, but then underpays the IRS. Underpayment comprised an estimated 12% of the annual gross tax gap from 2014 to 2016.
• 3. Nonfiling of Returns: A taxpayer fails to file a tax return when they are required to do so. Nonfiling represented an estimated 8% of the annual gross tax gap from 2014 to 2016.

Figure 1. Sources of Underreporting Tax Gap Underreporting accounted for an estimated 80% of the average annual gross tax gap from 2014 to 2016.

Source: IRS, Federal Tax Compliance Research: Tax Gap Estimates for Tax Years 2014-2016, Publication 1415, October 2022. Notes: Totals may not add due to rounding. Underreported tax liability accounted for 80% of the total gross tax gap. "Large" corporations are those with gross assets worth more than $10 million. The IRS estimated that estate tax underreporting cost the government an average of an additional $1 billion in lost annual revenues from 2014 to 2016.

In addition to formally estimating the tax gap and analyzing its components based on data that may take years to generate and analyze, the IRS publishes estimates of the tax gap for more recent years. The IRS projects the gross tax gap was $688 billion in 2021, $601 billion in 2020, and averaged $550 billion from 2017 to 2019. It also projects the net tax gap was $625 billion in 2021, $539 billion in 2020, and averaged $481 billion from 2017 to 2019.⁴ These estimates incorporate more recent administrative data on the nonfiling tax gap, but generally assume that underreporting (the main driver of the tax gap) continued at the rates estimated for 2014-2016 and that underpayment continued at the rate observed from 2017 to 2019.⁵ More detail on methodology is available in the Appendix.

Analysis

Available data indicate that tax nonpayment is more common for taxes on sources of income (and for tax expenditures) for which there is limited third-party information to verify claims. In other words, if taxpayers know the IRS has no information on a form of income, they may be more likely to underreport income from that source. In addition, research suggests that income underreporting is more common among the highest-income taxpayers, who may be more likely to receive these less-visible forms of income.

Distribution of the Tax Gap by Source of Income

Taxpayers are likelier to report their income subject to tax accurately when a third party, such as an employer or financial institution, also reports that income to the federal government. The IRS estimates that individual filers report 99% of income earned through wages or salaries accurately. For such income, employers not only report workers' compensation to the IRS, they also withhold workers' estimated income tax liability from their paychecks, and remit those taxes to the IRS. In contrast, filers underreport 55% of their income that is not subject to information reporting, such as that earned from sole proprietorships, rents, and royalties (Figure 2).

Figure 2. Estimated Share of Income Unreported, By Visibility of Income

Source: IRS, Federal Tax Compliance Research: Tax Gap Estimates for Tax Years 2014-2016, Publication 1415, October 2022. Notes: Income subject to substantial reporting and withholding includes wages and salaries. Income subject to substantial reporting, but not withholding, includes pensions and annuities, unemployment compensation, dividend income, interest income, state income tax refunds, and taxable Social Security benefits. Income subject to some reporting includes partnership and S corporation income, capital gains, and alimony income. Income subject to little or no reporting includes income from sole proprietorships, rents and royalties, farm income, Form 4797 income for the sale of business property, and other income.

Third-party information reporting discourages noncompliance, since the IRS can easily check whether one's reported income is accurate. The reporter also typically sends the payer a record of their reported income, which generally ensures the payer knows their income from that source and files their return accurately. Withholding estimated tax owed from the payer's income further likely ensures the filer will pay all legally required taxes on their income.

The different reporting rates for different sources of income suggest that individuals with similar total incomes that come from different sources may have different likelihoods of underpaying their federal tax liability. For example, an engineer who earned a $500,000 salary would have that income reported to the IRS by their employer, and the employer would withhold the appropriate income tax and remit it to the IRS, making the engineer much more likely to pay their income taxes on that salary than a sole proprietor who earned $500,000 through their business.¹² Similarly, a worker who received their compensation in cash with no accompanying information return may be more likely to underreport their income than a worker who earned the same amount through wages that their employer reported to the IRS.

The IRS identifies underreported income through the National Research Program (NRP), which conducts semirandom audits.¹³ However, NRP audits miss some underreporting from opaque income sources, particularly if some auditors are more experienced or skilled than others. The IRS tries to account for unreported income its NRP auditors miss through a process called detection-controlled estimation (DCE). Under DCE, the IRS effectively estimates how much underreporting all auditors would have detected for a given source of income had they been as successful as the most successful auditors, subject to other limitations. For example, if the IRS estimated that some auditors found 5% underreporting for a given income source while others found 10%, the IRS would assume the true rate is 10%.¹⁴ Independent research suggests applying DCE methodology roughly tripled the estimates of underreported income from 2006 to 2013, primarily from less-visible income sources.¹⁵

Distribution of the Tax Gap by Income Level

While the IRS does not estimate the distribution of the tax gap by taxpayers' income levels, research has found that those with high incomes are responsible for most misreported income. Johns and Slemrod (2010) studied the income distribution of unpaid taxes using NRP audits from 2001. They found that, on average, the highest-income 5% of households were responsible for 51% of all misreported adjusted gross income.¹⁶ DeBacker et al. (2020) found that NRP returns from 2006 to 2014 showed a similar distribution.¹⁷

These estimates may understate the underreporting of income by the highest-income taxpayers. Guyton et al. (2021) estimated that random audits such as those conducted through the NRP often fail to detect evasion through forms common among high-income individuals, such as offshore bank accounts and pass-through businesses. The researchers found that operational audits of a sample of the 0.01% of taxpayers who reported the most income in 2010 found more noncompliance than the tax gap estimation methodology suggested existed among that entire group. This finding suggests DCE failed to identify significant nonpayment among the highest earners.¹⁸ The researchers also found that NRP audits of taxpayers failed to identify offshore wealth for 93% of filers who soon thereafter reported international bank accounts because of unrelated changes in tax policy. This amount exceeds what normal annual turnover in offshore account ownership could explain.¹⁹

The authors' stylized model estimated that the lowest-earning 50% of taxpayers fail to report 7% of their income, while most of the highest-earning 1% fail to report 20%-25% of their income. The underreporting percentage then falls to 15% for the highest-earning 0.01%.²⁰

Policy Options for Reducing the Federal Tax Gap

Policymakers may seek to identify policy changes that would encourage taxpayers to comply with the tax code. Due to the limitations with tax gap estimates, such policies may or may not reduce the measured tax gap, even if they improve tax compliance.

Expanding existing mechanisms, including both audit and nonaudit options, could further encourage enforcement. Given that taxpayers underreport income from sources that withhold and remit taxes to the IRS much less often than income from other sources, policymakers could also look for opportunities to encourage or require withholding for more sources of income. Other options include simplifying the tax code, expanding taxpayer services, and strengthening regulations on paid tax preparers to prevent errors.

Expand Existing Enforcement Mechanisms

The IRS could increase the frequency and expand the scope of existing enforcement mechanisms meant to encourage compliance with the tax code, including audits, math error authority, and information reporting. Enforcement actions can generate revenue directly. In 2022, the Congressional Budget Office (CBO) projected that increasing spending at the IRS by $79.6 billion—including $45.6 billion for enforcement activities—through the Inflation Reduction Act (IRA; P.L. 117-169) would generate $203.7 billion in additional revenue over the same period.²¹

Enforcement may also raise revenue indirectly, by encouraging voluntary compliance with the tax code. If taxpayers believe that the IRS is likely to catch evasion or error, they may be more likely to accurately report their liability on time. If so, enforcement would reduce the tax gap by more than the amount collected through the enforcement action itself.

Audits (Examinations)

The IRS selects some tax returns for targeted examinations that are meant to ensure that the taxpayer has accurately reported their tax liability. In an examination, taxpayers provide the IRS with additional information on their income or eligibility for credits or deductions, which the agency uses to verify or correct the taxpayer's filing. Examinations empower the IRS to verify that taxpayers are paying their true tax liability, which can raise revenue directly. Taxpayers' awareness that they may be subject to an audit might also encourage voluntary compliance.

The Costs and Benefits of Audits

Audits may encourage voluntary compliance with the tax code by affirming to taxpayers that the IRS is likely to catch evasion or error. If so, enforcement would reduce the tax gap by more than the amount collected through the enforcement action itself.

Being audited appears to encourage some taxpayers to better comply with the tax code in subsequent years. A 2023 study found that taxpayers of all income levels who are subject to an in-person audit report more income on average for at least 14 years after the audit took place.²⁴ Additionally, a Spanish study showed that Spanish businesses avoided growing to a size that would make them more likely to face an examination, suggesting that the possibility of an audit may affect the behavior of taxpayers who have never faced an audit themselves.²⁵

In some circumstances, audits may discourage voluntary compliance. For example, in the year after a random audit, self-employed taxpayers who were found to owe additional tax reported roughly 64% more income, yet those who the audit cleared reported 15% less. That decline deepened to 21% three years after the audit.²⁶

Another study also found that those with self-reported income initially claimed more income in the year after a random audit than a control group did, but reported less than that control group within five years. Potentially sophisticated taxpayers—namely older taxpayers, those in the top income quintile, and those who used professional tax preparers—followed a similar pattern of initial increase and subsequent decline. The fact that audits were random may have kept taxpayers from thinking the audit indicated their preexisting behavior invited close scrutiny. If so, they would be more likely to return to preexisting evasion (or possibly to pursue it more aggressively) once the audit ended.²⁷

Research on preaudit contacts from tax agencies has found similar results. One study of Minnesota taxpayers showed that those with high incomes, and particularly with self-employment or farm income, reported less income after receiving a letter stating their return would be "closely examined." The authors theorized this could be because high-income filers may see less value in reporting certain income to avoid an audit once they view an audit as inevitable.²⁸

Because the tax gap represents unmet legal obligations, some may consider the existence of any gap to be a policy problem deserving of remediation. However, others argue that the cost of eliminating it entirely would outweigh the benefits. Slemrod summarized this drawback by saying that "just as it is not optimal to station a police officer at each street corner to eliminate robbery and jaywalking completely, it is not optimal to completely eliminate tax evasion."²⁹

The tax gap does not account for the inevitable costs associated with enforcing and complying with the tax code. The IRS spent an average of $4.8 billion per year on enforcement from 2014 to 2016, the years of the tax gap estimate.³⁰ The CBO projects that Congress will appropriate approximately $68 billion to the IRS from FY2024 to FY2033 for enforcement of the tax code, in addition to much of the $45.6 billion expenditure Congress made for tax enforcement through the IRA.³¹ Complying with the tax code also creates costs for taxpayers, such as the time required to calculate one's tax liability and prepare and submit forms to the IRS (or pay a preparer to do so), which the tax gap does not reflect.

Trend in Audits and the Inflation Reduction Act

The IRS has said it intends to use some of the enforcement funds it received through the IRA to expand audit coverage for large corporations, partnerships, and filers with high incomes or large wealth, and for other "areas where audit coverage has declined," such as employment taxes, excise taxes, and estate taxes. The agency says it will not audit households and small businesses earning below $400,000 per year at a higher rate than it did in 2018, the most recent tax year for which the agency has completed audits.³²

These audits would reverse a decline in audit rates from 0.9% of domestic individual income tax returns with positive income in tax year 2010 to 0.3% in 2019.³³ Audit rates fell most among high-income filers.³⁴ These decreases followed a 26% fall in inflation-adjusted outlays on enforcement from FY2010 to FY2022.³⁵

The Fiscal Responsibility Act (P.L. 118-5) rescinded $1.4 billion in funding for enforcement and another spending priority funded by the IRA, operations support. The White House reportedly also agreed to rescind an additional $10 billion each during the FY2024 and FY2025 appropriations cycles.³⁶

Math Error Authority

Congress could provide the IRS with expanded "math error authority," which would allow the IRS to correct certain items on a tax return before a refund is issued. Math error authority is the legal right to assess tax that exceeds what a taxpayer put on their filing without a formal deficiency notice under certain specific circumstances. For example, taxpayers can calculate their tax liability incorrectly, make a clerical error, or provide an incorrect Social Security Number or Taxpayer Identification Number. Taxpayers have 60 days to request that the IRS abate or reverse a correction that it made using math error authority. In such cases, the IRS automatically abates any additional tax and does not require taxpayers to submit any additional information along with this request. However, the IRS can still then conduct a traditional examination.³⁷

Proponents argue that expanding math error authority empowers the IRS without the IRS and the taxpayer incurring as much time and paperwork burden as they would under a normal deficiency notice. Past Administrations have at times requested that Congress expand the IRS's math error authority to enable the agency to fix more issues promptly.³⁸

Some observers, such as the National Taxpayer Advocate (NTA), have expressed concerns that the math notice process gives taxpayers fewer rights to contest a decision than the examination process does, particularly because many find the IRS's math error notices unclear.³⁹ While taxpayers typically have 90 days to contest notices of deficiency, they have only 60 days to contest math error notices.⁴⁰ The NTA has also said that applying math error authority to complex tax circumstances could increase the chances of IRS error that could burden taxpayers.⁴¹

Information Reporting

Lawmakers could expand information reporting on activities into which the IRS has little visibility. To do so, they could require third parties to provide the IRS with more information the agency can use to automatically verify the claims taxpayers make on their tax returns, including the amount of income they earn as well as their eligibility for credits or deductions.

Information reporting enables the IRS to identify more income underreporting or improper tax expenditure claims routinely using computing technology, without the need to conduct an examination. Further, if taxpayers know the IRS can automatically check whether the taxpayer has misrepresented themselves on their filing, they may be less likely to attempt to do so. As noted in the section "Distribution of the Tax Gap by Source of Income," the majority of the tax gap results from taxpayers underreporting their tax liability, and underreporting is most common for opaque sources of income.

Critics of expanded information reporting have expressed concerns. Some worry that the IRS would violate taxpayers' privacy by collecting too much information on them.⁴² Others say that the requirements would put undue burdens on the third parties who would report the information, or on the IRS, which might struggle to process the information on its antiquated information technology systems.

For example, the Patient Protection and Affordable Care Act (P.L. 111-148) included a requirement that businesses report payments to corporations in excess of $600, as they must do for payments such as salaries, interest, and profits. However, lawmakers later repealed this requirement before it took effect.⁴³ President Obama cited the paperwork burden the requirement could have placed on small businesses as his motivation for repealing the requirement.⁴⁴

Recent Expansions

Congress has attempted to expand information reporting in recent years. P.L. 117-58, the Infrastructure Investment and Jobs Act, required cryptocurrency brokerages (popularly known as "wallets") to report their customers' profits or losses using Form 1099-B starting in tax year 2023. It also required businesses to report cryptocurrency transactions of greater than $10,000 to the IRS, the same rule that applies to cash transactions. These rules will give the IRS more information on cryptocurrency transactions, making evasion of any federal taxes on such transactions more difficult. The IRS's tax gap estimate may not reflect revenue generated through these reforms because it does not account for evasion done using cryptocurrency.

Congress also required third-party payment services (e.g., Paypal, Venmo, and CashApp) to report more information on commercial transactions conducted through their platforms as part of P.L. 117-2, the American Rescue Plan Act of 2021. Federal tax law did not previously require these services to report the value of inflows and outflows to their customers' accounts if those flows were below $20,000 annually or involved fewer than 20 transactions. Under the new rule, the services must report inflows and outflows on all accounts with gross flows above $600. The rule only applies to payments for goods or services.⁴⁹ This reporting requirement will give the IRS more information on some taxpayers' business income, among the largest components of the tax gap. Congress scheduled this change to take effect for tax year 2022. However, the IRS has delayed implementation for two years "to facilitate an orderly transition" to the new system, and intends to use a threshold of $2,000 rather than $600 in 2024.⁵⁰

Proposals for Further Expansion

The GAO has encouraged the IRS to establish centralized leadership capable of assessing and adapting its use of information returns to meet the agency's needs.⁵¹ Additionally, the Treasury Department has proposed requiring American banks to report more information on their foreign customers to the IRS. The United States has information-sharing arrangements with foreign governments on bank accounts that citizens in one country hold in another. The purpose of these arrangements is to share information that will enable tax agencies to identify tax evasion through offshore accounts. These agreements often rest on the principle of reciprocity: foreign governments provide information regarding accounts within their borders in exchange for information from the U.S. government on accounts at U.S. banks. Treasury wants to strengthen reporting requirements for domestic banks to enable the United States to share more information with foreign governments in exchange for information on Americans' foreign bank accounts.⁵²

Tax Expenditure Transparency

In addition to understating their income, taxpayers can underreport their tax liability by claiming deductions or credits for which they do not qualify. Taxpayers must sometimes verify their eligibility for credits by giving the IRS certain information. For example, to claim a child for the Earned Income Tax Credit (EITC), a taxpayer must provide that child's Social Security number. However, other tax credits and deductions do not require the taxpayer to provide any information justifying the claim.

For example, taxpayers who spend the funds they hold in 529 plans do not need to pay federal taxes on gains on those savings provided the funds are spent on qualified education expenses. Yet most taxpayers do not need to substantiate this information during filing season. Taxpayers typically only need to substantiate eligibility if they face an audit.

Requiring information returns that verify whether taxpayers qualify for various tax benefits may discourage filers from claiming benefits erroneously. However, such a requirement would create an additional filing responsibility for taxpayers and additional costs for third parties that report the information, as well as for the IRS. The need for new filings may discourage some taxpayers from claiming tax benefits for which they are eligible.

Simplify the Tax Code

Lawmakers could also attempt to reduce the tax gap by simplifying the tax code. Complexity complicates the filing process and can confuse even professional tax preparers. In 2014, GAO sent employees posing as taxpayers to 19 tax preparers and asked them to prepare their tax returns. Only two of the paid tax preparers correctly identified the fictional taxpayer's refund.⁵³

One way to simplify the code would be to eliminate ambiguities that can either confuse taxpayers and cause errors or give plausible deniability to would-be tax evaders. For example, the IRS expects businesses and workers to consider a number of factors in determining whether a worker is an employee or an independent contractor. While employers withhold and remit estimated income taxes for their employees, they do not do so for independent contractors, making those workers more likely to underreport the income on their tax returns. As a result, ambiguity in the definition of an employee creates opportunities for tax underpayment.⁵⁴

Opportunities for error and evasion can also arise from complexities in the structure of the tax system itself, rather than ambiguities in definitions. For example, the federal government applies payroll taxes to wage income, but not returns to capital.⁵⁵ Business owners must distinguish what income they receive is compensation for their labor and what is profit, yet there may be no objective way to delineate the two. The tax code thus gives the owners an incentive to avoid paying the payroll taxes by attributing more of their income to returns to capital.

Simplifying the tax code can reduce the cost of taxpayer compliance, which could prevent errors. However, simplifying could also eliminate distinctions that Congress considers meaningful, and thus eliminate desired nuance in the tax code. Depending on the details, simplification could either raise revenue, be revenue neutral, or cost the federal government revenue.

Strengthen Regulations on Tax Preparers

Lawmakers could also regulate the competence of paid tax preparers and commercial tax preparation software. GAO reports that nearly 84 million taxpayers, or 43% of all filers, hired paid tax preparers in tax year 2021. While some of these preparers were professionals (e.g., certified public accountants) who had to meet minimum competency requirements, 53% of returns prepared by professional preparers were done by "unenrolled" preparers who were not subject to IRS registration, competency testing, or education requirements. Another 62.5 million taxpayers used commercial software to prepare their own returns.⁵⁶

Currently, most paid preparers are not credentialled and hence are not required to meet any standards or take required trainings. In 2011, the IRS announced it would begin requiring all paid tax preparers to meet minimum competency requirements. However, a group of tax preparers sued, and the U.S. District Court for the District of Columbia held that the IRS had overstepped its legal authority in regulating all paid tax preparers. The District of Columbia Circuit Court of Appeals upheld this decision.⁵⁷ (The IRS still offers a voluntary process for paid tax preparers to obtain an official IRS credential.)

The IRS imposes some regulations on commercial tax software, but the agency does not test the quality of this software routinely.⁵⁸ Congress could give the IRS explicit authority to establish and enforce minimum standards for paid preparers, and the IRS could review the quality of commercial tax software more extensively.

Requiring all preparers to pass minimum competency standards could reduce errors on tax filings, and thus the tax gap, because returns prepared by paid tax preparers have historically had more errors than returns prepared by regulated preparers or IRS-trained volunteers. Since unenrolled preparers are disproportionately likely to serve low-income clients and clients of color, such reforms could prove particularly effective for preventing errors and penalties for these groups. In 2021, 79% of returns that claimed the EITC and were filed by a paid preparer on behalf of a client were completed by unenrolled preparers. Those returns accounted for 90% of audited returns and 92% of dollars adjusted in EITC audits among all returns that were filed by paid preparers and claimed the EITC.⁵⁹

Nearly all returns are filed electronically, often with some sort of commercial software. Hence regulating commercial software more closely could also avoid confusing prompts that cause taxpayer errors.⁶⁰

Some opponents of regulations on tax preparers argue that such rules would prove burdensome for small preparers, giving large preparers an unfair advantage.⁶¹ If true, this effect could limit options and raise prices for consumers. Others have proposed limiting new regulations to currently noncredentialed preparers.⁶²

The Taxpayer Advocate Service, Biden Administration, and Trump Administration each proposed granting the IRS the statutory authority to regulate all paid tax preparers.⁶³ Proposals to tighten regulation of paid preparers have also included increasing penalties for tax preparers who make errors and for taxpayers who hire "ghost preparers," those who fail to identify themselves to the IRS on the tax return.

The IRS is also currently piloting an e-filing service, called "Direct File," that would compete with private e-filers. The program will operate in 13 states during the 2024 filing season. This competitive pressure could represent another means by which the IRS could influence the behavior of private tax preparation services. However, opponents question its likely cost-effectiveness.⁶⁴

Conclusion

The IRS estimates that taxpayers failed to pay 15% of their federal tax liability—an average of $496 billion annually—on time between 2014 and 2016. Underreported income taxes composed 80% of these unpaid federal taxes, particularly among income sources that are not subject to third-party withholding and reporting to the IRS. Underpayment explained another 12% of the total, while nonfiling accounted for the remaining 8%. Late payments and enforcement recouped an average of $68 billion annually, or 2% of the public's tax liability.

These unpaid taxes cost the federal government revenue that it could use to reduce the federal debt, lower taxes, or increase spending. Enforcing the tax code also costs the federal government financial resources, although recent estimates indicate that marginal increases in enforcement spending generate significantly more revenue than they cost.

While the tax gap is the best available estimate of unpaid federal taxes, it is incomplete. Research suggests that the semirandom audits the IRS uses to estimate underpayment fail to identify some underpayment by very high-income filers. The estimate does not include some forms of evasion, such as unpaid taxes on illegal activity, nor does it claim to measure legal tax avoidance. Furthermore, the tax gap relies on some modeling assumptions and older data sources that may bias both the total estimate and the form those unpaid taxes take.

To encourage tax compliance, policymakers could expand existing enforcement activity, such as examinations or math error authority. The IRA included $45.6 billion through FY2031 to bolster tax enforcement. However, Congress has already rescinded some of those funds and agreed to rescind more during the FY2024 and FY2025 appropriations cycles. Other options include requiring third-party information reporting on more sources of income, strengthening regulations on tax preparers, and simplifying the tax code to prevent honest errors.

Appendix. How the IRS Estimates the Tax Gap

The tax gap is the difference between the true tax liability and how much the IRS actually collects. Although the IRS knows how much revenue it collects, it must estimate taxpayers' true tax liability to determine the size of the tax gap. The IRS calculates each component of the tax gap differently. Some calculations rely solely on publicly available data and administrative data, whereas the agency produces others using economic models. The IRS collects new data to estimate the largest portion of the tax gap, underreporting of individual income, by semirandomly selecting returns for audit.⁶⁵

The IRS intends to use some of the funds made available by P.L. 117-169 to use "new data and enhanced analytics" to better understand the tax gap.⁶⁶ As a result, the methods detailed below may change in the coming years.

Individual Income Taxes

Nonfiling

The IRS uses data on household income and administrative data on tax filing to determine what share of households do not file tax returns even though they are required to do so. Specifically, the agency links data from the Census Bureau's American Community Survey—which gives detailed information on household income—to its own tax return data. The IRS calculates each nonfiler's tax liability and subtracts amounts withheld to determine the individual nonfiling tax gap.⁶⁷

Underreporting

Estimating individual income tax underreporting requires approximating filers' "true" tax liability. The agency does so by using the examinations it conducts through its National Research Program (NRP), a semirandom sample of examinations that the IRS conducts principally to calibrate its models that select individual returns for examinations.

The NRP examines roughly 14,000 returns annually. Since this level is approximately one-third of the number of returns the IRS audited through its prior process—known as the Taxpayer Compliance Measurement Program—the IRS pools its data over three-year periods. The sample disproportionately selects filers in groups that often have more complicated taxes, such as self-employed individuals, to help ensure the data on these groups are adequately specific.

NRP auditors may fail to identify every instance of underpayment. To account for any such underpayment, the IRS applies a process called detection-controlled estimation (DCE). DCE compares noncompliance detection rates for different tax issues among various examiners to try to account for the possibility that some auditors, such as those who are more experienced, identify more underpayment than others do. This calculation produces an estimate of the probability of a taxpayer underreporting income for each tax item.⁶⁸ The IRS then uses those probabilities to estimate the dollar value of underreported income tax liability in the population.⁶⁹ Independent researchers found that using DCE roughly tripled the estimated level of underreported income taxes.⁷⁰

This methodology does not identify tip income that recipients fail to report. Because customers often tip workers in cash, tips are subject to less information reporting than other wage and salary income, making them more prone to underreporting. The IRS assumes that tip income has the same noncompliance rate as that of sole proprietor income.⁷¹

Underpayment

Tax filings include each taxpayer's stated tax liability. The IRS can compare this stated liability to the actual amount of revenue it collects from each taxpayer to determine how much stated tax liability taxpayers failed to pay.

Corporate Income Taxes

Nonfiling

The IRS does not report a calculation methodology for corporate nonfiling, and implicitly assumes that corporate nonfiling contributes nothing to the tax gap.

Underreporting

The IRS does not randomly audit corporations to determine their true corporate income tax liability. Instead, the agency estimates corporate tax underpayment based on its normal ("operational") examinations and models of corporate tax underpayment. It does so separately for small corporations (those that report less than $10 million in assets) and larger corporations. The IRS stresses "there is considerable uncertainty surrounding the estimates of this component of the tax gap because of data limitations, lack of information from which to develop a reasonable method to adjust for undetected noncompliance, and other issues."⁷²

The model used to estimate underreporting by small corporations attempts to estimate five different equations jointly, using data on actual operational examinations. The IRS says that it audits less than 1% of small corporations in a given year, and thus uses data from a longer period to get a larger sample. The tax gap estimate for 2014-2016 is based on data from examinations of small corporations from tax years 2009 to 2016.⁷³

The IRS models unpaid large corporate tax liability by assuming that such liability follows a "Pareto distribution," meaning that most recommended additional taxes are concentrated in a small number of firms. The agency says it can make such an assumption if two conditions hold. First, firm size and corporate income must follow a similar distribution. Second, underreporting is roughly as common and significant in firms with low income as high income. Axtell (2001) examined these assumptions using 1997 data, and found that they held. With an estimate of the distribution of noncompliance in place, the IRS uses the results of operational audits to estimate the volume of noncompliance across the population of large corporations.⁷⁴

Underpayment

Tax filings include each corporate taxpayer's stated tax liability. The IRS can compare this stated liability to the actual amount of revenue it collects from each corporate filer to determine how much stated tax liability corporations failed to pay.

Employment (Payroll) Taxes

Nonfiling

Generally, the IRS can determine if businesses failed to file appropriate returns for employment taxes (also known as payroll taxes) using the same methodology as it uses for individual income tax nonfiling. The same is true for unpaid self-employment taxes. The IRS considers situations in which businesses withhold employment taxes but fail to remit them to the IRS in a timely manner to be underpayment cases rather than nonfiling cases.⁷⁵

Underreporting

The IRS uses data from an employment tax study to estimate the underpayment of employment taxes (also known as payroll taxes) for most workers. This study was officially a component of the NRP and used the same semirandom design. Since the study was last conducted from 2008 to 2010, the IRS adjusts the results using current population estimates. The IRS also applies its estimate of underreported tip income from the calculation of individual income tax underreporting to any employment taxes that employers and employees would owe on tip income. Due to a lack of data, the figure does not include underpayment of employment taxes to household and agricultural workers.⁷⁶

While employers and employees typically pay equal amounts of employment taxes on a given worker's wages, self-employed workers pay their entire self-employment tax themselves. As a result, the IRS can use the same methodology to estimate underreporting of self-employment tax liability as it does for individual income tax underreporting.⁷⁷

Underpayment

Tax filings include each taxpayer's stated tax liability. The IRS can compare this stated liability to the actual amount of revenue it collects from each taxpayer to determine how much stated tax liability taxpayers failed to pay.

Estate Taxes

Nonfiling

The IRS does not report a calculation methodology for estate tax nonfiling, and implicitly assumes that estate tax nonfiling contributes nothing to the tax gap.

Underreporting

The IRS says that it estimates the value of underreported estate taxes using "'risk-based' operational examination data and an econometric approach" that is "similar to the methodology used for the small corporation income tax gap estimates."⁷⁸

Underpayment

Tax filings include each taxpayer's stated tax liability. The IRS can compare this liability to the actual amount of revenue it collects from each estate to determine how much tax liability these estates failed to pay.