This article originally appeared in the March 4, 2019, issue of Tax Notes.
]
Kim M. Bloomquist, now retired, worked as a senior economist in the IRS Office of Research and as an operations research analyst at the Taxpayer Advocate Service. His earlier career included work as an economist in the IRS Chicago district office and the U.S. Army Corps of Engineers.
In this report, Bloomquist demonstrates that willfully ignoring geographic location in audit case selection does not ensure balanced regional coverage, especially when IRS policy favors auditing earned income tax credit taxpayers at higher rates than other taxpayers.
In a recent series of articles,1 ProPublica described how the decline in IRS enforcement budgets between 2010 and 2017 has caused a 42 percent drop in the number of audits for all taxpayers and an even steeper decline (69 percent) in audits of taxpayers with adjusted gross income of $500,000 or more. Over the same period, the audit rate also fell for taxpayers who claimed the earned income tax credit (mainly those with AGIs less than $25,000). However, the decline (36 percent) was only about half as large as that observed for the wealthiest taxpayers. Effectively, there was not only an overall reduction in audit activity, but also a realignment of IRS resources from high-income taxpayers to low- and middle-income EITC claimants.
A predictable consequence of this realignment of IRS enforcement is a growing regional bias in audit case selection. This is because EITC taxpayers are not randomly distributed throughout the country (Figure 1).
A 2016 Government Accountability Office report2 criticized the IRS Small Business/Self-Employed Division for not clearly defining the concept of fairness used in selecting tax returns to audit or developing measures to ensure that its return selection process is consistent with such a standard.3 In its response to the GAO report, the IRS stated that taxpayers’ geographic location is purposely ignored in audit case selection to “balance coverage across the country.” The purpose of this report is to show that willfully ignoring geographic location in audit case selection does not ensure balanced regional coverage, especially when IRS policy favors auditing EITC taxpayers at higher rates relative to other taxpayers.
In the next section I describe a method of estimating the number of county-level IRS tax audits using publicly available data. While readily acknowledging the limitations of relying on public data sources, I believe (based on a previous career working with IRS tax return population data) it is possible to reasonably replicate the spatial pattern of individual taxpayer audits using audit coverage rates published in the annual IRS Data Book in combination with county tax return data on the IRS website. The third section displays county-level estimates of audit intensity4 for tax years 2012-2015 and discusses how the IRS policy of auditing EITC taxpayers more heavily than other groups of taxpayers with known high rates of noncompliance (for example, non-farm business filers not claiming the EITC) leads to a regional bias in audit return selection.
Estimation of County-Level Tax Audits
The IRS publishes detailed tax return data for individual taxpayers by AGI category for all U.S. counties. These data reflect filing activity by tax year. The IRS also publishes examination coverage rates in its annual Data Book. Table 9a of the IRS Data Book displays audit coverage rates for business and nonbusiness tax returns and for tax returns with and without the EITC. Table 9b shows examination coverage for all individual tax returns (including EITC returns) by AGI category. The audit coverage rates in the IRS Data Book are calculated using fiscal-year audits in the numerator and calendar-year tax returns in the denominator. For example, the audit coverage rate of 0.6 percent for all individual taxpayers in fiscal 2017 is calculated as the number of fiscal 2017 tax audits (933,785) divided by the number of calendar year 2016 returns (149,919,416). A key assumption used in this report is that most individual tax returns filed in a calendar year are for the immediately preceding tax year. Thus, fiscal year 2017 audit coverage rates are assumed to consist primarily of tax year 2015 returns (which would, in most cases, have been filed in calendar year 2016).
Table 1 shows the exam coverage rates for tax year 2015 used in this report. The AGI categories correspond to those in the IRS county data. The exam coverage rate of 0.0137 for EITC returns is the weighted average of the coverage rate for EITC returns found in Table 9a of the 2017 IRS Data Book. The IRS county data includes EITC filers only for the first five AGI categories (with less than 1 percent of EITC filers in the first AGI category of less than $1). The tax year 2015 exam coverage rates for non-EITC filers shown in Table 1 include published IRS numbers and values derived by the author. The audit rate for the first AGI category (less than $1) and for AGI categories of $75,000-$100,000 and $100,000-$200,000 are taken from Table 9b of the 2017 IRS Data Book. The exam coverage rate for the AGI category of $200,000 or more is a weighted average of the coverage rates for returns with at least $200,000 shown in Table 9b of the 2017 IRS Data Book. Finally, the exam coverage rate of 0.0037 for non-EITC returns with AGI between $1 and $75,000 is a weighted average of four entries in Table 9a of the 2017 IRS Data Book: nonbusiness, non-EITC returns with total positive income less than $200,000 and business, non-EITC returns with total positive income less than $200,000 and total gross receipts less than $100,000.
AGI Category | EITC Returns | Non-EITC Returns |
|---|---|---|
< $1 | 0.0137 | 0.0255 |
$1 < $10,000 | 0.0137 | 0.0037 |
$10,000 < $25,000 | 0.0137 | 0.0037 |
$25,000 < $50,000 | 0.0137 | 0.0037 |
$50,000 < $75,000 | 0.0137 | 0.0037 |
$75,000 < $100,000 |
| 0.0045 |
$100,000 < $200,000 |
| 0.0047 |
≥ $200,000 |
| 0.0104 |
Source: 2017 IRS Data Book Table 9a and 9b and author’s calculations. Italicized entries are derived as described in the text. | ||
In the following analysis, I apply the previously described approach to derive exam coverage rates for tax years 2012-2015. I apply these rates to the IRS counts of EITC and non-EITC returns at the county level. Table 2 compares total actual audits of individual taxpayers for tax years 2012-2015 to estimated audits. The overall error rates for both EITC and non-EITC audits show that the estimated audit coverage rates appear reasonable for the nation, although county-level estimates may differ significantly from actual audits.5
Tax Year | EITC Audits | Non-EITC Audits | ||||
|---|---|---|---|---|---|---|
Actual | Estimated | Percent Difference | Actual | Estimated | Percent Difference | |
2012 | 480,989 | 488,563 | 1.57% | 761,490 | 773,826 | 1.62% |
2013 | 478,032 | 471,841 | -1.30% | 750,085 | 755,493 | 0.72% |
2014 | 428,207 | 442,794 | 3.41% | 606,748 | 629,036 | 3.67% |
2015 | 381,269 | 383,288 | 0.53% | 552,516 | 561,219 | 1.58% |
Total | 1,768,497 | 1,786,486 | 1.02% | 2,670,839 | 2,719,574 | 1.82% |
Source: IRS Data Book Table 9a and author’s calculations. | ||||||
Results and Discussion
The estimated number of audits are totaled over the four tax years and divided by the number of tax returns in tax year 2015. The resulting percentage is defined as a measure of audit intensity. Multiple years are used instead of a single year to increase the number of audits in the numerator for counties with small taxpayer populations.
Figure 2 displays this measure of audit intensity for tax years 2012-2015. A visual comparison of figures 1 and 2 shows, as expected, a similar spatial distribution for both audit intensity and the percentage of taxpayers claiming the EITC. Audit intensity is generally highest in the Southern states and some counties in the Northern Plains, Mountain, and Western states. Audit intensity is generally lower in the upper Midwest, Mid-Atlantic, and New England states. No audits were estimated for three counties with very small taxpayer populations over the four years covered in this analysis.6 This explains the audit intensity value of zero in the legend. The national average audit intensity for tax years 2012 to 2015 is 3.03 percent.
Table 3 lists the top 10 and bottom 10 counties with at least 1,000 taxpayers by audit intensity. Eight of the 10 counties with the highest audit intensity are in Mississippi. In tax year 2015, 51 percent of the taxpayers in these 10 counties claimed the EITC. The combined audit intensity rate in tax year 2012 to 2015 was 4.6 percent. It should also be noted that the 2017 population of these 10 counties is 79 percent nonwhite (nearly all black or African-American) according to the U.S. Census Bureau.7
In contrast, among the 10 counties with the lowest audit intensity only 10 percent of taxpayers claimed the EITC in tax year 2015 and just 7 percent of the 2017 population was nonwhite. The combined audit intensity in tax years 2012 to 2015 was 2.5 percent, or about half the audit intensity of the highest 10 counties.
County | State | Audit Intensity | Tax Returns | Estimated Audits |
|---|---|---|---|---|
Top 10 Counties | ||||
Humphreys | Mississippi | 4.81 | 3,410 | 164 |
Tunica | Mississippi | 4.70 | 4,740 | 223 |
East Carroll | Louisiana | 4.67 | 2,610 | 122 |
Coahoma | Mississippi | 4.65 | 9,610 | 447 |
Noxubee | Mississippi | 4.60 | 4,650 | 214 |
Holmes | Mississippi | 4.57 | 7,330 | 335 |
Quitman | Mississippi | 4.57 | 2,760 | 126 |
Sharkey | Mississippi | 4.57 | 1,860 | 85 |
Claiborne | Mississippi | 4.55 | 3,520 | 160 |
Greene | Alabama | 4.45 | 3,370 | 150 |
Bottom 10 Counties | ||||
Mercer | Ohio | 2.55 | 21,110 | 539 |
North Slope | Alaska | 2.55 | 3,760 | 96 |
Sumter | Florida | 2.54 | 55,010 | 1,398 |
Sherburne | Minnesota | 2.54 | 43,900 | 1,113 |
Sargent | North Dakota | 2.54 | 2,050 | 52 |
Storey | Nevada | 2.54 | 1,890 | 48 |
Washington | Wisconsin | 2.51 | 69,170 | 1,739 |
Calumet | Wisconsin | 2.51 | 24,060 | 603 |
Putnam | Ohio | 2.49 | 17,430 | 434 |
Denali | Alaska | 2.17 | 1,200 | 26 |
Source: IRS county data (taxpayers) and author’s calculations. | ||||
Summary and Recommendations
IRS budget reductions in recent years have led to a steep decline in the number of audits overall but a much smaller relative reduction in audit coverage for low-income EITC taxpayers compared with taxpayers with AGIs exceeding $500,000. This trend, if not halted and reversed, will only exacerbate the existing regional bias in audit return selection and further concentrate the economic and psychological burdens associated with IRS enforcement activities mainly on groups of taxpayers least able to afford it.
What could the IRS and tax policymakers do to change this situation? An obvious place to start would be to increase the IRS enforcement budget to enable more audits of high-income taxpayers. However, even with no increase in its budget, the IRS should broaden its audit selection criteria to include other significant areas of compliance risk and not focus so narrowly on EITC claimants. Lastly, the IRS should explicitly take into consideration any regional bias of its audit selection criteria, intentional or otherwise, by ensuring the audit coverage rate for each state does not vary substantially from the national average. For example, if the national audit coverage rate is 1 percent, then audits might be assigned to each state so that state-level audit coverage is between 0.95 percent and 1.05 percent. This would allow the IRS some flexibility to conduct special programs while ensuring that taxpayers in all states are treated fairly. While regional equity is not the sole factor the IRS ought to consider in achieving fairness for all taxpayers, the main point raised by this report is that the current policy of ignoring geography in selecting tax audits does not achieve the balanced coverage the IRS claims.
FOOTNOTES
1 See ProPublica, “Gutting the IRS” (Dec. 2018).
2 See GAO, “IRS Return Selection: Certain Internal Controls for Audits in the Small Business and Self-Employed Division Should Be Strengthened,” GAO-16-103 (Dec. 16, 2015).
3 The concept of fairness is embodied in the IRS mission statement: “Provide America’s taxpayers top-quality service by helping them understand and meet their tax responsibilities and enforce the law with integrity and fairness to all.”
4 In this report, audit intensity is defined as total estimated audits for tax years 2012 to 2015 divided by the total number of taxpayers in tax year 2015.
5 The author will provide the data used in this report on request. Email him at kbloomq@aol.com.
6 The reader is reminded that the data shown in Figure 2 are estimates and not actual audits. The author would welcome corrections from the IRS to address any gross inaccuracies in these results.
7 Census Bureau, “Annual Estimates of the Resident Population by Sex, Race, and Hispanic Origin for the United States, States, and Counties: April 1, 2010 to July 1, 2017” (June 2018).
END FOOTNOTES