At a Glance
The Internal Revenue Service (IRS) promotes compliance with tax laws in many ways: It verifies information provided by taxpayers, works to prevent the distribution of refunds that are claimed erroneously, audits tax returns, and attempts to collect unpaid taxes and unfiled returns, among other efforts. In this report, the Congressional Budget Office examines how those activities declined over the 2010–2018 period as the service’s resources decreased and how changes to the IRS’s budget could affect federal revenues.
- In its most recent report on uncollected taxes, the IRS estimated that an average of $441 billion (16 percent) of the taxes owed annually between 2011 and 2013 was not paid in accordance with the law. Most of the unpaid taxes were the result of taxpayers’ underreporting their income. Through enforcement, the IRS collected an average of $60 billion of those unpaid taxes annually, reducing the gap between taxes owed and taxes paid in those years to $381 billion per year, on average.
- The IRS’s appropriations have fallen by 20 percent in inflation-adjusted dollars since 2010, resulting in the elimination of 22 percent of its staff. The amount of funding and staff allocated to enforcement activities has declined by about 30 percent since 2010.
- Since 2010, the IRS has done less to enforce tax laws. Between 2010 and 2018, the share of individual income tax returns it examined fell by 46 percent, and the share of corporate income tax returns it examined fell by 37 percent. The disruptions stemming from the 2020 coronavirus pandemic will further reduce the ability of the IRS to enforce tax laws.
- CBO estimates that increasing the IRS’s funding for examinations and collections by $20 billion over 10 years would increase revenues by $61 billion and that increasing such funding by $40 billion over 10 years would increase revenues by $103 billion.
- CBO’s estimates are subject to considerable uncertainty and only capture the direct effect of enforcement activities. The IRS’s enforcement activities have an indirect effect on the tax gap by discouraging taxpayers from making misstatements on their returns.
The Internal Revenue Service Data Book, 2019 was released on June 30, 2020. Because the date of that release did not allow enough time for the Congressional Budget Office to incorporate those data in its analysis, this report is based on data from the Internal Revenue Service Data Book, 2018 and earlier years.
Unless this report indicates otherwise, all years referred to are federal fiscal years, which run from October 1 through September 30 and are designated by the calendar year in which they end.
Numbers in the text, tables, and figures may not add up to totals because of rounding.
To remove the effects of inflation, CBO adjusted discretionary funding related to federal personnel with the employment cost index for wages and salaries and expressed those amounts in 2018 dollars; other types of discretionary funding were adjusted with the gross domestic product price index and expressed in 2018 dollars. Dollar amounts other than discretionary funding were adjusted with the personal consumption expenditure index and expressed in 2018 dollars. Estimates of the tax gap and the estimated revenue effects of changes to funding for the Internal Revenue Service are in nominal (current) dollars.
The Internal Revenue Service (IRS) collected $3.5 trillion in taxes in 2018, nearly 95 percent of total federal revenues. To do so, it relied largely on taxpayers to report their income, calculate the amount of tax they owed, and remit that amount to the IRS through withholding or other payments. However, some taxpayers have failed to pay hundreds of billions of dollars in taxes, the IRS estimates. Policymakers have expressed interest in how changes in IRS funding, particularly for enforcement of tax laws, could increase the federal government’s tax revenues.
This report describes how the IRS encourages and enforces compliance with tax laws. It examines the IRS’s enforcement activities between 2010 and 2018 and analyzes how the decline in those activities reflects the decline in its funding and staff over that period. On the basis of the relationship between enforcement funding and revenues, the Congressional Budget Office estimates the effects an increase in IRS funding for enforcement could have on federal tax receipts.
The difference between the amount of taxes owed and the amount collected each year—often called the tax gap—is estimated periodically by the IRS. The gross tax gap is the amount that taxpayers do not pay by their filing deadline. As such, it measures the extent of noncompliance with the tax code. In its most recent report, the IRS estimated that the annual gross tax gap was $441 billion, on average, between 2011 and 2013.
The IRS ultimately collects some of that amount. The net tax gap, which is the gross tax gap reduced by the amount that the IRS collects through its enforcement activities, was an estimated $381 billion annually over that period. In addition, the IRS’s enforcement activities have an indirect effect on the tax gap by discouraging taxpayers from making misstatements on their returns. The size of the tax gap is also affected by whether income is visible to the IRS and by the complexity of the tax code, among other factors.
The IRS undertakes a variety of enforcement activities:
- Auditing tax returns,
- Collecting unpaid taxes,
- Obtaining tax returns from taxpayers who did not file returns on time,
- Correcting mathematical or clerical errors,
- Using software to flag questionable refunds, and
- Verifying information reported by taxpayers against information from third parties.
Appropriations for the IRS have fallen by a total of about 20 percent in real (inflation-adjusted) dollars between 2010 and 2018. With the exception of 2016, real appropriations have consistently fallen below the previous year’s level over that period. Because labor costs account for about 70 percent of the IRS’s budget, measures to reduce its workforce were instituted, including a hiring freeze. Those measures resulted in a 22 percent decline in the number of employees at the agency and a 30 percent decline in the number of employees working in enforcement roles. The number of revenue agents and revenue officers, highly specialized enforcement employees who handle the most complex examinations and collections cases, fell by 35 percent and 48 percent, respectively, between 2010 and 2018.
As the IRS’s budget and workforce declined, so did its examination rates for both individual and corporate income tax returns. (The examination rate is the number of examinations closed in a fiscal year divided by the number of returns filed in the previous calendar year.) The overall examination rate for all returns fell by about 40 percent between 2010 and 2018. Over that period:
- The examination rate for individual income tax returns dropped by about 46 percent. About 0.6 percent of individual income tax returns were examined in 2018.
- The examination rate for corporate income tax returns fell by about 37 percent. In 2018, 0.9 percent of corporate income tax returns were examined.
- Larger corporations were more likely to have their returns examined than smaller ones over the 2010–2018 period. However, the examination rates for large corporations—those with assets of more than $10 million—declined more steeply between 2010 and 2018 than examination rates for corporations with fewer assets did.
- Similarly, higher-income individuals were more likely to be examined than lower-income ones over the period. However, the examination rate for higher-income taxpayers fell, while the examination rate for lower-income taxpayers remained fairly stable. Nearly all examinations of lower-income taxpayers were initiated because of claims for the earned income tax credit.
- The amount of additional taxes and penalties the IRS recommended after examinations of corporate and individual income tax returns—before taxpayers appeal or challenge those recommendations—also fell from 2010 to 2018. The decline occurred because the IRS closed fewer examinations each year.
The amount of delinquent tax debt, or unpaid assessments, increased from 2010 to 2018. The amount of revenue received from that debt as a result of collections activities, however, remained between 8 percent and 10 percent of unpaid assessments over the period. The number of delinquent taxpayer accounts, resulting from returns filed without payment of all taxes due or examination assessments not paid promptly, generally increased from 2010 to 2018. The IRS is also responsible for securing returns that were not filed on time. The number of investigations of delinquent filers fell over the 2010–2018 period.
Trends are unlikely to reverse in the near future. The disruptions stemming from the 2020 coronavirus pandemic will reduce the IRS’s enforcement activities and pose new challenges for taxpayers in complying with tax laws.
On the basis of its analysis of the effects that different funding levels have had on IRS enforcement, CBO estimates that increasing the IRS’s funding for examinations and collections by $20 billion over 10 years would boost revenues by $61 billion, resulting in a $41 billion decrease in the cumulative deficit; increasing such funding by $40 billion over 10 years would boost revenues by $103 billion, resulting in a $63 billion decrease in the deficit.
CBO’s estimates for those two options are uncertain and only capture the direct effect of enforcement activities. Any indirect benefits of increasing enforcement, such as deterring taxpayers from violating tax laws, are excluded from the estimates.
Because of the budget scorekeeping guidelines used by the Congress, only the spending increases attributable to those options would be counted in a cost estimate. However, if an appropriation bill or another bill providing funding for one of the options were enacted, CBO’s next projection of the budget deficit would incorporate the estimated effects of the funding increase on tax revenues.
Although the IRS collected, on average, about $2.3 trillion in taxes annually between tax years 2011 and 2013, in its most recent report on the tax gap it estimated that taxpayers actually owed hundreds of billions of dollars more in those years. The IRS periodically estimates the tax gap using a variety of methods, including data from its examinations of tax returns, statistical models, and comparisons of the returns filed by taxpayers to external data or actual receipts.
The IRS estimated that the gross tax gap—the difference between the total amount of federal taxes that taxpayers owed and the amount that was paid on or before the filing deadline—averaged $441 billion annually from 2011 to 2013 (see Figure 1). That figure was 2.7 percent of the nation’s gross domestic product and about 16 percent of federal taxes owed. The gap occurred because taxpayers underreported their liability (for example, by not reporting all sources of income or overstating deductions from income), failed to include full payment with their return, or failed to file a required return in any of five categories: individual income, corporate income, employment, estate and gift, and excise.
Whereas the gross tax gap measures the extent of noncompliance, the net tax gap reflects the ability of the IRS to enforce the law. The net tax gap—defined as the amount of taxes that remains unpaid after the IRS has sought through administrative or enforcement actions to collect taxes owed—averaged $381 billion annually between 2011 and 2013.
The IRS analyzes the ways that taxpayers avoid paying the full amount of taxes they owe so that the agency can determine where to direct its resources for enforcement. About 80 percent ($352 billion) of the annual gross tax gap over the 2011–2013 period occurred because taxpayers underreported their liability (see Figure 2). They either understated their income or overstated tax credits, tax deductions, or income adjustments. The remainder of the gross tax gap was attributable to taxpayers who filed a return but did not pay their taxes in full ($50 billion) and to taxpayers who failed to file a required return ($39 billion).
The largest amount of underreported liability ($245 billion) occurred because individual taxpayers underreported their income (see Figure 3). Most of that underreported income was the result of underreported business income ($110 billion), underreported nonbusiness income ($57 billion), or overstated credits ($42 billion).
The IRS reduced the tax gap from 2011 to 2013 by $60 billion a year, on average, through administrative and enforcement actions. Of that $60 billion, the tax agency collected an average of $43 billion (72 percent) in individual income taxes and $10 billion (17 percent) in corporate income taxes.
The IRS’s ability to reduce the net tax gap through enforcement and other activities depends partly on its budget and staff, but other factors also have an effect: whether taxpayers’ income is reported to the IRS by a third party, whether tax is withheld from payments to the taxpayers, the complexity of the tax code, and the availability of resources that can facilitate compliance, such as IRS publications or paid preparers with expertise.
Visibility of Income and Withholding. Some organizations inform both taxpayers and the IRS of payments to taxpayers, making that income visible to the IRS. For example, employers report wages and salaries on IRS Form W-2. Such third-party information reporting can increase voluntary compliance by minimizing taxpayers’ recordkeeping burden and by making them less likely to underreport earnings. Reporting also allows the IRS to more easily verify amounts reported on a return. Items that are subject to substantial information reporting tend to be accurately reported on income tax returns.
In contrast, items that are subject to little or no third-party information reporting account for most of the underreported income (see Figure 4). For example, although the IRS receives information on some businesses’ gross receipts, it does not receive independent information on expenses. Noncompliant taxpayers can, therefore, inflate their expenses to minimize their net profit from a business.
In recent years, the scope of third-party information reporting has expanded. Payment settlement entities, such as banks and other processors of credit card transactions, are required to report certain payments to individuals on IRS Form 1099-K. When certain assets are sold, brokers and investment managers must include information on the original cost of the assets on IRS Form 1099-B, thus showing the amount of a transaction that is taxed as a capital gain. Some foreign financial institutions are also required to report information about accounts held by U.S. citizens or entities with substantial U.S. ownership.
Income on which taxes are withheld and that third parties report to the IRS, such as wages, accounts for a very small portion of the tax gap. Withholding narrows the tax gap because it allows for the collection of tax as liability accrues. A shift in income away from wages to payments to independent contractors in the so-called gig economy could increase the tax gap because taxes are not withheld on money paid to contractors (who are expected to remit quarterly estimated tax payments), and only certain payments are reported on Form 1099-K or on IRS Form 1099-MISC.
Complexity of the Tax Code. The complexity of the tax code makes compliance more challenging and increases areas of potential dispute with the IRS. Eligibility requirements for certain tax benefits can be confusing and make it more difficult for taxpayers to determine their tax liability. Such eligibility may also be difficult for the IRS to verify.
Another area of complexity is the varied treatment of different forms of income and expenses. For example, corporations can deduct compensation paid to employees as a business expense, but not dividends paid to employees. The difference between the corporation’s interpretation and the IRS’s interpretation of whether a payment is compensation or a dividend can lead to lengthy disputes that are costly for the taxpayer and the IRS.
Resources to Facilitate Compliance. Taxpayers can make mistakes filing a return or calculating liability because they do not know enough about filling out tax returns. To help taxpayers comply with tax laws, the IRS provides information through its website, in its publications, over the phone, and in local Taxpayer Assistance Centers. That assistance can narrow the gross tax gap by helping taxpayers understand how to correctly report their income, credits, deductions, and exemptions.
Some taxpayers seek the aid of paid tax preparers to help them understand the law and how it applies to them. About half of individual income tax returns are filed by a paid preparer. The effect on compliance depends on the skill and motivation of the preparers, who have very different levels of training and expertise. Some 55 percent of all paid preparers are not regulated by the IRS or by any professional licensing board. Regardless of skill, the motivation of preparers may affect whether their assistance improves compliance. For example, tax preparers who offer refund-anticipation products may have an incentive to claim a larger refund than the taxpayer is actually eligible for.
The IRS relies on a variety of approaches to reduce the tax gap. The IRS’s most significant enforcement activities are examinations, collection of unpaid taxes and unfiled returns, automated screening of returns as they are filed, and comparison of information on returns with third-party information (such as that provided on Form W-2). Together, those activities account for the vast majority of revenues the IRS collects after returns have been filed. Enforcement of tax laws also has an indirect effect on revenues by promoting greater voluntary compliance; however, that effect is difficult to measure.
The IRS has separate processes for determining the amount of taxes owed and for collecting unpaid taxes or unfiled returns (see Figure 5). The agency’s enforcement activities are described briefly here; more detail is provided in Appendix A. IRS enforcement relies not only on multiple approaches but also on the work of employees with various levels of skills.
As returns come in, the IRS screens them by using the Return Review Program, a software program that looks for indications of fraud or identity theft, and the Dependent Database, another software program that flags indications of identity theft related to the earned income tax credit (EITC) and other refundable credits. Returns that pass those filters are then screened for mathematical or clerical errors that the IRS can automatically correct.
The IRS assesses the amount of taxes due after it has processed the return and contacts the taxpayer to request additional payment or return any overpayment with a refund. If a return is then referred for examination, the assessed amount may change.
The IRS compares information on returns with third-party information on income that it receives from employers and payment processors. If there is a discrepancy, the IRS may contact the taxpayer to resolve it by mail through the Automated Underreporter (AUR) program, which can assess additional tax without a formal examination.
The IRS audits some taxpayers to determine whether they accurately reported income, deductions, and credits on their return. Examiners use various criteria to determine which taxpayers to audit. Typically, the IRS has up to three years from a return’s due date to examine it.
The IRS seeks payment of delinquent tax debts and the filing of required returns. After returns and payments have been processed, collections personnel determine whether additional taxes are due and notify taxpayers. Once an examination has concluded and taxes, penalties, and fees have been assessed, collections personnel are responsible for ensuring the amount owed is paid. The IRS also identifies taxpayers who were required to file a return but did not and follows up with those taxpayers. The IRS has the authority to place liens on a taxpayer’s property or seize their property to satisfy a tax debt (including garnishing wages from employers). Typically, the IRS has up to 10 years from the date taxes are assessed to collect those taxes.
Although the IRS has increased its use of automated tools, most enforcement activity relies on employees. Examinations and collection of unpaid assessments and unfiled returns require a large number of skilled employees. Three main types of employees are involved in examinations. Tax examiners conduct correspondence examinations for individuals and small businesses. They are trained by the agency to examine a limited range of tax topics. Tax compliance officers conduct limited-complexity, in-person examinations at IRS offices. Revenue agents conduct extensive, in-person field examinations at a taxpayer’s home or place of business.
The two main types of collections employees are contact representatives, who handle taxpayer queries about automatically generated notices from the Automated Collection System, and revenue officers, who contact taxpayers that have not responded to notices. Contact representatives are housed in several call centers within IRS campuses; they handle cases nationwide. Revenue officers, who meet taxpayers face to face, are located throughout the country.
Although the IRS does not measure how much its enforcement of tax laws deters taxpayers from violating those laws, the amount of voluntary compliance probably reflects the level of enforcement activity. In a common model of compliance with tax laws, taxpayers incorporate the risk of being caught violating the laws and the severity of the punishment they might receive into their decisions. Taxpayers’ perceptions of the risk of being caught can affect their decisions even if they themselves are not examined. (For more on how enforcement activities can deter tax noncompliance, see Box 1.)
The enforcement activities of the Internal Revenue Service (IRS) affect revenues directly, by collecting unpaid taxes, and indirectly, by influencing taxpayers’ behavior. Indirect effects are difficult to observe, and their magnitude is highly uncertain. They can be specific, influencing individuals who have been audited to change their behavior, or general, causing even taxpayers who were not audited to be more careful on their returns.
Audited taxpayers may change their behavior in positive or negative ways. They may become better informed about how to report their income and calculate their tax liability, thus increasing compliance in the future. Or they may use the opportunity to learn what the IRS is able to detect and what is permissible, which can reduce their compliance in the future. In addition, audited taxpayers may expect their audit risk to be lower in the near future, further reducing their compliance with tax laws.
Some researchers have found that for several years following an individual income tax audit, people tended to increase the amount of taxable wage and self-employment income they report on their tax returns. The effects were largest for those who were assessed additional tax after the audit, and the longevity of the effect differed by income source. The researchers found a small but sustained positive effect on reported wage income over the six years following an audit. The positive effect on reported self-employment income was larger but quickly diminished. In contrast, those researchers found, corporate taxpayers tended to increase their tax aggressiveness and reduce their reported tax liability as a share of income immediately following an audit, probably because they perceived a lower audit risk in the near future.
Taxpayers’ responses may also differ based on their perceptions of an audit. Among claimants of the earned income tax credit (EITC), audited taxpayers were less likely to claim the EITC or file taxes for a refund in subsequent years than were similar taxpayers who were not audited, even though only a small share of audited taxpayers were determined to be ineligible for the EITC. (Most audits in the analysis sample resulted in a disallowed EITC because of undeliverable mail, taxpayer nonresponse, or insufficient documentation from the taxpayer.)
Other researchers have found that higher-income taxpayers lowered their reported income and tax liability after being notified that they would face an audit, perhaps because they viewed the eventual audit as a negotiation. (Lower-income taxpayers tended to increase their reported income after being notified of an audit.)
Taxpayers may be more likely to comply with tax laws if they perceive a higher risk of being caught, even if they are not audited themselves. Among corporate taxpayers, an increase in the overall examination rate increased all taxpayers’ reported effective tax rate. Researchers have analyzed data from an experiment in which randomly selected firms with a high risk of noncompliance were contacted by the IRS. They found that although IRS contact increased the amount of employment tax remittances paid by other businesses with the same tax preparer, it also decreased remittances by subsidiaries of the contacted firm. In that analysis, on net, the indirect effects of such contact on the people who shared a tax preparer, ownership link, or geographic area with the contacted taxpayer were close to zero.
1. For an overview of recent studies on tax compliance, see Joel Slemrod, Tax Compliance and Enforcement, Working Paper 24799 (National Bureau of Economic Research, July 2018), p. 924, .
2. See Jason DeBacker and others, “Once Bitten, Twice Shy? The Lasting Impact of Enforcement on Tax Compliance,” The Journal of Law and Economics, vol. 61, no. 1 (February 2018), pp. 1–35, .
3. See Jason DeBacker and others, “Legal Enforcement and Corporate Behavior: An Analysis of Tax Aggressiveness After an Audit,” The Journal of Law and Economics, vol. 58, no. 2 (May 2015), pp. 291–324, .
4. See John Guyton and others, The Effects of EITC Correspondence Audits on Low-Income Earners, Working Paper 24465 (National Bureau of Economic Research, May 2019), .
5. Joel Slemrod, Marsha Blumenthal, and Charles Christian, “Taxpayer Response to an Increased Probability of Audit: Evidence From a Controlled Experiment in Minnesota,” Journal of Public Economics, vol. 79, no. 3 (March 2001), pp. 445–483, .
6. See Jeffrey L. Hoopes, Devan Mescall, and Jeffrey A. Pittman, “Do IRS Audits Deter Corporate Tax Avoidance?” The Accounting Review, vol. 87, no. 5 (2012), pp. 1603–1639, ; and Jason DeBacker and others, “Legal Enforcement and Corporate Behavior: An Analysis of Tax Aggressiveness After an Audit,” The Journal of Law and Economics, vol. 58, no. 2 (May 2015), pp. 291–324, .
7. See William C. Boning and others, Heard it Through the Grapevine: Direct and Network Effects of a Tax Enforcement Field Experiment, Working Paper 24305 (National Bureau of Economic Research, February 2018), .
Nearly all of the IRS’s funds are appropriated by the Congress. Appropriations for the IRS fell by about 20 percent (adjusted for inflation) between 2010 and 2019. About 70 percent of the IRS’s overall budget is for labor. The drop in funding thus resulted in a decline in the number of IRS employees over that period, particularly in enforcement.
In 2019, the Congress appropriated $11.3 billion (in current dollars) to the IRS, down from $11.4 billion in 2018, largely allocated among four accounts:
- Enforcement. The largest account, at 41 percent of the IRS’s appropriations in 2019, Enforcement funds the examination, collection, criminal investigation, and appeals activities. Enforcement activities rely on funds from both the Enforcement and Operations Support accounts.
- Operations Support. Agencywide expenses for office space, information technology maintenance and security, research, and strategic planning are funded by the Operations Support account, which received 35 percent of the IRS’s appropriations in 2019.
- Taxpayer Services. The Taxpayer Services account funds the infrastructure necessary for processing returns and refunds, as well as assistance and education for taxpayers as they prepare to file. It was allocated 23 percent of the IRS’s appropriations in 2019.
- Business Systems Modernization. This account funds upgrades to the agency’s taxpayer account and e-filing technology systems. Appropriations to this account in 2019 were 1 percent of the agency’s total appropriations.
Appropriations to the IRS over the past 10 years peaked in 2010, measured in both nominal (current) and real 2018 dollars (see Figure 6). Between 2010 and 2018, the agency’s appropriations decreased by 20 percent, measured in real dollars. The Enforcement account absorbed much of that decline in funding—a 29 percent drop in real resources during that period.
Moving appropriated funds between the IRS’s four accounts requires Congressional approval, though the agency has the flexibility to direct user fees and reimbursements for providing services to other agencies to any account. User fees and reimbursements added less than $0.5 billion to the IRS’s budget in 2018.
Because labor costs account for much of the IRS’s budget, the IRS reduced its staff by instituting a hiring freeze in 2011 and offering buyouts for early retirement in 2012. The result was a decline of 22 percent in the number of employees from its 2010 peak to 2018, mostly from attrition.
Employees funded by the Enforcement account absorbed much of the decline in IRS personnel: a 31 percent reduction in employees between 2010 and 2018. Employees who work the most complex examination and collections cases experienced especially large declines. Between 2010 and 2018, the number of revenue agents, who handle complex enforcement cases, fell by 35 percent, and the number of revenue officers, who manage difficult collections cases, dropped by 48 percent (see Figure 7).
The loss of 15,000 enforcement employees between 2010 and 2018 led to a significant reduction in the number of examinations and the number of follow-ups on discrepancies between returns and third-party data, as well as an increase in assessments that were not collected and unfiled returns that were not secured. Over that period, the number of examinations dropped by about 40 percent even as the number of returns filed grew by 5 percent.
Income tax returns filed by individuals and corporations account for the bulk of recommended additional tax from examination. The decline in examination rates for income tax returns over the 2010–2018 period led to a drop in the total amount of additional tax the IRS recommended following examinations. Taxpayers’ appeals of recommended additional taxes, which can affect the amount of revenues from examinations, remained a constant share of examinations.
The examination rate for individual income tax returns declined from 1.1 percent to 0.6 percent from 2010 to 2018 (see Figure 8). Most examinations of individual returns are conducted through correspondence, and the share of examinations handled through correspondence did not change significantly as the total number of examinations declined.
The percentage decline in the examination rate was larger for higher-income returns. For returns with more than $1 million in total income (before losses were deducted), the examination rate dropped from 8 percent in 2010 to 3 percent in 2018, a 63 percent decline (see Figure 9). The examination rate for returns with total positive income of less than $200,000, accounting for over 95 percent of individual returns each year, dropped to 0.6 percent in 2018 from 1.0 percent in 2010, a 45 percent decline.
Over one-third of all individual returns selected for examination in 2018 were chosen because they included an EITC claim. A former IRS commissioner noted that tax law related to the EITC is so complex that “even people trying to complete the returns accurately and their preparers have trouble figuring out who gets credit.” That complexity, combined with the focus during the past decade on reducing improper payments by government programs, has kept examination rates high for returns with EITC claims. (See Box 2 for more information.) The examination rate for returns with EITC claims is higher than for other returns that report less than $200,000 in total income, though the difference between the examination rates for those groups declined to 1.0 percentage point in 2018, from 1.7 percentage points in 2010 (see Figure 10).
The earned income tax credit (EITC) for low-income taxpayers is a refundable tax credit—that is, if the amount of the credit exceeds a filer’s tax liability, the taxpayer receives the excess amount. In 2018, claims for the EITC amounted to $73.6 billion. The government paid nearly 80 percent of that total, $58.6 billion, to individuals whose returns showed that their tax liability was less than the amount of the credit.
Some of those EITC payments were improper because they were made to taxpayers who were ineligible for the credit or because the government paid the wrong amount to eligible recipients. Improper payments, which cost the government an estimated $140 billion in 2017, have been a focus of legislation and executive action for over a decade. The Improper Payments Information Act of 2002, as amended by the Improper Payments Elimination and Recovery Act of 2010 and the Improper Payments Elimination and Recovery Improvement Act of 2012, increased federal agencies’ requirements to report improper payments. The act also required the director of the Office of Management and Budget to work with agencies to target the small subset of programs, including the EITC, that account for the majority of improper payments. Those high-priority programs must report additional information on the improper payments they issued and establish annual goals for reducing such payments.
To comply with reporting requirements, the Internal Revenue Service (IRS) uses the results of random audits from the National Research Program (NRP) to estimate the percentage of improper EITC claims. In 2018, the IRS estimated that 25 percent ($18.4 billion) of the $73.6 billion in EITC claims was improper. It recovered $1.2 billion of those improper payments through post-refund enforcement activity. The NRP sample revealed no instance of an underpayment to taxpayers who claimed the credit. However, many eligible taxpayers fail to claim the EITC, and nonpayments to such taxpayers are not incorporated in calculations of improper payments.
The high rate of improper EITC claims has several causes. The credit’s eligibility requirements are complex, and the IRS lacks third-party data to authenticate much of what taxpayers report to support their claim (for example, a child’s residence throughout the year or a taxpayer’s marital status.) That lack of data limits the IRS’s ability to verify eligibility without conducting an audit, and some taxpayers may not be able to provide documentation to prove their eligibility. The population that is eligible to claim the EITC undergoes significant turnover each year because wages and family circumstances change, so sending potential claimants notices to encourage compliance is difficult. Finally, taxpayers who claim the EITC are more likely than other filers to use paid return preparers who are not subject to the education requirements or qualifying examinations of tax professionals.
1. See Department of the Treasury, Agency Financial Report: Fiscal Year 2018, p. 194, (PDF, 9.8 MB).
2. See Department of the Treasury, “Payment where Earned Income Credit Exceeds Liability for Tax,” Budget for Fiscal Year 2020 Appendix, p. 953, (PDF, 572 KB).
44. Consistent with previous estimates, the Treasury uses EITC claims—including those that reduce the amount of tax paid—to estimate improper payments (though only the government’s outlays meet the definition of an improper payment).
5. If the $1.2 billion in recovered improper EITC payments was subtracted before calculating the improper payment rate, that rate would be 23.4 percent rather than the reported 25.1 percent. (Those estimates are based on NRP data; revenue recovered from operational audits involving the EITC is excluded.) See Department of the Treasury, Agency Financial Report: Fiscal Year 2018, footnote 4, p. 194, (PDF, 9.8 MB).
6. See Department of the Treasury, “Barriers,” Agency Financial Report: Fiscal Year 2018, Section III, p. 200, (PDF, 9.8 MB).
7. See Internal Revenue Service, Compliance Estimates for the Earned Income Tax Credit Claimed on 2006–2008 Returns, Publication 5162 (August 2014), Table 8, (PDF, 972 KB).
The examination rate for corporate income tax returns dropped to 0.9 percent in 2018 from 1.4 percent in 2010 (see Figure 11). Because of their complexity, most corporate income tax examinations are conducted in the field. Although the total number of corporate income tax examinations has declined, the share of examinations conducted in the field—that is, at the taxpayer’s home or workplace—has been roughly constant.
The rate of examination for corporations that reported assets of $20 billion or more declined to about 50 percent in 2018, down from almost 100 percent in 2010 (see Figure 12). For corporations with assets of less than $10 million, the examination rate fell to 0.6 percent from 1.1 percent.
Examinations generate enforcement revenue by proposing adjustments to a return and recommending additional tax and penalties, though a small number of cases result in a refund for the taxpayer. After a period when a taxpayer may challenge or appeal the audit findings, that additional tax is no longer recommended but required and becomes a tax assessment. As examinations declined, the total amount of additional tax (excluding penalties) recommended for individual and corporate income tax returns fell by 50 percent, from $46 billion in 2010 to $23 billion in 2018.
The amount of additional tax recommended after examinations of individual income tax returns fell steadily over the 2010–2018 period (see Figure 13). The average amount of additional tax that the IRS recommended did not change significantly, but the number of examinations declined. The decline was largest for taxpayers with income of $1 million or more (before losses, exclusions from income, or adjustments to income). The amount of additional tax recommended for that group fell to $1.9 billion in 2018 from $5.7 billion in 2010.
For corporations, the amount of additional tax recommended after examinations declined more steeply and was more volatile. That result was driven by examinations of corporations with more than $20 billion in assets because the closure of a few big cases can account for a large percentage of the total amount of additional tax recommended. In 2017, the closure of a small number of large corporate cases accounted for more than $2 billion in examination assessments.
Taxpayers have the right to challenge a recommendation of additional tax and penalties, either through an administrative appeal within the IRS or through a judicial appeal in the U.S. Tax Court, a U.S. District Court, or the U.S. Court of Federal Claims. A verdict can reduce or eliminate the additional tax. If the additional tax is reduced, any penalties that were calculated based on the amount of that tax are reduced as well. Courts may also reduce penalties independently, without changing the amount of additional tax, if taxpayers can show reasonable cause for not complying with tax law.
The overall number of appeals declined from 2010 to 2018, but the rate at which individual and corporate taxpayers challenged the results of their examinations did not change significantly over the period. Many factors may influence a taxpayer’s decision to challenge an examination result: the strength of the IRS’s case, the scope of the changes the IRS made to the return, the amount of additional tax the IRS recommended, and the taxpayer’s resources (for example, funds may be needed to hire assistance in the administrative or judicial appeals process). Any changes in these factors over the period did not significantly affect the rate at which taxpayers agreed to examination results.
The IRS does not rely entirely on formal examinations to make adjustments to tax liability. Some processes to check the accuracy of returns are largely automated and thus require fewer employee hours per return than an examination. They include the correction of mathematical and clerical errors and the identification of discrepancies between returns and third-party documents.
The IRS is authorized by law to automatically correct certain mathematical and clerical errors on returns and recalculate liability on the basis of such corrections. That “math error authority” is applied using computer software. The number of errors corrected automatically by the agency each year depends on the number and type of errors the IRS is authorized by the Congress to correct and the number of taxpayers who made them. In 2018, 1.9 percent of individual returns were corrected automatically with math error authority. The IRS’s math error authority was expanded in 2009 to enable it to automatically correct certain errors associated with the temporary Making Work Pay and Recovery Rebate credits. If errors associated with those credits are set aside, the number of math errors identified as a share of individual income tax returns declined from 2010 to 2018. That decline occurred because taxpayers made fewer correctable errors rather than because the IRS’s resources were constrained.
The IRS uses software after it processes returns to identify discrepancies between those returns and data supplied by employers and other third parties. Some discrepancies are selected for review as part of the Automated Underreporter program, which generates notices to taxpayers when a discrepancy is found in a return, proposing changes to the return based on the third-party information. The number of such notices depends on the number of AUR personnel available to review the flagged returns and handle taxpayers’ responses to notices. The number of employees in the AUR program declined by 40 percent, from 2,255 to 1,366, between 2010 and 2018. Improvements in the selection of cases to be reviewed allowed the IRS to increase the productivity of the AUR program’s remaining employees, but case closures nonetheless declined as a share of individual returns by 36 percent over the period (see Figure 14).
In 2018, taxpayers owed the IRS about $511 billion in delinquent tax debt, or unpaid assessments (see Figure 15, top panel). Unpaid assessments arise when taxpayers file a return without paying taxes owed in full, when examinations or automated enforcement activity result in additional tax that is not paid promptly, or when nonfilers are found to owe taxes. They include taxes and accrued penalties and interest for the current year, as well as any amounts owed from previous years that fall within the 10-year statute of limitations on collecting taxes. Only a portion of unpaid assessments is collectible; in some cases, the taxpayer cannot be located, is deceased (or, in the case of a business, defunct), or faces financial hardship.
The economy thus affects collections revenues—the tax, penalties, and interest that are received as a result of collections activity on returns filed with additional tax due or on overdue returns. A stronger economy not only increases tax liabilities but can also result in fewer uncollectible assessments. The combination of those factors suggests that in a growing economy, the IRS would collect a greater share of unpaid assessments. However, although the economy grew stronger between 2010 and 2018, revenues from enforcement activities increased only slightly. Between 2010 and 2018, collections revenue was within 8 percent to 10 percent of the growing amount of unpaid assessments owed to the IRS. (See Figure 15, center and bottom panels).
Most cases in collection are delinquent accounts from filers who either did not pay the tax they owed or paid only a portion of the amount due (see Figure 16). From 2010 to 2018, the IRS typically opened more delinquent account cases in a year than it closed, leading to a growing backlog. In recent years, the agency has transferred some accounts in the backlog to private collection agencies, leading to the closure of more cases.
Another set of cases in collection are investigations of people who did not file a return. The IRS identifies probable nonfilers through third-party information or prior-year filing information and sends them an automated notice about their failure to file. The number of notices sent has declined in recent years because the reduction in the IRS’s resources meant that fewer employees were available to identify nonfilers.
The decline in staff has also meant that there is little follow-up on the notices sent to nonfilers. The IRS assesses tax on nonfilers with an automated process that creates a substitute return with data from third parties (the Automated Substitute for Return, or ASFR). Even though substitute returns are created automatically, employees are necessary to respond to taxpayers who offer reasons for not filing or who want to correct the substitute returns, which tend to overstate taxpayers’ liability. The number of ASFR cases closed declined to 10,000 in 2018 from 1.2 million in 2010, and the program has been largely inactive since 2015, when the IRS assigned most of the ASFR’s staff to other functions. With the reduction in ASFR activity, enforcement activity for many high-income nonfilers has been reduced to a series of notices.
The disruptions stemming from the 2020 coronavirus pandemic will reduce the ability of the IRS to enforce tax laws and will present new challenges for taxpayers in complying with tax laws.
Reduced Enforcement Activities. The IRS announced a pause in many enforcement activities from April 1, 2020, through July 15, 2020. Specifically, the IRS suspended liens and levies, stopped initiating new field or correspondence examinations, and extended deadlines to make payments on installment agreements and submit supporting documentation for EITC claims. In addition, the closure of IRS facilities has interrupted the processing of paper correspondence from taxpayers, including documents related to audit and collections activities.
Increased Demands for Taxpayer Services and Operations. The IRS extended the April 15 deadline to file and pay federal income taxes to July 15, 2020, so the IRS will need to devote additional resources to processing tax returns and refunds over a longer time period. In addition, the agency has been tasked with administering the individual recovery rebates enacted as part of the Coronavirus Aid, Relief, and Economic Security (CARES) Act. The IRS has disbursed those payments quickly—in the two months after enactment of the CARES Act, the IRS paid nearly $267 billion to 159 million individuals. Those rebates and other new tax provisions will create further demands on IRS resources when 2020 tax returns are filed in 2021.
Reduced Taxpayer Assistance. The IRS suspended live telephone assistance and closed the walk-in Taxpayer Assistance Centers, reducing the resources available to help taxpayers comply with the law. Those actions will hinder taxpayers seeking assistance with new issues that have arisen from recent legislative and administrative changes as well as taxpayers whose issues predate the coronavirus pandemic.
Policymakers have expressed interest in how increases in IRS funding, particularly for enforcement activities, would increase tax revenues. (For a discussion of other changes that could increase revenues under the current tax regime, see Box 3.) Estimates of the additional revenue that would result from more spending—such as the estimates in this report—would not be included in a cost estimate because of scorekeeping guidelines used by the Congress. If additional funds were appropriated, however, their effects on both spending and revenues would be incorporated into the Congressional Budget Office’s next budget baseline. The estimates are necessarily uncertain because the link between spending on enforcement and the collection of revenues is not direct, and many factors can affect the IRS’s ability to use added funding to increase revenues.
In addition to changing the amount of appropriations for the Internal Revenue Service (IRS) to increase the amount of traditional enforcement activity, policymakers have other options to affect the amount of revenues the IRS brings in under the current tax regime.
In recent testimony to the Congress, the Government Accountability Office (GAO) described many potential changes that could reduce noncompliance with tax laws. A full analysis of those policies is outside the scope of this report, but in brief, GAO suggests that the IRS could do the following:
- Develop a strategy that uses data from its National Research Program (NRP) to update its compliance programs,
- Establish a quantitative goal for improving voluntary compliance,
- Analyze and use results of employment tax NRP examinations to improve employment tax compliance programs, and
- Make greater use of the automated Return Review Program to reduce fraud.
The tax gap (the difference between taxes owed and taxes paid) could also be reduced if policymakers increased the amount of information available to the IRS or expanded its authority by doing one or more of the following:
- Expanding third-party information reporting to cover more transactions,
- Requiring more taxpayers to electronically file tax and information returns,
- Expanding the IRS’s math error authority to other types of discrepancies on tax returns, and
- Giving the IRS the authority to regulate paid tax preparers.
Former IRS Commissioner Charles Rossotti proposed new reporting requirements for small- and medium-sized businesses and also suggested significantly increasing the agency’s technology spending to allow wider use of data analysis in fraud detection and enforcement activity (for example, enabling the IRS to automatically generate notices with information specific to each taxpayer). Those more ambitious proposals to reduce the tax gap would, in the judgment of the Congressional Budget Office, require additional statutory authority (for example, greater information reporting to the IRS) or more fundamental overhauls of the IRS’s existing information technology, audit, and enforcement functions.
1. See Testimony of James R. McTigue Jr., Director, Strategic Issues, Government Accountability Office, before the House Committee on Ways and Means, Tax Gap: Multiple Strategies Are Needed to Reduce Noncompliance, GAO-19-558T (May 9, 2019), .
2. See Charles Rossotti, “Recover $1.6 Trillion, Modernize Tax Compliance and Assistance,” Tax Notes Federal (March 2, 2020), p. 1411, .
CBO estimated the effect of increasing the IRS’s enforcement budget by $20 billion or $40 billion over a 10-year period, projecting how much of the additional revenue would be received in each year. Those estimates are based on the IRS’s estimates of the average amount of revenue that would be collected for every additional dollar of enforcement. However, CBO has not estimated the deterrent effect of increased enforcement on other taxpayers—and thus its estimates do not show increases in revenues as greater enforcement influences more taxpayers to comply with tax laws.
Estimated Revenue Effects of a $20 Billion Increase. CBO estimates that a $20 billion increase to the IRS’s appropriations for enforcement activities over 10 years would raise revenues by $61 billion over that period. On net, the increase in spending on enforcement would reduce the deficit by $41 billion over the 2021–2030 period (see Table 1).
The IRS’s funding would increase gradually, rising by $500 million each year for the first five years, and then remain at an additional $2.5 billion per year from 2026 to 2030. Each infusion of new funding would result in the start of new enforcement initiatives—expansions of audits and other activities that could improve compliance with the tax system. All of the new initiatives would be funded at the same level throughout the budget period. For example, 2021 initiatives would receive $500 million each year from 2021 to 2030, 2022 initiatives would receive $500 million from 2022 to 2030, and so forth.
CBO estimates that revenues would increase gradually over the 10-year period, reaching roughly $9 billion a year from 2027 through 2030. The agency’s assessment is based on the amount of time it generally takes new hires to be fully effective. In the first year of an initiative, the return in tax revenue per dollar of spending would be low because new employees need to be hired and trained. The return on the initiative would increase as employees finished their training and gained experience, and it would reach its maximum level in the third year of funding.
Although CBO’s estimates start with the IRS’s calculations of the revenue it would collect per dollar of enforcement spending, CBO made two adjustments to better approximate the marginal return on that spending. The first is an adjustment for taxpayers’ learning. After the third year of an initiative, CBO judges that taxpayers will have adapted to a new enforcement activity and developed ways to evade that enforcement. CBO therefore reduced the marginal return on each activity after the third year. The second adjustment incorporates the expectation that the IRS prioritizes enforcement activities that it projects to have the highest average return; therefore, the spending associated with the 2021 initiative would have the greatest return, and initiatives that start in the 2022–2025 period would have progressively lower returns.
The return from a particular initiative reflects the amount of revenue that will be collected from it over the next 10 years. CBO converted that return into a stream of revenue receipts on the basis of information from the IRS. Enforcement initiatives that start after 2021 would bring in some revenue outside the 10-year period; CBO’s estimate does not include that revenue.
Estimated Revenue Effects of a $40 Billion Increase. If the IRS was given twice as much additional funding for enforcement activities as was provided in the first option, CBO estimates that the return per dollar of spending would be less than twice as high, reflecting the expectation that the IRS would focus first on initiatives that generate the most revenue. A $40 billion increase in the IRS’s appropriations would thus boost revenues by $103 billion if directed to enforcement activities over that period. On net, the increase in the enforcement budget would reduce the deficit by $63 billion through 2030.
The pattern of funding for the $40 billion option would follow the previous option, beginning with $1 billion in additional funds the first year, increasing gradually by $1 billion per year for the first four years, and then remaining at $5 billion per year from 2026 to 2030. The adjustment to returns for taxpayer learning would also be the same. But the adjustment for reduced marginal returns would be even larger with each successive initiative because the activities with the highest average return would be undertaken even more quickly than under the first option. Additionally, hiring enough qualified new enforcement employees each year would become more difficult, and training less-qualified employees might involve more time and spending. The amount of added revenues would peak at close to $15 billion in 2028.
The revenue changes attributable to the two options for increasing enforcement would not be counted in formal cost estimates. Under the Congressional scorekeeping guidelines that govern the cost estimates CBO produces, added revenues or reductions in mandatory spending that might result from additional spending are not included. The Congress established those guidelines in large part to avoid crediting uncertain potential savings as an offset against certain upfront spending. The scorekeeping guidelines were included in the conference report for the Balanced Budget Act of 1997, which aimed to ensure consistent treatment of spending authority, appropriations, and outlays over time. The guidelines were last updated in 2015.
Two guidelines are especially relevant to estimates for legislation that involves enforcement of tax laws. Scorekeeping guideline 3 states: “Revenues, entitlements and other mandatory programs (including offsetting receipts) will be scored at current law levels . . . unless Congressional action modifies the authorizing legislation.” Put another way, potential revenues from legislation will be counted in a cost estimate only if those revenues result from changes in the tax code. Even though additional discretionary appropriations for IRS enforcement may produce budgetary savings (from increased federal tax receipts), such savings are not counted in a cost estimate.
Scorekeeping guideline 14 states: “No increase in receipts or decrease in direct spending will be scored as a result of provisions of a law that provides direct spending for administrative or program management activities.” That guideline prohibits budgetary savings from being counted if they result from funding in authorizing legislation for administrative or program management activities, including increased IRS enforcement.
The scorekeeping guidelines do not apply to CBO’s baseline budget projections or to its other projections such as the analysis of the President’s budget. So, although CBO does not include the revenue effects of changes in the IRS’s funding in cost estimates, the agency incorporates both the spending and revenue effects of enacted legislation in its next update of baseline budget projections.
CBO also adjusts baseline projections of revenues from taxes paid on a liability from a prior year, called back taxes, to account for increases or decreases in the real amounts appropriated for enforcement of tax laws. In general, revenues from back taxes are projected on the basis of the historical relationship between those revenues and overall tax liabilities. If the IRS’s resources for enforcement activities in future years are projected to be less than in the past, the expected ratio of revenues from back taxes to overall tax liabilities is lowered. (Similarly, that expected ratio would be increased if IRS resources were projected to increase.)
Four factors contribute to the uncertainty surrounding the revenue effects of increases in IRS funding. First, although the average return per dollar of enforcement is the best method available to measure the effect of increasing revenues, it is imperfect. Second, the productivity of additional funding would depend on taxpayers’ responses to increased enforcement. Third, the effect of additional funding on revenues would depend on the efficiency with which it was used. Fourth, revenues would be affected by how the IRS chose its caseload.
Use of Average Return per Dollar to Measure Marginal Returns. Although CBO adjusts the IRS’s average-return-per-dollar figure to better measure marginal return, CBO’s adjustments may not capture all the ways that changes in the IRS’s funding affect revenues (see Appendix B for details on CBO’s method for estimating changes in revenues). In particular, the average return per dollar does not incorporate the indirect effects of IRS spending, which are difficult to measure. Excluding the indirect effects of enforcement spending may understate the amount of revenue brought in as a result of greater enforcement.
Further complicating the calculation of indirect effects is the fact that enforcement activities with relatively low returns on spending may substantially affect taxpayers’ behavior. For example, the ASFR program for nonfilers has a low return on spending because the substitute returns tend to overstate the amount of taxes owed. IRS employees may need to correct the substitute returns (increasing the cost of the program), and they typically reduce the amount due (lowering the program’s impact on revenues). However, researchers judge that the program has had a large indirect impact by motivating nonfilers who were affected by the program to file returns in subsequent years.
Taxpayers’ Behavior. The returns on additional funding depend on taxpayers’ future behavior, including their responses to changes in tax laws and the nature of their employment. If more income was earned in sectors with less third-party reporting—for example, if more taxpayers were self-employed—more IRS resources would be needed to detect the likely uptick in unreported income.
The ways in which taxpayers interact with the IRS and the processes used by the agency also affect the results of enforcement activities. For example, tax returns that are filed on paper must be manually transcribed into the IRS’s electronic databases. That process is costly, and transcription errors can be introduced. In addition, only some of the information on a paper return is transcribed, which limits the amount of information available for enforcement activities. The continued increase in electronic filing and potential changes in how paper returns are processed (for example, by using optical character recognition technology or scannable barcodes) will affect the productivity of the IRS’s enforcement programs.
The IRS’s reliance on processes such as automated math error correction and partially automated discrepancy review under AUR can also affect revenues if taxpayers do not understand the IRS’s notices and cannot provide adequate responses. In both automated processes, taxpayers’ ability to appeal an error is more limited than it would be in a traditional examination.
Use of Additional Funding. The steep decline in the IRS’s funding after 2010 resulted in substantial staffing reductions. The return on additional funding would depend on the efficiency with which funding can be used—that is, the speed with which the IRS can hire and train new staff and allocate new employees to enforcement activities.
It is uncertain how long it would take for new hires to become productive. The IRS would not be able to bring new hires up to speed instantly, nor is it likely that the agency could hire them quickly. The Government Accountability Office reports that it can take a year or longer from the time an IRS supervisor notifies the division of a staffing need until the employee is on board. Enforcement staff need to develop specialized expertise to become effective, and the IRS estimates that it can take four to five years to train new hires to become experienced senior-level revenue officers. In addition, a growing share of the IRS’s staff members are eligible for retirement, particularly those in the Senior Executive Service, the federal government’s managerial branch.
Some of the IRS’s activities depend on where its employees are located. Correspondence audits and other routine enforcement functions are conducted from campuses around the country. Work that involves face-to-face contact with taxpayers, such as field audits and field collections, is conducted from local IRS offices. For those activities, the effectiveness of additional employees would depend on those employees’ location.
Choice of Cases for Examination. CBO’s revenue estimates incorporate the expectation that the IRS would choose to work the cases with the highest return first. But the IRS might choose a different mix of cases to meet other goals, such as balancing enforcement among different types of taxpayers. For example, the average return on correspondence cases is substantially higher than the average return on field examinations, even though correspondence cases bring in less revenue, on average. That is because field examinations are substantially more costly: They require experienced employees and take more time to complete. However, the IRS might choose to conduct more field examinations, which typically involve corporations or other taxpayers with business income, to ensure that those taxpayers comply with tax laws. Moreover, because different types of tax issues require examiners with different skills and levels of expertise, the IRS’s ability to change the types of cases it pursues is limited by the number of employees that have particular skills.
1. Those figures are measured as a decline in the number of full-time equivalents (FTEs). Because not all employees work full time in a given year, the IRS calculates the number of FTEs as the total number of hours worked divided by the number of hours that a full-time employee would work.
2. For details of how those estimates were created, see Internal Revenue Service, Federal Tax Compliance Research: Tax Gap Estimates for Tax Years 2011–2013, Publication 1415 (September 2019), (1.4 MB). Those estimates are the most recent available.
3. Payments counted in measuring the net tax gap include both those obtained through enforcement actions and those made after the due date without the IRS’s intervention. For example, a taxpayer might make a payment when filing an amended return.
4. Business income includes nonfarm proprietor income; flow-through income from partnerships, S corporations, and estates and trusts that is taxed at the individual level; rent and royalty income; and farm income. Those forms of income are reported on Schedules C, E, and F. Nonbusiness income includes wages and salaries, interest, dividends, pensions and annuities, unemployment compensation, Social Security benefits, and capital gains.
5. For example, see Joel Slemrod and others, “Does Credit-Card Information Reporting Improve Small-Business Tax Compliance?” Journal of Public Economics, vol. 149 (May 2017), pp. 1–19, .
6. Payments reported on Form 1099-K include payments received by the taxpayer from debit, credit, or stored-value cards or from a third-party settlement organization (such as PayPal). To receive a 1099-K from a settlement organization, the taxpayer must receive gross payments in excess of $20,000 and have more than 200 transactions. See Internal Revenue Service, “Understanding Your Form 1099-K” (March 11, 2020),
7. Certain types of income that are not subject to withholding may be subject to backup withholding if the payee’s taxpayer identification number is incorrect or if the taxpayer has previously failed to fully report his or her income on a tax return.
8. Form 1099-MISC is generally used for reporting payments greater than $600 to someone who is not an employee, such as someone who provides a service, or for other purposes. See Internal Revenue Service, “About Form 1099-MISC, Miscellaneous Income” (March 11, 2020), .
9. Complexity can also lead to taxpayers paying more than they owe. The estimated underreporting tax gap does not include such overpayments.
10. Paid preparers who are attorneys, certified public accountants, or enrolled agents must meet education requirements and pass qualifying exams to practice as tax professionals. See Internal Revenue Service, “Understanding Tax Return Preparer Credentials and Qualifications” (January 24, 2020), ; and Government Accountability Office, Paid Tax Return Preparers: In a Limited Study, Preparers Made Significant Errors, GAO-14-467T (April 8, 2014), Figure 2,
11. See Maggie R. Jones, Tax Preparers, Refund Anticipation Products, and EITC Noncompliance, CARRA Working Paper Series, Working Paper 2017-10 (Census Bureau, December 2017), (PDF, 545 KB).
12. Other IRS enforcement activities include investigating criminal violations of internal revenue laws and other financial crimes. Estimates of the tax gap do not include taxes on income derived from illegal activities or certain types of fraud, so criminal investigations are not included in CBO’s analysis of enforcement activities.
13. For a detailed chart of the tax system, see Taxpayer Advocate Service, Publication 5341: Taxpayer Roadmap (December 2019), (402 KB).
14. Tax examiners currently receive approximately 85 hours of training in income tax law when hired and can complete additional training modules. See National Taxpayer Advocate, Tax Law Questions, “Correspondence Examination: The IRS’s Correspondence Examination Procedures Burden Taxpayers and Are Not Effective in Educating the Taxpayer and Promoting Future Voluntary Compliance,” Annual Report to Congress 2018 (February 2019), .
15. For more discussion of how enforcement affects noncompliance, see Joel Slemrod, “Cheating Ourselves: The Economics of Tax Evasion,” Journal of Economic Perspectives, vol. 21, no. 1 (February 2007), pp. 25–48, .
16. The analysis focuses on IRS’s resources and enforcement activities from 2010 to 2018, though information going back to 2008 is also shown in figures to provide context for longer-term trends. IRS resources also declined significantly during the 1990s. For more details, see Alan Plumley and Eugene Steuerle, “Ultimate Objectives for the IRS: Balancing Revenue and Service,” in Henry Aaron and Joel Slemrod, eds., The Crisis in Tax Administration (Brookings Institution Press, May 20, 2004), pp. 311–346, .
17. See Government Accountability Office, Internal Revenue Service: Strategic Human Capital Management Is Needed to Address Serious Risks to IRS’s Mission, GAO-19-176 (March 2019), pp. 17–18, .
18. In CBO’s baseline projections, the agency adjusts discretionary funding related to federal personnel with the employment cost index for wages and salaries, and it adjusts other discretionary funding with the GDP price index. IRS appropriations in this report are adjusted for inflation with a blend of those indexes, weighted to reflect the percentage of each appropriation account that funds personnel.
19. The analysis here is based on the Internal Revenue Service Data Book, so the period ends with the last published year of data at the time of writing (2018). During this period, the IRS’s responsibilities have grown. For more discussion, see Brian Erard and Alan Plumley, “Doing More With Less? Using Data and Analytics to Overcome Shrinking Enforcement Budgets and Expanding Responsibilities” (paper presented at the 2018 Corporate Tax Management Conference on Tax and Technology), .
20. The Internal Revenue Service Data Book lists employees at the level of “budget activity,” a subcategory of appropriation account. The Enforcement account funds three budget activities: examinations and collections, investigations, and regulatory. See Internal Revenue Service, Internal Revenue Service Data Book, 2018 (June 30, 2020), .
21. The examination rate is calculated as the number of examinations closed in a fiscal year divided by the number of returns filed in the previous calendar year. That is because most examination activity takes place in the fiscal year after a return is filed.
22. The percentage of returns in a fiscal year that were examined because of an EITC claim is drawn from the Internal Revenue Service Data Book and divided by the number of returns filed the previous calendar year. See Internal Revenue Service, Internal Revenue Service Data Book, 2018 (June 30, 2020), .
23. See William Hoffman, “A Conversation With Former IRS Commissioner John Koskinen: Tax Administration,” Part 1, Tax Notes Talks (podcast, October 10, 2019), .
24. For the EITC, the amount of credit that is either claimed by an ineligible taxpayer or claimed in the wrong amount by an eligible taxpayer is considered an improper payment. The amount of EITC that is not claimed by eligible taxpayers is not included in that calculation.
25. The examination rate is calculated as the number of examinations closed in a fiscal year divided by the number of returns received in the previous calendar year. As a result, the rate can exceed 100 percent in a given year, if older returns are examined in the current fiscal year or if an examination started in a prior year takes longer than a year to close, as complex corporate examinations typically do. In some situations, the IRS examines multiple years of a corporation’s returns and then closes all of them at one time. In 2012, the examination rate for corporations with assets reported on their balance sheets of $20 billion or more surpassed 100 percent.
26. See Treasury Inspector General for Tax Administration, Trends in Compliance Activities Through Fiscal Year 2017, Reference Number 2017-30-072 (September 13, 2018), p. 6, (PDF, 1 MB).
27. See Internal Revenue Service, Internal Revenue Service Data Book, 2010–2018 (June 30, 2020), .
28. A study examining the outcomes of appeals by large public corporations found that as the IRS’s resources for conducting corporate examinations declined, the IRS may have prioritized examinations of large corporate taxpayers with weaker (more questionable) cases. Despite such prioritization, the study found that the reduction in the IRS’s resources had a negative impact on the amount of tax revenue it received from large corporations. See Michelle Nessa and others, “How Do IRS Resources Affect the Corporate Audit Process?” The Accounting Review, vol. 95, no. 2 (March 2020), pp. 311–338, .
29. Errors associated with the Making Work Pay and Recovery Rebate credits (authorized in the American Recovery and Reinvestment Act of 2009) totaled 10 million on returns filed in calendar year 2009 (7 percent of returns) and 7 million on returns filed in calendar year 2010 (5 percent of returns). Such errors continued to decline in later years.
30. See Treasury Inspector General for Tax Administration, “Automated Underreporter Program Tax Assessments Have Increased Significantly; However, Accuracy-Related Penalties Were Not Always Assessed When Warranted,” Reference Number 2015-30-037 (May 8, 2015), pp. 5–6, .
31. Amounts assessed following examination that the taxpayer does not agree with and amounts in appeals are also included. Treasury estimates that about 14 percent of the unpaid assessments in fiscal year 2016 were from examinations. For more details on unpaid assessments, see Treasury Inspector General for Tax Administration, Trends in Compliance Activities Through Fiscal Year 2016, Reference Number 2017-30-072 (September 11, 2017), (PDF, 3.95 MB).
32. Some tax debt is paid by transferring credits (for example, a refund on income taxes paid) to satisfy a debt from a past year. Collections revenue is the amount of delinquent tax debt collected excluding such transfers plus the amount of payments from investigations of nonfilers. See Internal Revenue Service, Internal Revenue Service Data Book, 2018 (June 30, 2020), . The figure for delinquent tax debt is drawn from the amount of gross accounts receivable calculated by the IRS Office of Research, Analysis, and Statistics and Chief Financial Officer. See Treasury Inspector General for Tax Administration, Trends in Compliance Activities Through Fiscal Year 2018, Reference Number 2019-30-063 (September 9, 2019), Appendix IV, Figure 2, (PDF, 729 KB).
33. See Treasury Inspector General for Tax Administration, A Significantly Reduced Automated Substitute for Return Program Negatively Affected Collection and Filing Compliance, Reference Number 2017-30-078 (September 29, 2017), pp. 9–10, (PDF, 990 KB)
34. See Treasury Inspector General for Tax Administration, High-Income Nonfilers Owing Billions of Dollars Are Not Being Worked by the Internal Revenue Service, Reference Number 2020-30-015 (May 29, 2020), (PDF, 3.9 MB)
35. See Internal Revenue Service, “IRS Unveils New People First Initiative; COVID-19 Effort Temporarily Adjusts, Suspends Key Compliance Program,” IR-2020-59 (news release, March 25, 2020), .
36. The IRS received an additional $250 million in appropriations for fiscal year 2020 in the CARES Act (Public Law 116-136) to facilitate the extension of the filing season and processing of the recovery rebates.
37. See Internal Revenue Service, “159 Million Economic Impact Payments Processed; Low-Income People and Others Who Aren’t Required to File Tax Returns Can Quickly Register for Payment With IRS Non-Filers Tool,” IR-2020-111 (news release, June 3, 2020), .
38. CBO previously estimated the revenue effects of increased appropriations for IRS enforcement. See Congressional Budget Office, Options for Reducing the Deficit: 2019–2028 (December 2018), p. 306, .
39. Other researchers have argued that increasing the audit rates in 2018 to the same levels as in 2011 would have raised $14 billion in revenues in 2018 and that dedicating additional audit resources only to higher-income taxpayers would have resulted in more revenues. See Natasha Sarin and Lawrence H. Summers, “Shrinking the Tax Gap: Approaches and Revenue Potential,” Tax Notes (November 18, 2019),
40. Funding would be directed to the appropriations accounts for enforcement and operating support to cover increases in costs for agencywide infrastructure, such as offices and computer hardware, for new employees.
41. See U.S. House of Representatives, Balanced Budget Act of 1997: Conference Report to Accompany H.R. 2015, House Report 105-217 (July 30, 1997), pp. 1007–1014, (PDF, 3.2 MB).
42. Funding for the IRS is generally provided through appropriation acts, not authorizing legislation.
43. In its baseline, CBO projects that individual discretionary appropriations, including those for enforcement activities, grows with inflation; hence, in real dollars, projected funding in all future years is equal to funding for the enforcement account in the most recent appropriation act.
44. The Treasury Department suggests that the indirect effect of deterrence on revenues is at least three times the direct effect. See Department of the Treasury, “Internal Revenue Service: Program Summary by Appropriations Account and Budget Activity,” Fiscal Year 2017 Budget in Brief, p. 15, (PDF, 1.95 MB).
45. See Saurabh Datta, Stacy Orlett, and Alex Turk, “Individual Nonfilers and IRS-Generated Tax Assessments: Revenue and Compliance Impacts of IRS Substitute Assessments When Taxpayers Don’t File” (paper presented at the IRS–Tax Policy Center Research Conference on Tax Administration, Washington, D.C., June 18, 2015), .
46. Government Accountability Office, Tax Fraud and Noncompliance: IRS Could Further Leverage the Return Review Program to Strengthen Tax Enforcement, GAO-18-544 (July 24, 2018), p. 23,
47. See Government Accountability Office, Internal Revenue Service: Strategic Human Capital Management Is Needed to Address Serious Risks to IRS’s Mission, GAO-19-176 (March 26, 2019), .
49. Ibid, Figure 2.
50. See Janet Holtzblatt and Jamie McGuire, “Effects of Recent Reductions in the Internal Revenue Service’s Appropriations on Revenues” IRS Research Bulletin, Publication 1500 (June 2020), (PDF, 9 MB).
51. See Charles P. Rettig, IRS Commissioner, letter to the Honorable Ron Wyden, Ranking Member, Senate Finance Committee (September 6, 2019), .
The Internal Revenue Service (IRS) uses a variety of approaches to prevent fraud, determine the correct amount of taxes owed, collect taxes that were not paid, and secure returns that were not filed. Automated enforcement efforts are used to screen for fraud and to determine tax liability when returns have arithmetic or clerical errors or conflict with third-party information. Examinations are used to determine liability for more complex return issues. For unpaid debts or unfiled returns, the IRS alerts taxpayers and seeks to collect payments or a filed return.
The IRS uses several software programs that rely on computer models to flag suspicious returns or to correct a return without a formal examination.
The IRS suspends the processing of refunds for taxpayers with questionable income tax returns using two fraud detection systems: the Dependent Database and the Return Review Program. The Dependent Database combines information from the Department of Health and Human Services and the Social Security Administration that could establish a relationship between a dependent and taxpayer. The IRS checks return information against those data and applies rules to identify those returns that may have been submitted by an identity thief using another person’s name and taxpayer identification number to file a fraudulent return or obtain a fraudulent tax refund.
The Return Review Program uses third-party information returns, the taxpayer’s previous returns, and several analytical methods to produce a score that reflects the likelihood of both identify-theft fraud and refund-related fraud such as inflating the amount of wage or self-employment income to receive a larger earned income tax credit (EITC). For returns with markers of potential identity theft, taxpayers must contact the IRS to verify their identity. For returns with markers of potential refund-related fraud, tax examiners review and verify income and withholding information on the return.
The IRS is authorized by law to correct certain mathematical or clerical errors (collectively referred to as math errors) on a tax return and issue a notice to the taxpayer with the new assessment, including any applicable penalties. Such errors include using the wrong entry from a table or schedule, omitting a form needed to substantiate an entry, claiming a deduction or credit in excess of the statutory limit, and omittng Social Security numbers or taxpayer identification numbers.
Math errors that the IRS is authorized to correct are identified and corrected immediately when a tax return is processed by the IRS. A notice is sent to the taxpayer with an explanation of the error and the correction, which could be favorable to the taxpayer (resulting in a smaller balance due or a larger refund) or unfavorable (resulting in an additional tax or smaller refund). The taxpayer has 60 days from the date of the math error notice to challenge the correction. Corrected refund amounts are paid to the taxpayer within two to three weeks of the notice, regardless of whether the correction is challenged. If the correction is not challenged, any additional tax is due 60 days from the date of the notice, and the corrected refund amount already issued becomes final.
After returns have been processed, the IRS matches third-party information returns, which provide information on the amounts paid to or processed for the taxpayer, against the processed returns and selects a portion of returns with discrepancies for the Automated Underreporter program (AUR). The AUR’s process is similar to the process for a correspondence examination. Taxpayers whose returns have been updated with third-party information are notified of the change and any additional tax, penalties, or interest. For example, if wages shown on a Form W-2 are greater than the amount of wages reported on the return, the taxpayer is notified that he or she owes additional tax based on the income shown on the W-2. The taxpayer can submit documentation to verify the amount on the return or accept the change and pay the additional tax.
If the taxpayer does not respond to the notice or does not have documentation that resolves the discrepancy, the IRS sends a notice of its intention to assess any additional tax due, known as a statutory notice of deficiency. That notice gives the taxpayer 90 days (150 days if the taxpayer does not reside in the United States) to petition the Tax Court if he or she would like to challenge the AUR results without paying the recommended additional tax. Otherwise, the IRS assesses the additional tax at the end of the 90-day or 150-day period.
Comparisons between processed returns and third-party information returns are made three times a year. For a return filed by April 15, the IRS performs the first match in July and begins notifying taxpayers of discrepancies in late October. In calendar year 2011, on average, 13 months elapsed between filing and notification of a discrepancy.
Examinations, or audits, which can require a substantial amount of employees’ time, are undertaken for some returns. Examinations can be conducted by correspondence or in person, depending on the type of taxpayer and the complexity of the case. Typically, the IRS has up to three years from the date a return was due or filed to assess additional tax.
The IRS creates an inventory of returns that have audit potential. There are many ways a return can be selected to be included in that pool, which vary by type of return and by the IRS unit that would conduct the audit. Some mechanisms for selection common to all types of income tax returns are the following:
Computer Screening. Computer algorithms score the likelihood of noncompliance on individual and corporate income tax returns. The Discriminant Inventory Function (DIF) system is used on individual returns and corporate returns of entities with assets of less than $10 million, and the Discriminant Analysis System (DAS) is used for larger corporate returns. On the basis of existing data, formulas are developed that assign weights to certain characteristics of returns. The score derived from the weights is used to rank returns, with a higher score indicating a higher probability of an examination generating a significant tax change. Returns with high DIF or DAS scores are screened by tax examiners, and some are selected for examination.
Data Discrepancies. Discrepancies between information provided to the IRS by third parties (such as states, employers, banks, payment processors, brokers, or other federal agencies) on income or dependents claimed on the return can prompt a screening of a return for potential audit.
Focus Area. In the case of large and international businesses, a return may be selected for examination because it has characteristics that match an issue-based area of focus, called a “campaign.” Campaigns identify narrow issues that represent a high risk of noncompliance and select returns with those issues for audit or for notices intended to educate the taxpayer. Such campaigns may focus, for example, on taxpayers who claim an individual foreign tax credit but do not meet the requirements, taxpayers with offshore private bank accounts, or taxpayers that claim a property deduction for an energy-efficient commercial building. The IRS also periodically identifies abusive tax avoidance schemes as areas of focus, such as syndicated conservation easements, in which investors claim a deduction for a charitable contribution based on an inflated value for the conservation easement.
Related Returns. When an examination is opened, prior or subsequent returns filed by the taxpayer, or returns filed by related taxpayers like business partners, may also be selected for examination.
Referrals. Taxpayers suspected of noncompliance can be referred for IRS examination by federal, state, or local government agencies, or by citizen whistleblowers. Other IRS programs can also refer suspicious returns for examination.
Mandatory Review. Certain types of returns are always examined because of law or IRS policy. For example, the IRS examines all returns that claim a refund of more than $2 million (or $5 million for C corporations) and submits a report to the Joint Committee on Taxation. Other returns that are routinely examined include those of certain IRS employees, the President, and the Vice President. Examiners also review amended returns, which must be manually processed.
Random Selection. The IRS’s National Research Program (NRP) selects a random sample of returns for examination to provide the agency with information about voluntary compliance. Though the primary purpose of NRP examinations is to gather statistically valid data on compliance that are used in estimating the tax gap and informing computer screening for audit, corrections made as information on a return is verified can result in recommended additional tax for the taxpayer, just as it would in an operational (non-NRP) examination. The NRP’s main study is an annual sample of individual income tax returns (Form 1040), which it has conducted since the program was started in 2000. The NRP has also conducted two smaller studies—a sample of S corporation tax returns (Form 1120-S) from tax years 2003 and 2004, and a sample of employment tax returns (Form 941) from tax years 2008 to 2010.
When an auditor receives a case for examination, he or she identifies issues on the return that could change tax liability. Managers then order examinations on the basis of available staff, funding and the IRS’s enforcement priorities.
Returns with issues that could be resolved with a limited amount of additional documentation from the taxpayer are conducted entirely through correspondence. Those examinations typically involve individual returns with claims for EITC or other refundable credits, itemized deductions, or expenses related to self-employment income. A correspondence examination begins with a notice letter, informing the taxpayer that his or her return is under examination and listing documentation the taxpayer needs to provide to resolve the issue. A taxpayer has 30 days to respond to a notice letter with supporting documents or to request a 30-day extension, which is ordinarily granted. The scope of a correspondence examination can expand if the auditor identifies additional issues in the return, but it is typically limited to verifying lines in the return that are related to the issue that prompted the audit. Correspondence examinations accounted for 81 percent of individual income tax examinations and 2 percent of corporate income tax examinations in 2018.
When an examination involves more complicated issues, it requires an in-person meeting between an IRS employee and the taxpayer to review records and allow the taxpayer to provide oral testimony. The meeting may take place at an IRS office, for examinations of limited complexity; more complex examinations may occur “in the field,” at the taxpayer’s home or business.
If the information a taxpayer provides during a correspondence, office, or field examination is sufficient to resolve all issues, the IRS accepts the original return as filed with no additional tax assessed. If not, the IRS proposes changes to the return in a letter with an audit report. Examiners are responsible for recommending additional tax and for adding civil penalties if applicable. The taxpayer has 30 days to agree to any recommended additional tax and penalties or to challenge the assessment through the independent appeals function of the IRS.
If the taxpayer does not agree to the audit report or appeal its findings within 30 days, the IRS sends a statutory notice of deficiency. That notice gives the taxpayer 90 days (or 150 days if the taxpayer does not reside in the United States) to petition the Tax Court if he or she would like to challenge the examination results without paying the recommended additional tax. Otherwise, the IRS assesses the additional tax at the end of the 90-day or 150-day period.
The length of time it takes to complete an examination varies by the examination and taxpayer type. On average, in 2014, an audit took about 270 days between when a return was filed and when an examination was closed. Among taxpayers who have self-employment income or are small businesses, field audits took an average of 310 days in 2014, whereas office audits took an average of 262 days. Correspondence audits of individuals and small businesses were generally shorter; in the 2017–2018 period, they averaged 190 days for individuals and and 229 days for small businesses, with refunds sometimes frozen during that time. The length of audits of large corporations ranged, on average, between 29 months and 48 months, in 2012. Refunds are generally not frozen for field or office audits.
The IRS collects unpaid taxes and secures tax returns that have not been filed. The agency has the authority to place liens on a taxpayer’s property or seize their property to satisfy a tax debt (including garnishing wages from employers). Typically, the IRS has up to 10 years from the date taxes were assessed to collect them.
Collections cases can originate from returns that did not include full payment of taxes, assessments resulting from examinations or automated enforcement efforts that are not paid promptly, and IRS programs that identify nonfilers based on third-party information returns.
Because of limited resources and the vast pool of potential cases, the IRS prioritizes resolving cases quickly and at the lowest cost. It first sends a number of computer-generated notices to the taxpayer about the unpaid tax or delinquent return before using more labor-intensive methods. Taxpayers who file a return but do not pay the full amount of tax due receive up to four notices over 20 weeks; individual income tax nonfilers receive up to two notices and have 14 weeks to respond.
For cases that remain unresolved after the initial notice phase, the IRS may open a taxpayer delinquent account (for a return that was filed with inadequate payment) or a taxpayer delinquent investigation (for an unfiled return). Cases are then assigned to different collection processes depending on whether they are likely to be collectible, the amount and type of tax owed, and the age of the account. Interest and penalties accrue until the debt is paid in full. Within each process, an automated system ranks cases so that higher-priority cases are handled first. Cases can also be shuffled between the various processes described below.
- IRS representatives in the Automated Collection System (ACS) make telephone contact with taxpayers and issue notices of federal tax liens and levies.
- In field collection, a revenue officer makes in-person contact with the taxpayer. Managers assign cases based on the characteristics of the revenue officer, including geographic proximity and expertise, and characteristics of the case, such as its potential for collectability.
- Cases that are awaiting assignment to a revenue officer for field collection or to the ACS are held in a queue.
- Cases that are predicted to be unproductive or have not been assigned after being in the queue for a year are put on hold, or shelved. The IRS typically does not further pursue shelved cases unless the taxpayer has additional tax debts or delinquent returns, or unless penalties and interest accruing on the shelved case become sufficiently large that the case is moved back into active collection. Some shelved cases are transferred to private collection agencies (see next section).
- Some cases involving taxpayers who did not file a required return are assigned to the Automated Substitute for Return program. For those cases, the IRS creates a return based on income reported on information returns. The assessed tax is calculated on the assumption that the taxpayer’s status is either single or married filing separately (if the taxpayer previously filed as married) and that he or she claims the standard deduction. That proposed tax liability may be higher than the taxpayer’s actual tax liability would have been if he or she had filed a return.
A collections case can be resolved in several ways. Taxpayers can pay the full amount owed or apply for an installment payment plan. They can request a reduction of the amount owed (known as an offer in compromise). If a taxpayer is facing financial hardship, the IRS can analyze his or her financial situation and potentially suspend active collection of tax debts. Otherwise, the IRS may seize a taxpayer’s property to satisfy the tax debt or place a federal tax lien on a taxpayer’s property. Generally, taxpayers have 30 days to appeal IRS collections actions after receiving a notice. Following the appeals determination, the taxpayer has 30 days to contest it in Tax Court.
The Fixing America’s Surface Transportation Act of 2015 required that the IRS use private collection agencies (PCAs) to collect unpaid assessments that the IRS is not currently pursuing. Cases that are eligible to be transferred are generally those that the IRS has shelved, those that it did not assign to an employee for collection, or those in which there has been no contact between the IRS and the taxpayer for over a year. Certain cases, such as those involving taxpayers who are under age 18 or who are victims of identity theft, cannot be transferred. The Taxpayer First Act of 2019 further restricted the IRS from transferring the tax debts of low-income taxpayers to PCAs.
PCAs receive a commission of 25 percent of the amount collected, and the IRS can retain another 25 percent of the amount collected to hire and train “special compliance personnel” to work in field collection or in the ACS. The IRS began transferring cases to PCAs in 2017. As of September 2018, PCAs had received responsibility for collecting about $5.7 billion worth of debt from more than 700,000 taxpayers. The PCAs had collected $89 million and resolved about 38,000 cases with full payment or with an installment agreement through that time. During that period, the IRS paid $16 million in commissions and spent another $50 million to implement and manage the program, resulting in a net gain of $22 million.
The use of PCAs can allow collection of tax debt the IRS does not have the resources to pursue. Revenue generated from the use of PCAs can be counted in the Congressional Budget Office’s cost estimate for legislation authorizing their use, unlike estimates of revenue generated from increased enforcement spending. However, previous IRS programs using PCAs in 1996 and from 2006 to 2009 ended because they were not cost effective. Cases assigned to PCAs are, on average, almost four years old—but after three years, debts tend to be uncollectible. The IRS predicts that those cases would have low returns, and given the agency’s limited resources, it is unlikely that the IRS would pursue them.
Some advocates for taxpayers are concerned that the use of PCAs may impose hardships on taxpayers. Although the IRS is required to take into account the taxpayer’s ability to pay his or her tax debt, those guidelines do not apply to PCAs, and they do not have the ability to offer alternatives to collection such as an offer in compromise. The use of PCAs may also make taxpayers more vulnerable to identity theft if taxpayers cannot distinguish between a phone call from a legitimate PCA or an impersonator.
1. The IRS views the Dependent Database and the Return Review Program as complementary detection systems for identity theft fraud. See Treasury Inspector General for Tax Administration, The Return Review Program Increases Fraud Detection; However, Full Retirement of the Electronic Fraud Detection System Will Be Delayed, Reference Number 2017-20-080 (September 25, 2017), (PDF, 336 KB). The Dependent Database is also used to select individual returns with refundable credits for examination, which occurs after the fraud detection process but sometimes before refunds are issued.
2. For more information about the Return Review Program, see Government Accountability Office, Tax Fraud and Noncompliance: IRS Could Further Leverage the Return Review Program to Strengthen Tax Enforcement, GAO-18-544 (July 24, 2018), .
3. See 26 U.S.C. §6213(g)(2) for a complete list of mathematical and clerical errors.
4. The AUR program automatically applies accuracy-related penalties in cases of substantial understating of income. An examiner evaluates whether the taxpayer had reasonable cause for understatement and, if so, can waive the automatic penalty. See Treasury Inspector General for Tax Administration, Automated Underreporter Program Tax Assessments Have Increased Significantly; However, Accuracy-Related Penalties Were Not Always Assessed When Warranted, Reference Number 2015-30-037 (May 8, 2015),
5. See Government Accountability Office, Tax Refunds: IRS Is Exploring Verification Improvements, but Needs to Better Manage Risks, GAO-13-515 (June 4, 2013), Figure 5,
6. The Small Business/Self-Employed Division examines corporate returns of entities with assets of less than $10 million and individual returns that do not claim refundable credits. About 20 percent of examinations closed by the division in 2014 were selected for examination because of the return’s DIF score. See Government Accountability Office, IRS Return Selection: Certain Internal Controls for Audits in the Small Business and Self-Employed Division Should Be Strengthened, GAO-16-103 (December 16, 2015), Figure 3, . DAS is the main selection method of the Large Business and International Division, which examines corporate returns with assets of more than $10 million. See Government Accountability Office, IRS Return Selection: Improved Planning, Internal Controls, and Data Would Enhance Large Business Division Efforts to Implement New Compliance Approach, GAO-17-324 (March 28, 2017), .
7. Compliance data collected by the National Research Program from examinations are used to improve the DIF algorithm.
8. The Wage and Investment Division examines individual returns claiming refundable credits. In 2014, 59 percent of that division’s examinations were selected using the Dependent Database, which uses internal IRS data as well as external data on child custody from the Department of Health and Human Services and birth information from the Social Security Administration to identify discrepancies on returns. See Government Accountability Office, IRS Return Selection: Wage and Investment Division Should Define Audit Objectives and Refine Other Internal Controls, GAO-16-102 (December 17, 2015), .
9. Since January 2017, the Large Business and International Division has introduced a total of 53 compliance campaigns that guide the selection of returns for audit. For a list of all of the division’s active and retired campaigns, see Internal Revenue Service, “Large Business and International Compliance Campaigns” (accessed June 17, 2020), .
10. See Internal Revenue Service, Publication 556: Examination of Returns, Appeal Rights, and Claims for Refund (September 2013), (1.2 MB).
11. See Joint Committee on Taxation, “Joint Committee Statutory Refund Review” (accessed June 17, 2020),
12. Since 2007, the National Research Program’s individual compliance study has combined examination results from a rolling three-year period, with annual sample sizes of about 13,000 returns. Previously, this study selected about 45,000 returns from a single tax year. See Internal Revenue Service, “IRS Updates National Research Program for Individuals,” IR-2007-113 (news release, June 6, 2007), (19 KB)
13. See Government Accountability Office, IRS Correspondence Audits: Better Management Could Improve Tax Compliance and Reduce Taxpayer Burden, GAO-14-479 (June 5, 2014), .
14. The amount of time allotted for an audit by a tax compliance officer is much shorter than the amount allotted for an audit by a revenue agent because tax compliance officers address less complicated issues. See National Taxpayer Advocate, “Office Examination: The IRS Does Not Know Whether Its Office Examination Program Increases Voluntary Compliance or Educates the Audited Taxpayers About How to Comply in the Future,” Annual Report to Congress 2018 (February 2019), vol. 1, pp. 153–163, .
15. For example, if an individual taxpayer underreported income and the recommended additional tax from examination of his or her return is greater than $5,000, the taxpayer may be assessed a penalty of 20 percent of the recommended additional tax for substantially understating income. In 2018, $1.5 billion in accuracy-related penalties was assessed on individual tax returns (from any tax year), and $0.2 billion in accuracy-related penalties was assessed on corporate tax returns. See Internal Revenue Service, Internal Revenue Service Data Book, 2018 (June 30, 2020), .
16. See Government Accountability Office, IRS Return Selection: Certain Internal Controls for Audits in the Small Business and Self-Employed Division Should Be Strengthened, GAO-16-103 (December 16, 2015), Table 3, .
17. See National Taxpayer Advocate, “Correspondence Examinations: The IRS’s Correspondence Examination Procedures Burden Taxpayers and Are Not Effective in Educating the Taxpayer and Promoting Future Voluntary Compliance,”Annual Report to Congress 2018 (February 2019), vol. 1, pp. 126–141, .
18. See Government Accountability Office, Corporate Tax Compliance: IRS Should Determine Whether Its Streamlined Corporate Audit Process Is Meeting Its Goals, GAO-13-662 (August 2013), Figure 1,
19. For more information on the collections process for individual income tax nonfilers, see Saurabh Datta, Stacy Orlett, and Alex Turk, “Individual Nonfilers and IRS-Generated Tax Assessments: Revenue and Compliance Impacts of IRS Substitute Assessments When Taxpayers Don’t File” (paper presented at the IRS–Tax Policy Center (TPC) Research Conference, Washington, D.C., June 18, 2015), .
20. Unpaid tax debt is subject to penalties and interest. Interest accrues on any unpaid tax from the date the return is due until the date of full payment, at the interest rate on federal short-term debt plus 3 percent. Penalties for failure to file or failure to pay are calculated as a percentage of the tax due, and grow each month that the return remains unfiled or the tax debt unpaid. See Internal Revenue Service, “Topic No. 653: IRS Notices and Bills, Penalties, and Interest Charges” (July 1, 2020),
21. The IRS also includes one personal exemption in calculating tax liability on a substitute return. However, the 2017 tax act (Public Law 115-97) temporarily suspended personal exemptions for tax years 2018 to 2025.
22. If the financial analysis shows that a taxpayer’s income is not sufficient to provide a minimum standard of living, the IRS places the taxpayer in “currently not collectible” status and does not actively try to collect that person’s tax debt. The debt will continue to accrue penalties and interest, and income tax refunds may be used to offset the debt. See Taxpayer Advocate, “Currently Not Collectible” (accessed June 22, 2020), .
23. The appeals process differs by the type of collections action. For more details, see Internal Revenue Service, Collection Appeal Rights, Publication 1660 (revised January 2020), (657 KB).
24. See Government Accountability Office, Tax Debt Collection Contracts: IRS Analysis Could Help Improve Program Results and Better Protect Taxpayers, GAO-19-193 (March 29, 2019), .
25. See Congressional Budget Office, cost estimate for the conference agreement on H.R. 22, the FAST Act (December 2, 2015), .
26. The earlier programs differ from the current one in several ways, including the structure of payments to PCAs and the types of cases the IRS transferred to PCAs.
27. See Treasury Inspector General for Tax Administration, Private Debt Collection Was Implemented Despite Resource Challenges; However, Internal Support and Taxpayer Protections Are Limited, Reference Number 2018-30-052 (September 5, 2018), (PDF, 3.3 MB).
28. See National Taxpayer Advocate, “Private Debt Collection: The IRS’s Expanding Private Debt Collection Program Continues to Burden Taxpayers Who Are Likely Experiencing Economic Hardship While Inactive Private Collection Agency Inventory Accumulates,” Annual Report to Congress 2018 (February 2019), vol. 1, pp. 277–294, .
The Congressional Budget Office estimated the effect of changes in revenues collected by the Internal Revenue Service (IRS) by calculating the amount that revenues would increase for each additional dollar of enforcement spending. The starting point for its projections was the estimates that the IRS typically provides as part of the President’s annual set of budgetary proposals. The IRS estimates a return-on-investment (ROI) factor for spending on new enforcement initiatives by calculating the expected revenues that would be raised from taxes, interest, and penalties as a result of the new initiatives divided by their additional cost. That additional cost is based on the cost of the employees required to implement the new initiatives—that is, the pay grade of those employees and the number of hours they work. The IRS’s ROI factors have ranged from 5.2 to 9.2 since 2016.
CBO used IRS estimates from multiple years of the President’s budget to estimate a general ROI factor equal to 6.4 for broadly increasing the IRS’s enforcement funding. CBO anticipates that the return on enforcement would rise to 6.4 over the first three years of an initiative as employees finished training and became more skilled in enforcement. In CBO’s assessment, that return would then decline in the later years of a given set of initiatives and for initiatives that start in later years of a funding proposal. The decline reflects CBO’s expectation that the IRS would focus first on the cases that bring in the most potential revenue (the amount of spending times the expected return) and then move to cases that are likely to yield less revenue. In addition, CBO expects that taxpayers would adjust their behavior over time to avoid the new compliance activities. Finally, CBO estimates how much revenue would be collected and when the IRS would receive that revenue.
To illustrate its approach, CBO calculated how spending and revenues would be affected by each new set of initiatives that would result from a $20 billion increase in the IRS’s appropriations over 10 years (See Table B-1). In this illustration, initiatives that begin in 2021 receive $500 million in funding each year from 2021 to 2030 and generate $20 billion in revenues over that period. The 2022 initiatives similarly receive $500 million in funding each year from 2022 to 2030, generating $15 billion in revenue by 2030. Revenues are lower for each successive set of initiatives because CBO projects that they would focus on lower-yielding enforcement activities and because less of the potential revenue would be collected within the 10-year period. (For example, additional revenues generated by the 2021 initiatives during the first six years are projected to be almost twice the revenues generated by the 2025 initiatives in their first six years.)
This report was prepared at the request of the Ranking Member of the Senate Budget Committee. In keeping with the Congressional Budget Office’s mandate to provide objective, impartial analysis, the report makes no recommendations.
Kathleen Burke and Shannon Mok wrote the report, with guidance from Janet Holtzblatt (formerly of CBO), Joseph Rosenberg, and John McClelland. Matthew Pickford provided useful comments. Brian Erard of B. Erard & Associates, Dayanand Manoli of the University of Texas at Austin, and Erin Towery of the University of Georgia also provided helpful comments. The assistance of external reviewers implies no responsibility for the final product, which rests solely with CBO.
Jeffrey Kling, Wendy Edelberg (formerly of CBO), and Robert Sunshine reviewed the report. Elizabeth Schwinn was the editor, and Casey Labrack was the graphics editor. An electronic version is available on CBO’s website .
Phillip L. Swagel