A taxpayer is indicted on a tax-evasion charge under IRC §7201. The mainstream coverage of the indictment focuses on the alleged conduct, the dollar amounts at issue, the potential maximum sentence (“up to 5 years per count, plus fines”). What the coverage rarely walks is the actual resolution path: the percentage of federal tax-fraud cases that go to trial versus settle by plea, the typical sentence ranges under the U.S. Sentencing Guidelines, the differences between the §7201 felony evasion charge and the lesser §7203 misdemeanor failure-to-file charge, the role of the tax loss figure in driving the sentence. The numbers exist, are published, and produce a materially different picture of how these cases actually resolve than the headline coverage implies.
The DOJ Tax Division publishes annual case-disposition statistics; the U.S. Sentencing Commission publishes sentencing data by offense type; the U.S. Sentencing Guidelines themselves specify how the tax loss number translates into the offense level that drives the sentence range. Pulling these sources together produces the actual landscape practitioners and defendants should understand. Most tax-fraud cases do not go to trial. The majority that resolve by plea produce sentence ranges that are bounded by specific provisions of the Guidelines, and the relevant inputs (tax loss, criminal history, acceptance of responsibility) are concrete enough that a realistic outcome estimate is possible early in the case. This article walks the architecture.
The §7201 vs §7203 distinction
IRC §7201 is the felony tax-evasion statute. The elements are: (a) a tax due and owing, (b) an affirmative act of evasion (filing a false return, hiding assets, false statements, structured transactions designed to evade), and (c) willfulness, the defendant’s conscious purpose to evade. The maximum statutory penalty is 5 years’ imprisonment per count, plus fines and the cost of prosecution. Each tax year typically constitutes a separate count.
IRC §7203 is the misdemeanor failure-to-file or failure-to-pay statute. The elements are: (a) required to file a return or pay a tax, (b) failure to do so, (c) willfulness. The maximum statutory penalty is 1 year imprisonment per count, plus fines. Each tax year is a separate count.
The DOJ Tax Division generally pursues §7201 charges when the conduct involves affirmative acts of evasion; §7203 charges when the conduct is willful non-filing or non-payment without affirmative evasion acts. The choice of charge dramatically affects the eventual sentence, §7201 carries 5x the maximum statutory exposure of §7203, and the Guidelines treat the two offenses differently in the underlying offense-level computation.
The Sentencing Guidelines architecture
U.S. Sentencing Guidelines §2T1.1 covers tax offenses under §7201, §7203, and related statutes. The Guidelines compute an offense level based primarily on the tax loss figure, with the offense level translating to a sentence range when combined with the defendant’s criminal history category.
The tax-loss table in §2T4.1 produces the base offense level. Approximate ranges:
| Tax loss | Base offense level |
|---|---|
| \$2,500 or less | 6 |
| \$2,501 – \$6,500 | 8 |
| \$6,501 – \$15,000 | 10 |
| \$15,001 – \$40,000 | 12 |
| \$40,001 – \$100,000 | 14 |
| \$100,001 – \$250,000 | 16 |
| \$250,001 – \$550,000 | 18 |
| \$550,001 – \$1.5M | 20 |
| \$1,500,001 – \$3.5M | 22 |
| \$3,500,001 – \$9.5M | 24 |
| \$9,500,001 – \$25M | 26 |
| Above \$25M | continues upward in \$X bands |
Specific offense characteristics add levels in some cases: sophisticated means (+2), use of a corporate entity structure to evade (+2 in some applications), failure to report income from criminal activity (+2). Acceptance of responsibility under §3E1.1 generally subtracts 2-3 levels for defendants who plead guilty and demonstrate timely acceptance.
A defendant with a \$300,000 tax loss, no specific offense characteristic enhancements, full acceptance-of-responsibility credit (-3), and Criminal History Category I (the default for first-time offenders) would compute roughly: base level 18, minus 3 for acceptance, equals offense level 15. Under the Guidelines sentencing table, offense level 15 / Category I produces a range of 18-24 months.
What the plea-versus-trial split actually looks like
DOJ Tax Division annual statistics show that the vast majority of federal tax-crime cases resolve by plea rather than by trial. The DOJ Tax Division’s most recent published case-disposition data, covered in the agency’s Fiscal Year 2023 Annual Report and consistent with the multi-year trend reflected in the FY 2018 through FY 2022 reports, shows that approximately 90-95% of charged tax cases proceeding past indictment resolve by guilty plea, with a small remainder going to trial. The conviction rate at trial in tax cases is itself high. The U.S. Sentencing Commission Annual Sourcebook (for FY 2022 published in 2023, and earlier Sourcebooks in the same series) reports the trial conviction rate for tax offenses at well above 90%, again consistent with the multi-year pattern. The exact figures shift year over year as case mix changes, but the general structure, overwhelming plea rate, high trial conviction rate among the few that proceed, is stable across the published series.
Several factors drive the plea rate. The Acceptance of Responsibility reduction (§3E1.1) is meaningful, a 2-3 level reduction off the offense level translates to a substantially shorter sentence range. The cost of trial defense in a federal tax case is high, often six figures, and the marginal probability of acquittal on the §7201 elements is modest given the IRS-CI investigative architecture (which generally develops cases for 1-3 years before referring to DOJ). The certainty of a plea outcome, combined with the favorable Guidelines treatment, makes the plea the dominant path even for defendants who maintain factual innocence on specific elements.
The cases that do go to trial typically involve either (a) defendants who genuinely contest the underlying facts (not just the legal interpretation), (b) defendants for whom the marginal sentence difference between plea and trial is small enough that the plea-incentive disappears, or (c) cases with novel legal theories where appellate-level review may produce a favorable ruling.
The actual sentence ranges that result
The combination of Guidelines computation and plea-rate pattern produces a relatively consistent sentence landscape:
For tax-loss cases under \$250,000, sentences under §7201 frequently produce probationary or short-incarceration outcomes (6-18 months), particularly for first-time offenders with full acceptance of responsibility and no aggravating offense characteristics.
For tax-loss cases in the \$250,000-\$1.5M range, sentences typically run 18-36 months, meaningful incarceration but well below the §7201 statutory maximum per count.
For tax-loss cases above \$1.5M, sentences scale up materially. Cases in the \$5M-\$25M tax-loss range produce sentences in the 4-7 year range; cases above \$25M can run 7-10+ years.
These ranges are approximate and assume Criminal History Category I, no specific offense characteristic enhancements, and full acceptance of responsibility. Defendants with prior convictions, with cases involving sophisticated means or money laundering, or who proceed to trial without prevailing typically receive sentences materially above the equivalent-loss plea figures.
Restitution under 18 USC §3663A is generally ordered in tax-fraud cases, restitution covers the actual tax loss to the United States, plus interest. Civil penalty assessments under IRC §6663 (75% civil fraud penalty) typically follow separately from the criminal case, often added on the IRS civil side after the criminal resolution.
What this means for defendants
Two practical points:
The first is that the dollar of tax loss is the single most consequential number in the case. Sentencing positions, plea-negotiation leverage, and the realistic outcome range all anchor on the tax-loss figure. Defendants who can credibly contest the IRS’s tax-loss calculation, by demonstrating that some of the alleged loss reflects legitimate deductions, basis adjustments, or substantiation that the agent missed, produce materially better outcomes than defendants who concede the loss figure and litigate only the willfulness element.
The second is that the cost-benefit of plea-versus-trial is calculable. A defendant facing a Guidelines range of 24-30 months with full acceptance, facing a likely sentence in the 36-48 range without acceptance, can compute the trade-off explicitly. The cost of trial defense (often \$200K-\$500K+ for a federal tax case), the marginal probability of acquittal on the specific elements (generally modest given the IRS-CI investigation pattern), the cost of additional months of incarceration if convicted, these inputs produce an expected-value calculation that drives the resolution decision more than the inclination to “fight” or “concede” abstractly.
The headlines treat tax-fraud cases as binary outcomes, convicted or acquitted, prison or freedom. The actual resolutions are graduated, math-driven, and resolved by plea in roughly 19 of every 20 indicted cases. Understanding the architecture changes how the resolution path can be navigated.
Authority: IRC §7201 (felony tax evasion); IRC §7203 (misdemeanor failure to file or pay); IRC §7206 (false statement on return); IRC §6663 (civil fraud penalty); U.S. Sentencing Guidelines §2T1.1 (tax offense base computation); U.S. Sentencing Guidelines §2T4.1 (tax loss table); U.S. Sentencing Guidelines §3E1.1 (acceptance of responsibility); 18 USC §3663A (restitution); DOJ Criminal Tax Manual; DOJ Tax Division annual statistics (case dispositions); U.S. Sentencing Commission Annual Report and Sourcebook; United States v. Klein, 247 F.2d 116 (2d Cir. 1957) (Klein conspiracy charge); United States v. Pomponio, 429 U.S. 10 (1976) (willfulness standard); Cheek v. United States, 498 U.S. 192 (1991) (good-faith-belief defense to willfulness).