Site icon Washington University Law Review

A World Without Federal Sentencing Guidelines

bmallory26
Merchant_A World Without Federal Sentencing GuidelinesDownload

Sam J. Merchant[1]*

Abstract

Most participants and observers of the criminal system perceive the Federal Sentencing Guidelines as excessively harsh. A foundational question has persisted since the creation of the Guidelines: is a guideline-based regime actually preferable, or should we embrace complete judicial discretion in sentencing? For decades, analysts have resorted to hypothetical cases to explore this issue. But a little-known world exists in which real federal sentences are imposed without any reference to sentencing guidelines: U.S. Sentencing Guideline §2X5.1. This Article is the first to compare actual sentences imposed with and without guidelines for the same offenses and same types of offenders.

The analysis reveals that judges tend to impose harsher sentences in the world without sentencing guidelines. Variability is also more pronounced in sentences without guidelines: after a conviction for child endangerment, some parents received two years in prison and others received fifteen (even after adjusting for severity). Two Black men convicted of a fourth and fifth non-accident DUI received ten years in prison while other offenders received probation (the median sentence is around twelve months’ imprisonment). Recent Supreme Court cases affecting tribal jurisdiction, and the January 6, 2021, “Capitol Breach” cases, have led to an explosion in the number of these cases in many districts.

This Article argues that sentencing within a guideline framework, or within a data-based framework when guidelines are inapplicable, provides more certainty and minimizes unwarranted disparities. The conclusions offer critical insights to states or other systems that do not currently have sentencing guidelines or do not meaningfully collect sentencing data. Finally, this Article offers recommendations to judges, the United States Probation Office, and the Sentencing Commission to help advance a more just and efficient sentencing system.

Introduction

Craig Alan Morrison and his girlfriend Amanda Lyn Walker brought Ms. Walker’s three-year-old son to the emergency room saying that the boy had injured himself after jumping off his bed and hitting his head on a scooter.[2] Doctors observed additional bruising on the child that did not align with the adults’ story.[3] A pediatric physician examined the boy and concluded that he had been physically abused, which the adults denied.[4] Police investigated and the two were ultimately charged in federal court with child abuse.[5] Both were convicted at trial.[6] As set forth later in this Article, statistically they should expect time in prison, but what would be an appropriate prison term? There are no federal sentencing guidelines for child abuse.

The judge sentenced Mr. Morrison to 300 months (25 years) in prison.[7] Prior to sentencing, defense counsel cited what limited data he could locate regarding comparable cases from within the Tenth Circuit, which showed sentences that were much shorter than 300 months.[8] The trial judge declined to rely on those cases, and the Tenth Circuit affirmed, stating that what matters in analyzing unwarranted disparity is “nationwide disparities,” not intracircuit disparities.[9] The problem was that the nationwide average for an offense like Mr. Morrison’s is between 35 to 60 months.[10] The 25-year sentence was almost twice as long as sentences where the child dies as a result of physical child abuse.[11] But because the Sentencing Commission has not collected or released data on these kinds of sentences, neither the parties, nor trial judge, nor appellate court knew (until this Article) that the 300-month sentence vastly exceeded the nationwide average. As discussed throughout this Article, other examples of disparate sentences abound in cases like these—those sentenced without guidelines.

Throughout most of America’s history judges had wide discretion to sentence within broad statutory ranges of punishment (for example, anywhere between probation and 20 years’ incarceration) “without meaningful legal guidance” and “typically without offering a detailed explanation for the sentence.”[12] Congress passed the Sentencing Reform Act of 1984 (SRA) to address “unwarranted sentencing disparities” that Congress had observed in that system.[13] In the SRA, Congress created the United States Sentencing Commission (Sentencing Commission) and directed it to create sentencing guidelines that would result in similar offenders engaging in similar conduct being sentenced within the greater of 25% or 6 months of each other.[14] Thus, to Congress, if similar offenders commit similar offenses and are not sentenced within 25% or 6 months of each other, an unwarranted disparity between the offenders might exist.[15]

There is a sentencing guideline for most offenses that are prosecuted in federal court. But there are some offenses that have no sentencing guidelines. Judges have struggled with sentencing these cases for decades.[16] Section 2X5.1 (§2X5.1) of the United States Sentencing Guidelines generally provides that for offenses that are not covered by a direct guideline, judges should apply the most analogous guideline. If there is not an analogous guideline, judges should impose a fair sentence under the broad sentencing considerations that Congress has enumerated (i.e., sentence without guidelines).[17] One would hope for consistency in the decision of whether and how to analogize, but instead, for the same underlying crime, sometimes judges apply guidelines and sometimes they do not. Judges who apply guidelines often apply different guidelines in different cases to the same type of offense conduct. This inconsistency exists inter-district, intra-district, and sometimes with the same judge. The inconsistency can also lead to a difference of years of imprisonment. This Article proposes a clear, restated test based on appellate case law to hopefully bring consistency to the decision of whether and which guidelines apply.

But opportunity accompanied this inconsistency. Data from hundreds of §2X5.1 cases, compiled here for the first time, provide a type of natural experiment to analyze how judges sentence with and without guidelines—for the exact same offenses and type of offender. Debates have persisted for decades about the efficacy, wisdom, and complexity of the Guidelines,[18] but no prior analysis has compared actual cases sentenced under the guidelines to virtually identical cases sentenced without guidelines. At most, comparisons might analyze guideline sentences with pre-guideline sentences,[19] or ask judges how they might, hypothetically, sentence a real or fictional case.[20] This is of course qualitatively different than studying real cases, where before a judge stands an offender in a jumpsuit and leg irons (with whom the judge is already familiar based on presentence proceedings), a victim, grieving families on both sides, and two seasoned attorneys passionately advocating opposite positions.[21]

Detailed §2X5.1 sentencing data reveal an “analogous-guideline gap,” or substantial difference in sentences depending on whether a judge analogizes to a guideline. In almost every analyzed category, judges sentenced harsher when they sentenced without guidelines, sometimes ten to fifteen times harsher. If nothing else, the analogous-guideline gap reinforces the imperative of a uniform test for determining whether guidelines apply, as that single decision can have an enormous impact on the resulting sentence. The data also suggest much more unexplained variability in sentences when judges do not use guidelines—the magnitude of the difference between sentences was much greater when judges sentenced without guidelines. Even after adjusting for severity and criminal history, without guidelines some similarly situated offenders received probation or a short prison sentence and others received decades in prison for the same offense.

Beyond justice and fairness concerns, the variability shown in this Article risks delegitimizing federal judges, federal sentencing, and the courts as a whole.[22] Unwanted variability is not tolerated because it is “thought acceptable but because it ha[s] remained unnoticed.”[23] I argue that a primary reason for disparity in §2X5.1 cases is a lack of access to data in comparator cases, meaning parties are negotiating and judges are sentencing blindly.[24] This Article exposes the variability and recommends measures to reduce it.

Finally, use of and interest in § 2X5.1 is rapidly increasing. The Supreme Court’s recent decision in McGirt v. Oklahoma[25] has led to a significant increase in the number of these cases, and because most of these offenses occur on tribal lands, the disparity in sentencing can disproportionately affect tribal citizens. The guideline has also been invoked in many of the January 6, 2021, “Capitol Breach” cases because offenses like seditious conspiracy and civil disorder do not have direct guidelines.[26] Proposals to replace the complicated “categorical approach,”[27] which impacts thousands of defendants with prior convictions, have borrowed principles from §2X5.1, meaning §2X5.1 case law and sentences can serve as bellwethers for how an alternative categorical approach might look. Additionally, some of the Sentencing Commission’s proposed options for simplifying the Guidelines Manual predict that, if enacted, simplified guidelines would lead to an increased reliance on §2X5.1 because there would be many more offenses without guidelines.[28]

Part I summarizes the applicable authorities and clarifies confusion about the correct tests that apply in §2X5.1 cases. Part II presents and analyzes §2X5.1 sentencing data for the first time, revealing harsher sentences and unexplained variability when judges do not use guidelines.[29] Because Part II provides granular detail, Part III summarizes the findings for those readers who only desire an overview of the findings. Part IV offers pragmatic recommendations for judges, the Sentencing Commission, and the U.S. Probation Office to address the problems described in Parts I and II. Ultimately, §2X5.1 cases serve as a cautionary tale against fully unrestrained discretion in sentencing and show that guardrails and accessible data are imperative in criminal sentencing.

I. Background

A. The Federal Sentencing Guidelines

To contextualize how §2X5.1 sentences are failing to satisfy Congress’s stated purposes of sentencing, a brief overview of the guidelines is appropriate. At least as of the 1950s, “every state and the federal system used indeterminate sentencing.”[30] In an indeterminate system, judges had wide discretion to sentence within broad statutory ranges of punishment (for example, anywhere between probation and 20 years’ incarceration), and an executive-branch agency largely determined actual sentences by granting parole or good-time allowances.[31] In this system, a judge could sentence an offender to ten years in prison, only to have the offender released on parole after three or four years[32] (thus making the sentence “indeterminate”).

After seeing their sentences regularly shortened, judges began trying to predict the likelihood of parole or good-time credits and increasing sentences accordingly.[33] But of course, if contingencies like parole or good-time credit failed to obtain in the way the judge had anticipated, the offender would serve an unfairly long sentence.[34] Observers noted “demonstrably disparate treatment of similarly situated defendants” with hardly any oversight by appellate courts.[35] Several studies spread in the 1970s and 1980s, decrying disparities in sentencing and strongly criticizing judicial and parole discretion.[36] Of particular concern were possible disparities based on race and gender.[37]

The Sentencing Reform Act of 1984 and accompanying creation of the Sentencing Commission were “the most significant change to the federal criminal justice system in American history.”[38] The novel[39] approach created a bipartisan Commission and instructed it to create sentencing guidelines with the general aim of “avoiding unwarranted sentencing disparities among defendants with similar records who have been found guilty of similar criminal conduct.”[40] The SRA gave sentencing judges “factors to be considered in imposing a sentence” including characteristics of the offender and the offense.[41]

The Federal Sentencing Guidelines (Sentencing Guidelines or Guidelines), first effective in 1987,[42] essentially quantify criminal conduct and offenders’ criminal history, putting a numerical score on the offense (ultimately, the “total offense level,” a number between 1 and 43) and the offender’s criminal past (the “criminal-history category,” between 1 and 6).[43] Judges then refer to the Sentencing Table, a biaxial, 258-cell grid in the Guidelines, to determine the “guideline range” of punishment based on the offender’s criminal-history category and the total offense level.[44] After the landmark case United States v. Booker, the Sentencing Guidelines are advisory and not mandatory.[45] Judges nonetheless are required to accurately calculate the guideline range under the Sentencing Guidelines,[46] and studies have shown that the guideline ranges have a “gravitational pull” on judges’ sentences,[47] likely in part due to the reality that so long as a sentence is within a guideline range, it is “presumptively reasonable” and less vulnerable to attack on appeal.[48]

In contrast to wide statutory ranges of punishment for most crimes, if a guideline applies to a crime, the presumption is that every offender will be sentenced within 25% or 6 months of similar offenders who engaged in similar conduct.[49] The substantially narrower imprisonment range is intended to facilitate less unwarranted disparity and more transparency, predictability, and proportionality.[50]

The Commission faced significant challenges from the start. Congress stopped short of passing a new, comprehensive, and comprehensible criminal code. As a result, “the Commission was required to work with a disarray of conflicting, vague, inconsistent, and redundant criminal statutes that characterized the federal criminal law.”[51] The Sentencing Commission worked within the existing complexity of the Federal Criminal Code and created an express guideline for most federal offenses, but for various reasons discussed throughout this Article, many crimes exist for which there are no sentencing guidelines. These include certain federal crimes (including contempt, inciting a riot, seditious conspiracy, and drug trafficking on a vessel), and state crimes that are assimilated into the federal system (including child abuse or neglect, DUI, burglary, and indecent exposure). These assimilated state crimes make up the largest number of federal sentences without guidelines.[52]

B. The Assimilative Crimes Act

Unlike the present system, under-inclusiveness plagued criminal codes in early America. Joseph Story remarked as early as 1816 that because federal courts had no criminal jurisdiction under common law,[53] and because states lacked jurisdiction on federal lands and on the high seas, a whole host of crimes were “dispunishable” including rapes, arsons, batteries, and conspiracy to commit treason.[54] Rather than embark on the futile and impractical[55] attempt to account for every conceivable crime, Story’s pragmatic solution was to simply incorporate state law—which was far more comprehensive—into federal court if those state crimes were committed on federal lands or waters. He drafted the text of a bill, which was incorporated into a broader bill that Daniel Webster sponsored and that eventually passed on March 3, 1825.[56] Story called the assimilated-crimes section “the most important section of the whole bill.”[57]

The Assimilative Crimes Act of 1825 (ACA) provides that if conduct occurs within a federal enclave[58] that is not otherwise punishable by federal law but is punishable by a state law in the area where the conduct occurred, the government may prosecute the state law in federal court.[59] The ACA is a flexible, gap-filling provision: “[t]he ACA’s basic purpose is one of borrowing state law to fill gaps in the federal criminal law that applies on federal enclaves.”[60] The elements of the state law are incorporated or “assimilated” into the federal criminal code and a “like offense” to the state offense results.[61] After a federal court finds guilt in an ACA case, a person is “subject to a like punishment” to what they would have faced in a state prosecution.[62] Sentencing judges must identify the state minimum and maximum imprisonment range and then apply federal sentencing policy within that range.[63]

Like the ACA, the General Crimes Act (GCA)[64] extends federal laws (including the ACA) to tribal lands (with certain exceptions not relevant here), and the Major Crimes Act (MCA)[65] permits the federal government to prosecute tribal defendants in federal courts for certain “major” offenses committed on tribal lands that historically might have gone unpunished under tribal justice systems.[66]

Because assimilated crimes incorporate the elements of state laws into federal court, there are no direct sentencing guidelines for these offenses. But just as the federal criminal code utilizes the ACA as a gap filler where federal law is silent or would not apply, the Federal Sentencing Guidelines provide a gap filler to account for criminal conduct that the Guidelines do not directly cover: United States Sentencing Guideline (U.S.S.G.) §2X5.1.

C. Federal Sentencing Guideline §2X5.1

1. Background and Purpose

U.S.S.G. §2X5.1 “tell[s] the sentencing court how to go about determining a sentence for a felony for which there is no offense guideline.”[67] This guideline can be broken into two primary parts:

  1. If the offense is a felony for which no guideline expressly has been promulgated, apply the most analogous offense guideline.
  2. If there is not a sufficiently analogous guideline, the provisions of 18 U.S.C. § 3553 shall control.

Any guidelines and policy statements that can be applied meaningfully in the absence of a Chapter Two offense guideline shall remain applicable.[68]

The gap-filling guideline is necessary to address new crimes, rare crimes,[69] crimes that cover a wide swath of conduct,[70] or the myriad of state offenses that one would not expect the federal Sentencing Guidelines to cover.

To illustrate the judge’s decision under §2X5.1, consider the following fact pattern which is loosely based on an actual case. Two young mothers, each with a toddler in tow, meet at R Bar and Restaurant in Lemmon, South Dakota. The two friends take advantage of happy hour while they dine, each ultimately consuming two margaritas, a Corona, and a shot of tequila. As they leave R Bar in separate cars, they encounter a fork in the road a few miles east. Mother 1 proceeds north on US-49 and enters North Dakota, and Mother 2 continues east on US-12.

A few miles later, Mother 1 sideswipes a car that is stalled on the shoulder of US-49. Her toddler suffers minor bruises from the incident. Because the mother is a member of a tribe and the event occurred within the boundaries of the Standing Rock Reservation, the mother is charged in federal court with child endangerment under North Dakota law.[71] There is no federal sentencing guideline for child endangerment because there is no federal crime of child abuse, but the parties stipulate in a plea agreement that after applying §2X5.1, the most analogous guideline is U.S.S.G. §2A2.3(a)(2) (Assault), which has a base offense level of 4. After agreed enhancements, the child’s bodily injury (+4) and the child being a vulnerable victim (+2),[72] and a two-level reduction for the mother’s acceptance of responsibility (-2),[73] Mother 1’s total offense level is 8. Assuming a fictional criminal-history category (CHC) of 1, Mother 1’s guideline range would be 0 to 6 months of incarceration.[74] (Hereinafter, references to a sentence or a range in “months” means months of incarceration, not probation or other alternatives.) Mother 1 is sentenced to three years of probation.

A few miles after the fork in the other direction, Mother 2 also side swipes a vehicle on her way home and her toddler also suffers minor bruises. She is charged in federal court with child endangerment under South Dakota law. There is no plea agreement, no agreement on a sentence, or agreement on an analogous guideline. At sentencing, the judge concludes that there is no analogous guideline and refers only to the statutory range, anywhere from zero to fifteen years in prison.[75] Mother 2 (CHC-1) is sentenced to 30 months in prison.

Section 2X5.1 presents judges with their own fork in the road: whether to analogize to a guideline or not. The guideline road presents a relatively narrow range of punishment. Down the no-guideline road, wide statutory ranges of punishment await: 0–20 years for first-degree home invasion in Michigan,[76] 2–30 years for a fourth DWI in Louisiana,[77] or a $500 fine to life imprisonment for child neglect in Oklahoma.[78]

Given the gravity of the judge’s decision of which road to traverse, an overview of the approaches is a crucial starting point.

2. Determining “The Most Analogous Offense Guideline” Under §2X5.1

The first part of §2X5.1 instructs judges to determine the most analogous offense guideline.[79] Judges have long lamented the difficulty in doing so: the Eighth Circuit has characterized attempts to fashion sentences in these cases as trying “to determine which round hole best accommodates a square peg.”[80] The Third Circuit agreed, quoting this language and calling the §2X5.1 sentencing judge’s task “admittedly difficult.”[81] The Seventh Circuit has called §2X5.1 cases “by definition ‘unusual.’”[82]

Further difficulty likely stems from the wording of the guideline.[83] Commentators have described §2X5.1 as “somewhat contradictory”[84] because it seems to put the cart before the horse. The guideline first instructs judges to identify the “most analogous” guideline before it addresses what to do if there are not any “sufficiently analogous” guidelines. As set forth later in this Article, the clearer order would be to identify any sufficiently analogous guidelines and then choose the most analogous one from that group.

The following table shows the standards that circuit courts have applied to §2X5.1. My survey of reported cases from appellate courts shows that an “elements-based test” is the most common approach. As set forth in Section II.B below, a vast majority of §2X5.1 cases in the past five years occurred within elements-based-approach circuits, meaning these circuit and district courts have had more opportunities to hone the elements-based approach. This increased vetting arguably makes it preferable to existing alternatives.[85]

Table 1: Summary of § 2X5.1 Approaches

Elements-Based Approach

Other[86]

Third, Fifth, Seventh, Eighth, Ninth, Tenth, and Eleventh Circuits

First, Fourth, and Sixth Circuits

United States v. Betts, 99 F.4th 1048 (7th Cir. 2024); United States v. Jackson, 862 F.3d 365 (3d Cir. 2017); United States v. Saac, 632 F.3d 1203 (11th Cir. 2011); United States v. McEnry, 659 F.3d 893 (9th Cir. 2011); United States v. Rakes, 510 F.3d 1280 (10th Cir. 2007); United States v. Calbat, 266 F.3d 358 (5th Cir. 2001); United States v. Osborne, 164 F.3d 434 (8th Cir. 1999).

United States v. Brady, 168 F.3d 574, 577 (1st Cir. 1999) (analyzing whether analogy is “reasonable”); United States v. Hendrickson, 822 F.3d 812, 830 (6th Cir. 2016) (analyzing guidelines and conduct); United States v. Terry, 86 F.3d 353, 358 (4th Cir. 1996) (analyzing guidelines and conduct, limited to only conduct charged in indictment).

Because the elements-based approach is by far the most common, it will be the focus of this Article. The Eighth Circuit originated the key components of the test in 1999.[87] This Article restates that test as follows:

This is the procedure for identifying the most analogous guideline for offenses without guidelines, distilled from several cases and introducing simpler steps and vocabulary.

For example, consider a case of physical child abuse on the Navajo Reservation in Arizona. Prosecutors would charge the state offense of child abuse in Arizona federal court. This state offense has no federal sentencing guidelines, and §2X5.1 instructs judges to determine the most analogous guideline. One would put the elements of the Arizona child-abuse offense next to the elements of federal assault and see whether the two are plausibly analogous. For simplicity, the elements below are based on relevant jury instructions, but parties might disagree about the proper elements of certain offenses:

Table 2

Offense of conviction

Prospect Statute 1

Prospect Statute 2

A.R.S. § 13-3623(B)(2)

(Child Abuse)

18 U.S.C. § 113(a)(4) (Assault by Striking or Wounding)

18 U.S.C. § 113(a)(5) (Simple Assault of Person Under Age 16)

1. [under circumstances other than those likely to produce death or serious injury,]

2. [intentionally] [knowingly] [recklessly] [with criminal negligence],

3. [caused the (child) (vulnerable adult) to suffer physical injury.] [caused the (child) (vulnerable adult) to suffer abuse.] [caused or permitted the person or health of the (child) (vulnerable adult) to be injured, while having the care or custody of the (child) (vulnerable adult).] [caused or permitted the (child) (vulnerable adult) to be placed in a situation where the person or health of the (child) (vulnerable adult) was endangered, while having the care or custody of the (child) (vulnerable adult).][94]

First, the defendant assaulted [name of victim] by intentionally [[striking] [wounding]] [[him]] [her]]; and

Second, the assault took place on [specify place of federal jurisdiction].[95]

First, the defendant assaulted [name of victim] by intentionally using a display of force that reasonably caused [him] [her] to fear immediate bodily harm;

Second, [name of victim] was under the age of 16 years at the time of the assault; and

Third, the assault took place on [specify place of federal jurisdiction].[96]

The above table is a non-exhaustive list of prospect statutes used only for demonstration purposes and more prospects could exist. The elements of the above federal statutes are plausibly analogous to the state offense in physical abuse cases,[97] so the federal guidelines associated with these assault offenses are candidates for most analogous guidelines. Both 18 U.S.C. § 113(a)(4) and § 113(a)(5) are covered by U.S.S.G. §2A2.3, making that guideline sufficiently analogous. Because (in this truncated example) it is the only sufficiently analogous guideline, it is the most analogous guideline.

Part I has hopefully clarified the correct application of the complicated § 2X5.1 Guideline. As explained below, the varying tests that have applied in § 2X5.1 cases, the inconsistent application of the same test (the elements-based approach), and inconsistent plea agreements have resulted in frequent “crossover” offenses, or offenses for which judges sometimes apply guidelines and sometimes judges do not. These offenses provide a vehicle to analyze guideline and no-guideline sentences in real cases.

II. U.S.S.G. §2X5.1 Sentencing Data

A. Methodology and Dataset

When processing the data from §2X5.1 sentencing documents, the Sentencing Commission does not code these sentences in the same way that it codes others.[98] Based on the current coding,[99] these cases are not accurately tracked in the Sourcebook or other Sentencing Commission reports,[100] nor are they searchable on the Sentencing Commission’s public databases, IDA and JSIN. In many ways, these are “ghost” cases that have fallen through the cracks. Among other problems, this means that parties and judges cannot look to these prior cases and compare with a case before them. I make recommendations to remedy this problem in Part IV, but for this Article I manually compiled, coded, and reviewed each case in the dataset.

Using Sentencing Commission data from fiscal years 2016–2021,[101] I extracted information from the five key sentencing documents that federal law requires the chief judge of each district to send to the Sentencing Commission: the judgement and commitment order (J&C), the statement of reasons (SOR), any plea agreement, the indictment or other charging document, and the presentence investigation report (PSR).[102]

I compiled the dataset in two stages, first searching by statute and then by guideline. I began by compiling federal sentences within the review period that cited the ACA, GCA, or MCA as a statute of conviction.[103] Class A misdemeanors were excluded because a direct guideline covers those offenses.[104] I narrowed the results to any case in which §2X5.1 controlled the sentence,[105] and then manually coded each state statute from the J&C so that cases could be analyzed by state statute as needed.

A second query consisted of any case that cited §2X5.1 in any way in imposing the sentence.[106] These results captured any §2X5.1 sentence that was not an assimilated crime,[107] or for which the sentencing documents did not cite the ACA, GCA, or MCA (perhaps due to local practice or a scrivener’s error) but still invoked §2X5.1. I then combined and deduped the two datasets.[108] I excluded six cases that were reversed on appeal and were not resentenced under §2X5.1 within the review period.[109] The resulting dataset consisted of 434 cases that were sentenced within the review period.

I grouped §2X5.1 cases into two categories depending on whether the underlying offense was state or federal:

  1. “Assimilated”: convictions assimilating state law and sentenced under §2X5.1.
  2. “Fed. Stat.”: convictions under other federal criminal statutes and sentenced under §2X5.1.

I separated cases into these two categories because the offenses in each category are qualitatively different: cases in the first category are all state offenses that no federal law criminalizes (and, being state offenses, have no direct federal sentencing guidelines). The second category includes offenses that are criminalized under federal law but do not have a direct sentencing guideline. The conduct in the two categories should never overlap. Unless there is a “substantial difference” between the kind of behavior covered by a federal law and a state law, the prosecutor should only charge the federal offenses—not the state offense.[110]

I then grouped cases into two subcategories within the two above categories based on whether the judge analogized to a guideline:

  1. “AG”: cases in which the sentencing judge applied an analogous guideline to the offense (thereby sentencing the offender under the Guidelines).
  2. “NAG”: cases in which the sentencing judge determined that no analogous guideline covered the offense (thereby sentencing without direct guidelines).

A majority of the sentences for federal offenses that analogized to a guideline (Fed. Stat. AG cases) were for convictions under 46 U.S.C. § 70503, maritime drug smuggling offenses that were uniformly analogized to §2D1.1 (Drug Offenses) (N=103).[111] Because the purpose of the analysis is to compare offenses sentenced with and without guidelines and these drug offenses were all sentenced under the same guideline, the offenses were not within the purview of the analysis and were excluded.[112]

Because many of the filings relied upon in this Article are confidential and remain under seal, offender names, judge names, and case numbers are not cited. Where fact patterns or details are referenced, they are sourced from unsealed or otherwise publicly available filings like court opinions, charging documents, J&C’s, and plea agreements that were not filed under seal.[113]

B. Most Common Jurisdictions and Types of Cases

Most §2X5.1 cases occurred within the Ninth (N=71), Tenth (N=61), Fifth (N=46), and Eighth (N=59) Circuits:[114] The states[115] or territories with the most §2X5.1 offenses during the review period are set forth below. As one might expect, these jurisdictions typically have either large tribal reservations, military bases, or national parks:[116]

Table 3: Top §2X5.1 States/Territories

Assimilated: NAG

#

Fed. Stat.: NAG

#

New Mexico

22

D.C.

4

South Dakota

17

Virginia

4

Oregon

15

Texas

3

Oklahoma

12

Alabama

2

Texas

10

Florida

2

Assimilated: AG

#

Fed. Stat.: AG

#

Arizona

29

Florida

9

South Dakota

14

Texas

7

North Dakota

12

New York

7

Michigan

10

Colorado

6

Texas

10

Puerto Rico

4

Washington

10

    

As broken down more fully below, the most common assimilated NAG offense was state child abuse. Other common assimilated NAG offenses included felony DUI[117] (typically fourth or fifth offense) and felony eluding. The most assimilated AG offense in these top jurisdictions was, interestingly, also child abuse. The other commonly assimilated AG offenses were burglary or theft.

The most common Fed. Stat. NAG offenses were piloting an aircraft without a pilot’s license,[118] offenses related to failure to register as a foreign agent,[119] contempt,[120] and unauthorized removal of classified documents or materials.[121] The most common Fed. Stat. AG offenses were piloting an aircraft without a valid pilot’s license,[122] bankruptcy fraud,[123] and civil disorder[124] or inciting a riot.[125] (analogized guidelines described in Section II.D, infra). With drug cases removed, there were relatively few Fed. Stat. §2X5.1 cases (N=95)—there were over twice as many assimilated cases (N=236).

C. Convictions for Assimilated State Crimes

Assimilated state crimes accounted for 71% of non-drug §2X5.1 offenses,[126] and the analysis revealed several concerning issues. Particularly, judges are in disarray in child abuse and felony DUI cases, and unwarranted sentencing disparities appear to exist. The Sentencing Commission’s Tribal Issues Advisory Group identified § 2X5.1 child-abuse offenses as its top priority for the Sentencing Commission to address.[127] The data confirm this priority. Child abuse was the most common § 2X5.1 offense, and all but 6 child-abuse cases occurred on tribal land, meaning this issue and any associated disparity disproportionately impacts tribal citizens (offender, victim, and public).

After child abuse (N=92), the next most common assimilated cases were burglary (N=30), DUI (N=28), and eluding (N=28). The Article focuses on crossover offenses, offenses that were sometimes sentenced under guidelines and other times not. All assimilated burglary cases were sentenced under the Guidelines (analogizing to § 2B2.1 (Burglary)), and 15 of the 16 eluding NAG cases occurred in Oregon, meaning little interdistrict analysis of crossover cases could be done. Thus, for assimilated offenses this Article focuses on child abuse and DUI offenses.

1. Child Abuse

There were 92 total child-abuse sentences during the review period. Comparing sentences in these cases required additional subcategories. For typical federal offenses sentenced under the Guidelines, to compare like offenses and offenders, one can compile all sentences imposed under a particular statute, filter by guideline, total offense level, and/or criminal history and then observe the results. This is not possible in §2X5.1 cases because (1) the data were not maintained in a readily useable format,[128] (2) there are multiple, conflicting state statutes to reference, and (3) in almost half of the child-abuse cases (N=47), judges did not use guidelines.

I separated the cases by state statute, identified whether the judge analogized to a guideline, and when judges applied guidelines, included which guideline the judge applied (Appendix A). The first question Appendix A elicits is why, for the same offense (child abuse) did some judges conclude that the offense has an analogous guideline and some judges conclude that it did not? All courts in those jurisdictions apply the elements-based approach, except Michigan, North Carolina, and Wisconsin (together accounting for only 5 of the 92 child-abuse cases). Potentially applying different tests[129] for whether to analogize did not appear to make a meaningful difference on outcome (although the number of cases from non-elements-based jurisdictions is small).

Arizona, Nebraska, New Mexico, Oklahoma, and South Dakota are particularly confounding: how are the same statutes within the same states considered both AG and NAG in federal court? In some instances, the answer is clear: the statute is broad and conduct varies widely. In the single Nebraska NAG case, a mother was intoxicated with her unrestrained children in the car, fled from police, and wrecked her car. In the Nebraska AG cases, two adults admitted to locking a child in a storage room and then becoming intoxicated and forgetting to check on the child. Both undoubtedly constitute egregious behavior, but one can see how a judge might conclude that different guidelines (or no guideline) might apply in these very different cases—even though all were convicted of child abuse. On the other hand, similar conduct (physical abuse/beating of a child) was considered both a NAG and an AG in Oklahoma, New Mexico, and South Dakota. As these cases demonstrate, child-abuse cases cover a wide array of unlawful conduct, and making all of the cases fit within the Federal Sentencing Guidelines reintroduces the analogy of a square peg in a round hole.[130]

Conflicting state statutes add further complexity when trying to assess horizontal disparity, or disparity across federal courts. Some state statutes separate the conduct into categories and charge child-abuse offenses accordingly. For example, Nebraska Revised Statute § 28-707(1) provides in relevant part:

  1. A person commits child abuse if he or she knowingly, intentionally, or negligently causes or permits a minor child to be:
    1. Placed in a situation that endangers his or her life or physical or mental health;
    2. Cruelly confined or cruelly punished; [or]
    3. Deprived of necessary food, clothing, shelter, or care.[131]

Quite clearly, subsection (a) constitutes child endangerment, (b) cruel confinement or physical child abuse, and (c) child neglect. The single Nebraska NAG case (intoxicated and eluding arrest with unrestrained children in the car) was charged under subsection (1)(a), and the two Nebraska AG cases (locking children in a storage room) were charged under subsection (1)(b). Arizona Revised Statute § 13-3623 similarly provides different subsections depending on whether the act was done “intentionally or knowingly,” “recklessly,” or “with criminal negligence.”[132] But even these statutes do not completely align: Arizona has different subsections depending on culpability, and Nebraska combines the states of mind and has subsections for different acts.[133]

Other state statutes conflate all of these categories. For example, South Dakota Codified Laws § 26-10-1 simply states in relevant part: “Any person who abuses, exposes, tortures, torments, or cruelly punishes a minor in a manner which does not constitute aggravated assault, is guilty of a Class 4 felony.”[134] Thus, there is no subsection to charge or analyze depending on the specific conduct—all offenses are lumped into one broad “child abuse” statute. North Dakota has one statute for abuse[135] and another for neglect,[136] but none specifically for endangerment.[137]

Because these cases are infrequent in many federal districts, judges often cannot lean on prior practice within a home jurisdiction, naturally leading them to other states for guidance.[138] But these varying, incompatible state statutory regimes for child-abuse offenses undoubtedly cause confusion or make it improbable that a judge would be able to compare like cases.

After thorough review of all charging documents, plea agreements, PSR’s, J&C’s, and SOR’s in these cases, recurring themes emerged in the offenders’ conduct. To allow for more meaningful analysis of similar conduct, I separated cases into three categories based on the offense conduct recounted in the PSR,[139] regardless of the statutory language: (1) physical child abuse, (2) child endangerment, or (3) child neglect.

“Physical child abuse” cases involved injury based on an affirmative act of intentional, abusive, physical contact with the child (often beating, excessive spanking, shaking, or intentionally dropping). “Child endangerment” cases involved behavior that was considered reckless or invited a substantial risk of harm, sometimes (but not always) leading to injury. Most of these cases involved DUI, sometimes involving automobile accidents with children in the car.[140] “Child neglect” cases involved omissions or neglect, but not intentional physical abuse. These cases typically included caretakers becoming intoxicated and neglecting to care for a child, or generally failing to provide adequate health or welfare for the child. These categories get us much closer to comparing similar offense conduct.

Analyzing the dataset using these categories, there were 41 cases of “physical child abuse,” 43 cases of “child endangerment,” and 8 cases of “child neglect.” Of the 41 physical child-abuse cases, judges analogized to guidelines more often: 27 were AG cases, with only 14 NAG cases. Almost the opposite is true of child-endangerment cases: judges analogized in 14 cases and found no analogous guidelines in 29 cases. Child-neglect cases were evenly split, with 4 NAG cases and 4 AG cases. Judges clearly have an easier time finding analogous guidelines in physical child-abuse cases, but not in endangerment or neglect cases. And this makes sense, as explained in the next sections.

a. Physical Child Abuse

As defined above, physical child abuse requires an intentional act. It is therefore often conceptually easier to analogize such behavior to assault, which is covered by sentencing guidelines. In fact, in every AG case of physical child abuse in the dataset (27 of 27 cases), judges analogized to either the guideline for assault (§2A2.3) or aggravated assault (§2A2.2). In the remaining 14 NAG cases, each of these involved intentional, physical abuse, but the judges determined that there was no analogous guideline and imposed a sentence solely under 18 U.S.C. § 3553(a).[141]

Disparity exists based on whether a judge decides to analogize these cases to a sentencing guideline. Put another way, had the judges applied guidelines in these NAG cases (or not applied guidelines in AG cases), it almost certainly would have impacted the sentences. The following table summarizes sentences in physical-abuse cases:

Table 4: Sentences for Physical Child Abuse

 

NAG

AG

NAG Difference

Average

46.9 months

23.3 months

101%

Median

24 months

12 months

100%

As this table shows, sentences in physical child-abuse cases were twice as long when judges did not sentence under a guideline and instead sentenced without guidelines.

Probation or time served were imposed in 7 AG cases (26%) and 5 NAG cases (36%). The following table excludes those cases and summarizes only sentences in which the judge imposed a term of imprisonment (beyond time served):

Table 5: Sentences for Physical Child Abuse

(Probation Excluded)

 

NAG

AG

NAG Difference

Average

72.9 months

28.6 months

155%

Median

60 months

12.5 months

380%

Deciding whether to analogize to a guideline can have a profound impact on the final sentence. Excluding probation, the median sentence in a physical-child abuse case without guidelines is several times longer than when a guideline applies. This sentencing differential between guideline and no guideline sentences for similar offenses is what I call the “analogous-guideline gap”—a gap created depending on whether a judge analogizes to a guideline.[142]

Comparing guideline and no-guideline sentences is not the only way to measure disparity. Comparing sentences to the bottom of the guideline range can also reveal disparity because most federal sentences are at or below the bottom end of the guideline range[143] (meaning the bottom of the guideline range is sometimes called the “presumptive sentence”).[144] The Sentencing Commission has found that guideline minimums have a “gravitational pull” on sentences, supported by the fact that when judges sentence within the Guidelines, they impose the guideline minimum 58.4% of the time, and when judges depart or vary from the Guidelines, they are nearly 21 times more likely to do so downward rather than upward.[145] Using the bottom of the applicable guideline is sometimes “considered to be a superior method of capturing the many individual guidelines calculations that are part of a typical federal sentencing calculus—as opposed to including each individual guideline calculation issue as a separate independent variable.”[146] Comparing child-abuse sentences to guideline minimums can reveal whether judges are softer, harsher, or in line with sentences in most other cases. We can measure this for both AG and NAG cases, which serve as a window into pure judicial discretion (NAG cases) and the Guidelines’ application to these offenses (AG cases).

To compare sentences in physical-abuse cases to the bottom of the guideline range, I began by calculating the most frequent guideline ranges in AG cases for these offenses. In physical-abuse AG cases, judges analogized to assault guidelines in every case (aggravated assault 14 times, and assault 15 times), making those the reference guidelines (§2A2.2 and §2A2.3). The average criminal-history category was between 1 and 2 (1.7 in NAG cases, 1.2 in AG cases), so I used both CHC-1 and CHC-2 to calculate the guideline range in the most common cases. The resulting guideline ranges are set forth below, and the guideline minimums in those ranges become the reference points. We can then compare those guideline minimums to averages from actual sentences imposed to observe how judges sentenced in relation to these minimums.

The average sentence in AG cases was 8.9 months when judges analogized to assault, and 36.7 months when judges analogized to aggravated assault. The average NAG sentence was 46.9 months. Because we do not know whether judges would have analogized the physical-abuse NAG cases to assault or aggravated assault,[147] both possibilities are presented below, as well as the average difference between the actual sentences and the guideline minimums.

Table 6: Physical Child Abuse: Difference Between Guideline Minimum, AG, and NAG Sentences (Assault)[148]

 

Guideline Minimum

AG Average Difference

NAG Average Difference

Assault: CHC-1

0 months

+8.9

+46.9

Assault: CHC-2

4 months

+4.9

+42.9

Table 7: Physical Child Abuse: Difference Between Guideline Minimum, AG, and NAG Sentences (Aggravated Assault)[149]

 

Guideline Minimum

AG Average Difference

NAG Average Difference

Aggravated Assault: CHC-1

21 months

+15.7

+25.9

Aggravated Assault: CHC-2

24 months

+12.7

+22.9

As these tables show, sentences in these cases substantially exceed the guideline minimum or “presumptive” sentence,[150] particularly in NAG cases (sentenced approximately two years longer than the presumptive sentence, even assuming the judge analogized to the harsher aggravated-assault guideline in all cases). To control for different variables, contextualize the data, and address outliers, the cases are further broken down below.[151]

i. Child-death cases

Some of the harshest sentences can occur when the child dies as a result of the abuse. So that the reader can appreciate whether these cases might skew the results, they (and all notable outliers) are briefly discussed in each child-abuse section. These cases also expose some of the most extreme instances of disparity when judges do not apply guidelines.[152] As will also be shown, comparing child-death cases to child-injury cases reveals that child-injury cases that are sentenced without guidelines are often sentenced harsher than child-death cases sentenced with guidelines.

With guidelines. There were 2 AG child-death cases within the physical abuse category, both involving the same intoxicated parents who were physically fighting while the mother, holding the baby, used the child as a shield. The child subsequently died as a result of injuries incurred during the fight. The mother (CHC-1; total offense level (TOL) of 18) received a sentence of 30 months, and the father (CHC-1; TOL 22) received 48 months. In both cases, the plea agreement was silent on a guideline or range and the judge analogized the case to aggravated assault, §2A2.2.[153]

Without guidelines. There were 3 physical-abuse NAG cases involving death within the review period. A baby was found dead one morning and the father (CHC-2) later admitted to a history of abuse and dropping the child multiple times (though he did not admit to being the direct cause of the child’s death). He received a sentence of 24 months. Prosecutors charged both parents in another case with the death of a child where there was evidence of a history of physical abuse, including cigarette burns. The father (CHC-1) pleaded guilty to negligently putting his child at risk, not physical abuse causing death, and was sentenced to 180 months, and the mother (CHC-1) was convicted at trial and sentenced to 180 months.

These NAG cases reveal probable unwarranted disparities. In the two NAG cases of the fathers pleading guilty, both cases involved the death of a child and admissions of negligently putting the child at risk (the child’s dangerous mother). But one was sentenced to 24 months and the other to 180 months.[154] More puzzling, the father sentenced to 24 months admitted to a history of physical abuse, while the father receiving 180 months did not. Both had plea agreements, and neither stipulated to a sentence or guideline range. Finally, the father sentenced to 24 months had a higher criminal-history category (2 vs. 1).

The prior cases (24 months vs. 180 months) were both NAG cases and reveal possible disparity when judges do not have guidelines. But comparing AG and NAG cases also presents a potential analogous-guideline gap, or disparity based solely on whether the judge analogizes to a guideline. For example, compare the sentence in the NAG case of the father who received 180 months to that of the father who received 48 months (CHC-1; TOL 22) in the AG child-death case involving the fighting parents. In the AG case, the judge analogized the case to aggravated assault, §2A2.2. Had the judge analogized the 180-month NAG case to aggravated assault, the offender’s total offense level would have probably been around 18,[155] resulting in a guideline range of 27–33 months’ incarceration. Even under the guideline for the much more severe offense of voluntary manslaughter, his base offense level would have been 29.[156] After reducing three levels for acceptance of responsibility,[157] his guideline range would have been 63–78 months, less than half of his actual sentence. In effect, the judge sentenced this offender closer to a conviction for second degree murder, while being convicted of abuse and admitting only to negligently putting the child at risk.[158]

ii. Adjusting physical abuse cases for severity

An important consideration when analyzing potential disparity is factoring in the severity of the offense when possible.[159] This is essential when comparing conduct in these physical abuse cases because of the broad spectrum of conduct that constitutes physical abuse. Because AG cases are sentenced within a guideline, severity is taken into account (incorporated into the total offense level), but NAG cases lack that benefit. For a truly meaningful comparison, the severity of the offense must be measured in some way.[160]

Statutory maximums are sometimes used to measure severity but serve as a poor proxy for severity in assimilated cases due to the wide ranges of statutory punishments for similar conduct across states.[161] Guideline minimums are also sometimes used.[162] However, because NAG cases have no guideline minimums, this measure is unavailable.

I addressed this issue by creating a severity scale that captured and categorized types of conduct in these cases. It was loosely inspired by the specific offense characteristics in the aggravated assault guideline, §2A2.2(b).[163] The severity scale is consistent with prior approaches[164] including the Minnesota approach[165] in its no-guideline cases (which may have inspired §2X5.1 in 1985[166]) and takes minor cues from the Department of Justice’s National Survey of Crime Severity.[167]

Physical-abuse cases were given a severity rating based on the following:

  1. Low-risk or unspecified[168] conduct resulting in no physical injury.
  2. Striking, biting, dropping, or similar acts causing minor bodily injuries.
  3. Striking, biting, dropping, or similar acts causing broken bones or other serious bodily injuries.
  4. Choking, use of a dangerous weapon, or other serious conduct causing or threatening bodily harm.
  5. Abuse causing life-threatening injuries.
  6. Abuse causing death.[169]

Most physical-abuse sentences fell neatly within one of these six categories. While no measure of severity is perfect, this framework permits a much more accurate comparison of like offenses. The few cases with sex-related conduct were excluded, as they are too incompatible with the conduct in the other cases. The results are presented in the table below, which also controls for the type of offender (via criminal history categories).

Beginning with cases in which judges analogized to a guideline:

Table 8: Summary of Physical Abuse Sentences Adjusted for Severity (AG)

Severity

CHC

Number of Cases

Range

Average Imprisonment

Median Imprisonment

1

1

1

Probation

Probation

Probation

1

2

1

4.5

4.5

4.5

2

1

6

Probation–24

11.2

11

2

2

2

12

12

12

2

4

1

11

11

11

3

1

2

Probation–20.6

8.8

5.7

4

1

2

12–13

12.5

12.5

4

2

1

Probation

Probation

Probation[170]

5

1

6

11.2–180

59.5

33

5

2

1

35.5

35.5

35.5

6

1

2

30–48

39

39

For analogous-guidelines sentences, the severity of the sentence largely aligned with the severity of the offense conduct. As the severity and criminal history increases, the sentence increased relatively consistently.

Under the same parameters, the same was not true in no-guideline cases:

Table 9: Summary of Physical Abuse Sentences Adjusted for Severity (NAG)

Severity

CHC

Number of Cases

Range

Average Imprisonment

Median Imprisonment

2

1

2

60

60

60

2

2

1

30

30

30

2

4

1

24

24

24

4

1

1

84

84

84

4

3

1

180

180

180

5

1

2

Probation–14.2

7.1

7.1

5

3

1

Probation

Probation

Probation[171]

6

1

2

24–180

102

102

6

2

1

180

180

180

This can be analyzed visually in the following graph:

Figure 1: Physical Child Abuse Adjusted for Severity

Comparing AG and NAG cases after controlling for severity suggests both an analogous-guideline gap (the gap between the two lines) and disparate sentencing. Any gap that exceeds 25% may violate Congress’s expectation that similarly situated offenders committing similar crimes should have sentences within 25% of each other.[172] Put plainly, while AG sentences tend to gradually increase with severity and criminal history, NAG sentences are all over the place, and tend to be much harsher. The ranges of punishment are far more extreme in NAG cases.

In the NAG cases, judges sentence offenders to longer terms of imprisonment for level 2 offenses than for many higher-level offenses (and all AG offenses), even after controlling for criminal-history category. Judges sentenced probation and seven months for level 5 offenses in NAG cases, but up to 60 months for level 2 offenses. In virtually all cases, judges sentenced no-guideline cases more harshly by some multiple of guideline sentences of the same severity. And when they did not (level 5 offenses), there is an argument that those sentences should have been longer, to the extent that parity with AG sentences is desirable.

If the measure of disparity involves an analogous-guideline gap, that is present in these cases. The following table (which, like all data in this Article, can serve as a reference point in future cases) shows the extent of the analogous-guideline gap among cases with the same severity and CHC:

Table 10: Comparing Average Physical Abuse Sentences

Adjusted for Severity

Severity

CHC

AG

NAG

% Difference

1

1

Probation

N/A

–

1

2

4.5

N/A

–

2

1

11.2

60

436%

2

2

12

30

150%

2

4

11

24

118%

3

1

8.8

N/A

–

4

1

12.5

84

572%

4

2

Probation

N/A

–

4

3

N/A

180

–

5

1

59.5

7.1

(743%)

5

2

35.5

Probation

–

5

3

N/A

Probation

–

6

1

39

102

162%

6

2

N/A

180

–

A final measure of disparity is to identify the average severity and criminal histories and compare AG and NAG sentences near those averages. The average severity and criminal-history category in NAG cases were slightly higher (4 and CHC-2) than in AG cases (3.2 and CHC-1.6). But even generously comparing sentences that were imposed near these averages reveals disparity. The sentences near these averages in AG cases averaged 8.8 months, and no sentence at a severity level of 3 or below exceeded an average of 12 months (regardless of CHC). See Table 6. On the other hand, NAG sentences at the lower severity level of 2 were 24 months (CHC-4), 30 months (CHC-2), and 60 months (CHC-1)—all multiples of the average AG case of like severity and CHC. See Table 7. In summary, the average NAG sentence is far longer than the average AG sentence, even after adjusting for severity.

Of course, this is a relatively small number of cases and one should be cautious making broader conclusions. But this is every known case within a five-year period, so we can at least conclude that the data suggest unwarranted disparity in these cases. If one’s measure of disparity includes the range of punishment for like offenses and offenders, disparity currently exists (particularly in NAG cases). Disparity is present if the measure considers the deviation from the guideline minimum (the most common sentence), or from Congress’s expectation of like offenses and offenders being sentenced within 25% or 6 months.[173] If one’s measure of unwarranted disparity includes asymmetry between sentence, severity, and criminal history, it is also present.

It is important to note at this point that disparity is not solely attributable to pure judicial discretion; plea agreements played a substantial role (although of course, the sentencing judge ultimately has discretion[174] to accept or reject all or part of a plea agreement). For example, in 93% of AG sentences in child-endangerment cases, the parties agreed on either the exact sentence, guideline, or maximum sentence. In 63% of physical-abuse AG cases, plea agreements provided the exact analogous guideline, and in all of those cases the judge ultimately applied the agreed guideline. In 63% of all child-neglect cases, the parties agreed to either an exact sentence or narrow range, and in each the judge imposed a sentence consistent with the agreement. Unpacking all of the granular data from plea agreements is beyond the scope of this Article,[175] but it is enough for present purposes to note that the parties (who are often negotiating blindly due to the lack of access to data in these cases)[176] agree to many of these sentences.

b. Child endangerment

Child endangerment cases involve behavior that is reckless or invites a substantial risk of harm but does not involve an affirmative act of intentional, physical abuse. There were 43 cases of child endangerment. In contrast with physical child-abuse cases, sentences (at first) appear longer when judges analogize to a guideline in child endangerment cases:

Table 11: Sentences for Child Endangerment

 

NAG

AG

NAG Difference

Average

8.7 months

16.1 months

(86%)

Median

Probation

3 months

–

Probation or time served were imposed in a larger percentage of these cases than physical-abuse cases: 8 AG cases (57%) and 16 NAG cases (55%). This is because these cases typically involve less severe conduct, like DUI with a child in the car (often without accident) or reckless but not intentionally abusive behavior. The following table excludes probation and time-served sentences and summarizes only sentences in which the judge imposed a term of imprisonment beyond time served.

Table 12: Sentences for Child Endangerment

(Probation excluded)

 

NAG

AG

NAG Difference

Average

18.7 months

35.7 months

(91%)

Median

18 months

24 months

(33%)

While it appears that sentences are harsher under the guidelines for child-endangerment cases, a significant outlier case, discussed below, provides additional context.

i. Child-death cases

One case stands as a clear outlier among the other child-endangerment cases. A baby died in his father’s custody, but the cause was unclear. The baby spent time with his mother and father, separately. The records do not reflect evidence that the father’s direct actions killed the baby, but he pleaded guilty to child abuse based on his failure to protect his child. He admitted to knowing the unsafe conditions that the baby lived in when the baby was in his mother and her boyfriend’s care, including physical abuse and intoxication. After analogizing to the assault guideline (§2A2.3), the father (CHC-3) had a guideline range of 12–18 months, but the nonbinding plea agreement agreed that the parties would recommend a sentence of 120 months, which the judge imposed.

Similar child-death cases within the child endangerment category resulted in much lower sentences (a 33-month sentence in an AG case that analogized to aggravated assault, and 36-months sentence in a NAG case). There is presumably some reason outside of the public record and sentencing documents warranting this disproportionately long sentence (for example, evidence of the offender’s direct involvement in the child’s death).[177] The next highest sentence in child-endangerment AG cases is 33 months. Or the parties (and ultimately, the judge) were not aware that a 120-month sentence would be a substantial outlier, absent the existence of more severe conduct.

It is important to highlight that in this outlier case, the parties agreed to a sentence that is ten times higher than the guideline minimum, and the judge almost certainly analogized to the wrong guideline for the actual offense conduct. In fact, to reach the 120-month sentence imposed within the Guidelines and without a substantial upward departure, given the offender’s criminal-history category of 3, he would need a total offense level of 30 under the Guidelines. This implicates much more severe guidelines, like §2A1.3 (Voluntary Manslaughter, BOL 29) or §2A2.1 (Assault with Intent to Commit Murder; Attempted Murder, BOL 33). Effectively, this offender was sentenced as if he were convicted of one of these more serious, intentional offenses, not simply endangering the life of his child or simple assault. As explained later, analogizing to the wrong guideline gives a false impression of the sentence for future litigants and judges seeking comparators, and can skew the data coming to the Sentencing Commission.

In total, four of the AG child-endangerment cases (29%) and five of the NAG cases (17%) involved the death of a child. Excluding the outlier discussed above, the average sentence in the 3 remaining child-death AG cases was 11 months, with a median sentence of probation.[178] In the 5 NAG cases, the average sentence was 11.4 months, with a median of 12 months. In general, the culpable parent in a child-death case typically received around a year of imprisonment when the parent’s behavior amounted to child endangerment.[179] In sum, other than the outlier case, child-death cases do not appear to meaningfully impact the aggregate data in child-endangerment cases.

ii. Impaired driving

The most common variation of child-endangerment offenses involved intoxicated driving with children in the car (N=19) (44%), and most of the time judges did not analogize to a guideline in those cases (79%).[180] When judges did analogize, each offender received a similar sentence: two offenders (CHC-1) received probation, one (CHC-3) received 8 months’ time-served, and one (CHC-1) received two days incarceration and 60 days of home confinement. When judges did not analogize, the average sentence in these cases was 6 months, with the median sentence of probation. Most of these offenders received probation (11 of 15; 73%), and where imprisonment was imposed, the offenders either had a high criminal-history category (3 and 6), or the incident involved a crash that injured the child.

Unlike physical-abuse cases, because the offense behavior varied so significantly and no consistent patterns emerged,[181] endangerment and neglect cases were not adjusted for severity. Additionally, the varying conduct in these cases means that it is less likely that they can properly be analogized to a guideline under the elements-based approach. In general, other than the sentences discussed above, judges typically imposed a sentence of imprisonment for knowledge of egregious abuse and failure to protect, discharging a firearm near a child, or other uncommon and extreme behavior, but otherwise imposed probation in child-endangerment cases (probation sentenced in 51% of all endangerment cases).

c. Child neglect

Child neglect cases in the dataset involved omissions or neglect, but not intentional, physical abuse. There were 8 cases of child neglect, virtually all with relatively high sentences:

Table 13: Sentences for Child Neglect

 

NAG

AG

NAG Difference

Average

54 months

51 months

6%

Median

48 months

54 months

(12%)

Although the population is much smaller in these cases, there appears to be a smaller analogous-guideline gap. Judges sentenced probation in only one of these cases (a NAG case in which a father left a baby unattended in a car outside of a liquor store while he got intoxicated). Removing that probationary sentence resulted in the following:

Table 14: Sentences for Child Neglect

(Probation Excluded)

 

NAG

AG

NAG Difference

Average

72 months

51 months

41%

Median

84 months

54 months

56%

This Article is hesitant to make sweeping conclusions based on such a small number of cases.

But offering data for the first time should help interested parties in the future compare similar sentences for similar conduct. Most of these cases involve severe cases of malnutrition, untreated obvious medical needs, long periods of neglect, and similar conduct.[182]

2. Felony DUI

Judges considered most DUI cases to have no analogous guidelines (N=24), but importantly, all except 1 of these NAG cases involved straightforward DUI and no injury to a victim.[183] Most cases involved intoxicated, non-military offenders who inexplicably found themselves at the entrances to military bases. Some tried to flee, some hit guardrails, but other than the single case described in Note 182, no victims were injured.

In each of the six DUI AG cases, the offenders were involved in accidents, sometimes fatally killing victims. Sentences in these cases ranged from 1 to 40 months, and there were no sentences of probation. Judges analogized the cases to either assault (§2A2.3), aggravated assault (§2A2.2), or involuntary manslaughter (§2A1.4).[184]

Table 15: Sentences for DUI

 

NAG

AG

NAG Difference

Average

21.8 months

15.8 months

37%

Median

12 months

15 months

(25%)

Table 16: Sentences for DUI

(Probation Excluded)

 

NAG

AG

NAG Difference

Average

34.1 months[185]

15.8 months

116%

Median

21 months

15 months

40%

Sentences between all AG and NAG DUI cases were similar in the aggregate (Table 13), and NAG sentences were on average twice as long as AG sentences with probation excluded. But recall that the NAG cases were almost all straightforward DUI cases with no identifiable victims, and all of the AG cases involved victims who suffered varying levels of injuries (sometimes death) as a result of a DUI accident. This means that when they did not use guidelines, judges sentenced non-injury DUI cases harsher than judges who sentenced injury-DUI cases with guidelines. This is similar to the observation that in child-abuse cases, without guidelines some judges sentenced some child-injury cases longer than child-death cases with guidelines.[186]

In the DUI NAG cases, all were felony DUIs, meaning the offender had several prior DUI convictions. Just as with the physical child-abuse cases, it is important to compare like offenses and like offenders, and it makes little sense to compare the sentence of someone convicted of a third DUI with someone convicted of their sixth. Accordingly, the table below controls for the number of DUIs.[187]

Table 17: Sentences for NAG DUI Cases Controlled by Offense Number

DUI #

# of Cases

Average Sentence

Median Sentence

# of Probation Sentences

Average (Probation Excluded)

Median (Probation Excluded)

Three DUIs

5

6

0

4

30

30

Four DUIs

6

26

9

2

39

15

Five DUIs

10

28

21

2

35

24

Thirteen DUIs

1

12

12

0

12

12

As one might hope to see, sentences tended to increase in length with more prior DUI offenses. However, there are three outliers in these categories.

First, for offenders in the “Three DUI NAG” category, every offender received a sentence of probation except one Black offender[188] in the Western District of Texas (CHC-2) who received a sentence of 30 months. The criminal history categories of other offenders in this category ranged from 1 to 5, and all received probation. There are no facts in the PSR or other documents to suggest that some aggravating factors existed in the case of the Black offender who received 30 months. Accordingly, although there might be reasons outside of the sentencing documents, that sentence appears to be disproportionately long.

In fact, this NAG sentence is longer than all but one of the sentences in the DUI AG cases that involved accidents, injuries, and deaths. In the sole, longer AG case, the victim died after the offender’s DUI accident and the judge analogized the DUI to involuntary manslaughter (40 months; CHC-5). In the other Three DUI AG case, a man was intoxicated and fell asleep at the wheel, hitting and killing a person in a car parked on the shoulder. After also analogizing to involuntary manslaughter, he (CHC-1) received a sentence of 12 months. Thus, the Texas Three DUI NAG offender involving no accident was sentenced to almost three times longer than a Third DUI AG offender who killed someone.

For offenders in the Four DUI category, the average sentence was 26 months, but one sentence was a substantial outlier. Two offenders within the Four DUI category received probation (a Black woman and a Caucasian male), and the average sentence of three other offenders was 12 months (CHC-1, 3, and 3).[189] However, a Black male in the Western District of Louisiana[190] was sentenced to ten years for his fourth DUI (by far the longest sentence in the category). Like many such offenders, he had several prior DUI-related offenses, and the only other potentially aggravating factor in the records was that he wrecked into bollards outside of a military base during his DUI offense. But several other offenders with similar backgrounds and DUI accidents received much shorter sentences. For example, two Caucasian offenders who were convicted of their fifth DUI also wrecked into government property at military bases. One offender was sentenced to 36 months (CHC not calculated) and another (CHC-1) received a sentence of 6 months. Again, there could be reasons that do not appear in the sentencing documents, but this 120-month sentence vastly exceeds sentences of other offenders with similar conduct and criminal histories.

There is a similar irregularity among fifth-time DUI offenders. A Black male offender in the Western District of Texas[191] received a sentence of ten years, when the average sentence among other fifth-time offenders (excluding this ten-year sentence) was 18 months. Nothing in the offense conduct stands out as significant, and officers arrested the offender without incident. Many other Five DUI cases involved more severe conduct, like accidents or fleeing. Nonetheless, this person’s sentence was over five times longer than similar offenders committing similar offenses.[192]

In the final category, one offender (CHC-5) in the Western District of Washington received 12 months after his thirteenth DUI.

The DUI dataset is visually represented below:

Figure 2: DUI Sentences

AG cases are represented with a white point, the NAG cases with a black point. As this plot reveals:

In summary, DUI cases revealed some of the most extreme disparity in the entire analysis. This could be because while a state judge (and negotiating counsel) likely sees several DUIs per week, a federal judge, and only one in an ACA-relevant jurisdiction, might only see a few over an entire career. State systems also have robust options that are alternatives to incarceration in DUI cases,[193] tools traditionally absent in the federal system.[194]

D. Convictions Under Federal Statutes

The most common federal statute AG and NAG cases are identified in Section II.B.[195] For most of these offenses, each type of offense was uniformly either an AG or a NAG, with few instances of crossover. This indicates that judges struggle less with §2X5.1 in Fed. Stat. cases. For example, in addition to uniformly analogizing maritime drug offenses to §2D1.1, every instance of bankruptcy fraud[196] (N=23)[197] was analogized to §2B1.1 (Larceny), and every violation for failure to register as a foreign agent[198] (N=7) was considered a NAG.

The only crossover offenses within the review period were piloting an aircraft without a pilot’s license[199] (AG N=6, analogized to §2B1.1 (Larceny); NAG N=4) and civil disorder or rioting[200] (AG N=8, analogized to §2K1.4 (Arson), §2A2.4 (Obstruction of Justice), or §2B1.1 (Larceny); NAG N=2).

The most common offense did not present a substantial analogous-guideline gap:

Table 18: Sentences for Piloting Without a License

 

NAG

AG

Average

5

6.8

Median

3

7

Table 19: Sentences for Piloting Without a License

(Probation Excluded)

 

NAG[201]

AG

Average

14

13

Median

14

12

In the civil disorder/rioting cases (mostly related to conduct occurring during protests/riots), with only 2 NAG sentences during the review period,[202] the number is too small to render any comparative conclusions. Both NAG cases resulted in sentences of probation, while the average sentence in AG cases was 18 months. Some might quibble with the sentences in these cases on the merits,[203] but there are not enough cases to conclude that any disparity is rooted in the application of §2X5.1. Like contempt, the conduct underlying these offenses can vary so wildly that, so long as the judges properly apply the elements-based approach described above and judges have access to sentencing data from similar cases, unwarranted disparity can be avoided.[204]

Compared to assimilated crimes, Fed. Stat. AG and NAG cases were sentenced within a narrower range, judges were more consistent in AG/NAG determinations, there were smaller analogous-guideline gaps, and there were far fewer notable disparities or outliers. Section 2X5.1 appears to be operating more effectively for federal offenses than for assimilated state crimes.

E. Trends

The Sentencing Commission’s 2022 sentencing data was only preliminary at the time of drafting this Article and was therefore not incorporated into the primary analysis. It was nonetheless reviewed to assess consistency with data within the dataset, to show trends over time, and to highlight any noteworthy cases. The following chart shows the number of assimilated offenses sentenced under §2X5.1 over time:

Figure 3: Assimilated Offenses Sentenced Under § 2X5.1

This chart shows the decrease in the number of these sentences (as with all cases) as a result of the COVID-19 pandemic.[205] It also shows that judges analogize to guidelines more often than not—particularly in 2022 (79% of assimilated cases in 2022). Sentences in 2022 were generally consistent with those in the dataset—many analogies likely invoked the incorrect guidelines under the elements-based approach, many understated the seriousness of the offense,[206] and disparity was widespread.[207]

The 2022 data confirm another important trend: the dramatic increase in the number of §2X5.1 sentences in Oklahoma after the Supreme Court’s landmark ruling in McGirt v. Oklahoma.[208] This is revealed when comparing the states with the highest number of §2X5.1 sentences:

Figure 4: Number of Assimilated Sentences Over Time
(Top States)

In most districts, §2X5.1 sentences have remained relatively constant (factoring in COVID-19), and before McGirt, Oklahoma federal districts averaged around one §2X5.1 sentence per year for assimilated offenses. After McGirt, that number is 20.[209] This aligns with data from the Administrative Office of the U.S. Courts, showing a substantial uptick in felony criminal filings in Oklahoma federal districts after McGirt:[210]

Figure 5: Total Oklahoma Felony Filings

McGirt resulted in around a 230% average increase in felony filings in the Eastern and Northern Districts, and approximately a 33% reduction between 2021 and 2022, likely due to the Supreme Court’s decision in Castro-Huerta.[211] However, as the previous chart shows, §2X5.1 sentences continued to increase even as total filings fell after Castro-Huerta (likely reflective of the districts being at capacity and prioritizing serious offenses like child abuse).[212] Oklahoma cases now make up virtually half of all assimilated §2X5.1 sentences.[213]

FY2022 also saw a substantial increase in §2X5.1 cases based on violations of federal statutes. This was a result of many of the offenders in various George Floyd protests (11 sentences) and the January 6, 2021, “Capitol Breach” cases (11 sentences) being sentenced in 2022. Most in the former group were sentenced under §2K1.4 (Arson) with a mean sentence of 23.7 months and a median sentence of 24 months, and most in the latter group were sentenced under §2A2.4 (Obstructing or Impeding Officers) or §2A2.2 (Assault) with a mean sentence of 20 months and a median sentence of 6.5 months.[214]

III. Summary of Findings

The dataset consisted of 434 cases sentenced under §2X5.1 between 2016 and 2021. The dataset was divided into two categories—federal offenses that do not have a direct guideline (N=198), and assimilated state offenses (N=236). The sentences were then divided again into subcategories: (1) sentences for which judges applied an analogous guideline to an offense (AG) and sentences for which judges found that there were no analogous guidelines (NAG). This Article focuses on assimilated state offenses.

The most common crossover assimilated offenses were child abuse and felony DUI. For these offenses, sometimes the same courts (and occasionally the same judges) applied different guidelines or no guidelines for the same types of offenses. This appears to be based on (1) confusion around correctly applying §2X5.1 (Section I.C clarified the test), and (2) binding or non-binding plea agreements agreeing to various guidelines in order to reach an agreed guideline range. The inconsistent application of different or no guidelines resulted in disparity. This disparity manifests as (1) an analogous-guideline gap (disparity depending on whether the judge applied guidelines or sentenced without guidelines); (2) the difference between the guideline minimum (the most common sentences for most other federal offenses) and actual sentences in these cases; and (3) a much wider range of sentences when judges do not apply guidelines.

For the most common type of child-abuse cases (physical child abuse), sentences averaged twice as long when judges did not apply guidelines. Child endangerment cases involved outliers, discussed in Part II.[215] Sentences were 37% longer in felony DUI cases when judges did not apply guidelines. Average sentences are presented below, with both total averages and probation excluded (P.E.). The difference between the NAG and AG sentences is the analogous-guideline gap:

Figure 6: Summary of Average Sentences

A summary of median sentences is also presented below:

Figure 7: Summary of Median Sentences

When comparing physical-abuse cases to the guideline minimum (the most common sentence for all federal offenses), when judges did not apply guidelines they sentenced years longer than when they applied assault guidelines.[216] Notably, judges also sentenced longer than the guideline minimums when they did apply guidelines in these cases (around 9 months longer than the assault-guideline-minimum and 19 months longer than the aggravated-assault-guideline minimum). This could suggest that judges do not believe that assault guidelines adequately capture the severity of child-abuse offenses.

There was also more significant variability in sentences without guidelines. For example, as depicted in Figure 1, while sentences for physical child abuse imposed with guidelines increased consistently with the severity of the offenses, sentences imposed without guidelines varied substantially. Similar variability existed in DUI sentences, with some offenders receiving probation and others receiving ten years in prison after seemingly identical DUI convictions.

Finally, the number of §2X5.1 cases has been relatively constant in most districts but has grown dramatically in the District of Columbia (as a result of the January 6, 2021, Capitol Breach) and in Oklahoma federal courts (as a result of McGirt v. Oklahoma). Proposals to amend the Guidelines could also lead to more cases applying §2X5.1. For example, §2X5.1 has been relied upon in proposals to reform the complicated “categorical approach,”[217] as well efforts to simplify the Guidelines.[218] Adoption of either reform would further increase attention to §2X5.1 and its existing body of law and commentary, as judges would look to §2X5.1 for guidance in these new areas. This Article endeavors to provide important clarity about existing cases that have relied on §2X5.1 in anticipation for the guideline’s increased use.

IV. Recommendations

Having a fair and consistent process for sentencing these difficult cases can minimize unwanted variability.[219] The above analysis and a review of relevant case law guide several best practices and recommendations directed to judges, the Sentencing Commission, and the United States Probation Office.

A. Judges

While the Sentencing Commission, Probation Office, judges, and parties should work together to achieve fair sentencing, it is ultimately the judge’s signature on the judgment.

Judges should independently conduct a “most analogous guideline” analysis and must reject incorrect guideline calculations. The sentencing judge must accurately calculate the guideline range.[220] That duty entails correctly analogizing to the correct guidelines.[221] As previously discussed, the application of a guideline is ultimately a legal conclusion that the judge renders. Regardless of any party agreement or PSR calculation, sentencing judges should independently confirm whether the case is an AG or a NAG, and if it is an AG, should confirm that the correct guidelines are being applied. Judges should apply the elements-based approach in most jurisdictions. If the judge concludes that it is appropriate in an AG case to sentence outside of the advisory guideline range (based on the judge concurring with a plea agreement or for any other legitimate reasons), the judge can simply depart or vary.

It is important to appreciate the impact of applying incorrect guidelines, even if doing so by party agreement. Instead of analyzing the offense conduct and relevant conduct to reach an appropriate guideline calculation, the cases within the review period show that parties often do almost the inverse—they agree on a sentence and then engage in what I call cafeteria analogizing. To reach guideline calculations in plea agreements, they take a scoop of one guideline (which sometimes bears no relation to the offense conduct), ignore other relevant guidelines, add an enhancement, and season with a reduction for acceptance of responsibility (sometimes -2, sometimes -3)[222] all to confabulate a guideline range that coheres with a sentence they had agreed to long before.[223] “[L]ying to a court, or deceiving it by withholding relevant facts, is wrong, period, and there is never a valid justification for it.”[224]

A troubling consequence of improper guideline stipulations is that all of this legally incorrect and inconsistent information is forwarded to the Sentencing Commission, concealing from the Sentencing Commission the actual basis for the sentence and thwarting the feedback loop that the Sentencing Commission has to fix deficient guidelines. Some unpacking is required here. Congress requires that the Sentencing Commission “periodically shall review and revise, in consideration of comments and data coming to its attention, the guidelines,”[225] and in furtherance of this duty, Congress gave the Sentencing Commission the power to, among other things, “serv[e] as a clearinghouse and information center for the collection, preparation, and dissemination of information on Federal sentencing practices.”[226] This works in tandem with every chief district judge’s duty to timely provide the Sentencing Commission with the charging documents, plea agreements, J&C, SOR, and PSR in every case.[227] But if judges are analogizing to the wrong guidelines, the Sentencing Commission is oblivious to the fact that there might be a problem with the correct guidelines in these cases.

For example, as noticed in its 2022–2023 amendment cycle, the Sentencing Commission amended the Guidelines to give defendants with no criminal history potentially lower guideline ranges for certain offenses based on data showing that these offenders are less likely to reoffend, and showing that a vast majority of judges are sentencing those offenders below the guideline range.[228] Data revealed to the Sentencing Commission the asymmetry between the Guidelines and judges’ sentences, and an amendment would bring the Guidelines into better alignment with judges’ actual sentencing practices.[229] The Sentencing Commission can analyze the hypothetical result of a one, two, or three-point reduction (or other possible amendment) and observe how those changes would align with where judges currently sentence.[230]

But this analysis would be impossible if judges did not first correctly calculate the guidelines in those cases and then consistently depart or vary downward, flagging the issue for the Sentencing Commission. When, as in many of the assimilated cases in the dataset, judges apply arbitrary guidelines that yield the range that the judge or parties consider fair, problems with the correct guideline are concealed from the Sentencing Commission’s corrective processes. By continuing to analogize to the incorrect guidelines, judges skew the data that the Sentencing Commission receives (and that future litigants and judges will rely upon as comparators). The optimal approach is to accurately calculate the guidelines and if the judge agrees with the parties’ stipulated sentence length or range, impose a sentence that conforms. If judges regularly depart or vary from a guideline range, the issue will be more likely to be flagged for the Sentencing Commission’s corrective processes.

Judges should not force sentences into the Guidelines when they do not belong. As cited through this Article, many cases—particularly, many assimilated crimes—simply do not fit comfortably within the Guidelines. Section 2X5.1 cases are “by definition ‘unusual’”[231] and for some offenses, analogizing to a guideline is like trying to fit a square peg into a round hole.[232] While the elements-based approach results in proper analogies to assault guidelines in most cases of physical child abuse, drug guidelines in maritime trafficking cases, or larceny guidelines in bankruptcy-fraud cases, the same is not true for child endangerment, child neglect, contempt, or DUI cases.

In many of these cases, rather than straining analogies to guidelines, judges should simply analogize to guidelines less often. For the square-peg offenses, judges should avoid the “fear of judging” without guidelines.[233] But this conclusion likely only holds if more information is made available to judges on comparator cases. If the other changes in the Article are adopted, data on comparator sentences will begin to develop a “heartland” of sentences.[234] Some commentators have proposed a data-based “presumption of unreasonableness” on appeal if a sentence falls outside of a congressionally mandated range of reasonable sentences.[235] Congress’s 25% rule could be a starting place, rendering a sentence above or below 25% (or 6 months) of the average or median sentence of similarly situated offenders presumptively unreasonable—a presumption overcome by a judge’s valid and documented reasons.[236]

Judges should provide more information in §2X5.1 cases. Judges tend to be eager to assist their colleagues in the administration of justice. This should particularly hold true in the relatively small world of sentencing difficult §2X5.1 cases.[237] Including a modest amount of additional information in these cases can go a long way toward ensuring more just sentences overall.

In §2X5.1 sentences, judges should provide more specific reasons for the sentence in the SOR (likely under “Additional facts justifying the sentence”). This would include why a judge is or is not analogizing, whether other factors (e.g., credit for serving a tribal sentence) were considered, and in no-guideline cases should provide more specific information about the reason for the sentence. Assuming some of the other recommendations in this Article are adopted, the judge would have access to information on sentences in comparable cases and could state where (and why) the present sentence fits among the comparators. Importantly, if the sentence is an outlier from the “heartland” of related cases, the SOR should explain this, for the benefit of defendants, future judges, and interested entities.[238]

Judges should also clearly include in the J&C the state statute of conviction (including subsection) so that the Sentencing Commission can code this data.[239] Finally, if the statute does not already do so for child-abuse cases, judges should specify in the SOR whether the judge believes that the conduct constitutes physical abuse, endangerment, or neglect. This will allow accurate data gathering (discussed below) and permit future comparisons of like cases.

With these few lines of information, judges can work collectively to help minimize unwarranted disparity in these difficult cases.

B. The United States Probation Office

Probation officers should conduct a truly independent analysis of the applicable guidelines. “The probation officer in the federal system is an employee of the judicial branch, and traditionally has acted as a confidential adviser to the court.”[240] The United States Probation Office has “developed a remarkable expertise in the guidelines and the case law interpreting it,” and officers have sometimes been called the “‘guardians of the guidelines’ based on their neutral, unbiased role in implementing the guidelines sentencing regime in our adversarial system.”[241] The probation officer plays a crucial role in ensuring that the guideline range is correctly calculated.[242]

The officer’s PSR is “the epicenter of the guidelines sentencing process.”[243] Indeed, Congress requires probation officers to submit a PSR in the manner that Federal Rule of Criminal Procedure 32(c) requires,[244] and Rule 32(d) provides that the PSR must, among other things:

(A) identify all applicable guidelines and policy statements of the Sentencing Commission;

(B) calculate the defendant’s offense level and criminal history category;

(C) state the resulting sentencing range and kinds of sentences available;

(D) identify any factor relevant to:

(i) the appropriate kind of sentence, or

(ii) the appropriate sentence within the applicable sentencing range; and

(E) identify any basis for departing from the applicable sentencing range.

This instruction to “identify all applicable guidelines” entails the probation officer correctly identifying all guidelines, not identifying the wrong guidelines (even those based on party agreement) and rendering calculations therefrom.

The probation officer should therefore conduct a full, independent analysis of the correct guideline range, regardless of any party stipulations or agreements. This includes an independent AG/NAG analysis, with data from the Sentencing Commission on AG/NAG outcomes in similar cases if necessary. In this way, the probation officer acts as another line of defense against the application of incorrect guidelines. The issue cannot evade both the adversarial process and the probation officer’s independent analysis.[245]

The probation officer should include additional information in the PSR. Much of this is already done in most districts based on cases in the dataset but the practices should be uniform.

First, either in the PSR or in a confidential supplement, in §2X5.1 cases the probation officer should include average and median federal sentences for the offense of conviction.[246] If the Sentencing Commission adopts the recommendations below, gathering the data should not be burdensome. If the Sentencing Commission does not, the probation officer should contact the Sentencing Commission to obtain this information. (This Article can be used for data from FY2016–2021). The probation officer should identify outliers or other relevant information for the judge and should try to identify sentences or a range of sentences for similar offenses and offenders. The best practice is to give adequate notice to the parties and an opportunity to object.[247]

Second, the probation officer should always calculate an offender’s criminal-history category—even in NAG cases.[248] Judges are accustomed to utilizing the officer’s calculations as an insight into the offender’s criminal past, the information is essential for the Sentencing Commission’s analyses, and the CHC is key when the sentence is used as a comparator in future cases. Additionally, it prepares the judge if the judge ultimately disagrees that the case is a NAG and elects to analogize to a guideline.

Third, the probation officer should always list the offense of conviction by subsection, where applicable. This is important for comparing like offenses with offenses that differ by subsection. For example, as discussed in Section I.C, the Nebraska child-abuse statute separates child abuse by subsection ((a) child endangerment, (b) cruel confinement or physical child abuse, and (c) child neglect). Clearly specifying the subsection will allow for more accurate comparison of like offenses. If there are doubts (sometimes charging documents do not state the applicable subsection), the probation officer can confirm the subsection with counsel, which the judge can confirm on the record and reflect in the J&C.

Finally, in child-abuse cases, regardless of whether the statute has subsections, the probation officer should indicate in the “Offense” section at the beginning of the PSR whether the conduct constitutes physical child abuse, child endangerment, or child neglect. As this Article has shown, sentences vary dramatically depending on which of these categories accurately describes a case, and this information is important for comparing a case to prior and future cases. When state statutes break offenses into these categories, the task is easier. When state statutes do not, the probation officer should utilize their best judgment given the offense conduct, and the parties are free to object to the probation officer’s conclusion. As set forth above, the judge’s conclusion on this issue will be reflected in the SOR, and in the future this variable can be used to compare like offenses. With this information, a body of sentences will begin to form the “heartland” of such cases,[249] and an interested party can finally obtain general sentencing information on, for example, a CHC-2 offender pleading guilty to child neglect.

C. The Sentencing Commission

The Sentencing Commission should modify how it codes and processes §2X5.1 cases. A judge’s sentence affirms both society’s need and authority to punish, but central to that ritual is the judge’s informed discretion.[250] The Sentencing Commission gives judges access to sentencing information upon request, but even if a judge asks, gathering §2X5.1 data is currently a difficult and time-consuming task due to the unique coding of these cases. Important details like the applicable state statutes and the type of offense are not specifically coded. Responding to an information request would currently require a manual review of each relevant indictment, PSR, plea agreement, J&C, and SOR in scores of cases. Given the wide range of sentences in these cases, it is not sufficient to identify one or some comparable cases, the reviewer needs all comparable cases. Relatedly, current coding does not allow the data to populate on the Sentencing Commission’s public databases (IDA and JSIN[251]), which judges, probation officers, advocates, and other interested parties use to see prior sentences from similar cases.

To provide judges and interested parties easier access to accurate data, the Sentencing Commission should consider the following changes to processing these cases:

These, and perhaps other modest coding changes will advance many other recommendations in this Article and will allow the Sentencing Commission to promptly provide accurate sentencing data to the appropriate entities. This will allow data to be returned in seconds instead of days or weeks (if at all).

The Sentencing Commission should consider rewording §2X5.1. The Sentencing Commission should make minor adjustments to the structure of §2X5.1 to reduce confusion. The language and word order should cohere with the more logical structure of how the background commentary[256] and judges conduct a “most analogous guideline” test: determine whether there are any sufficiently analogous guidelines and then apply the most analogous offense guideline. Minor rewording of subsection (a) would result in:

If the offense is a felony for which no guideline expressly has been promulgated, determine whether there are any sufficiently analogous guidelines and if so, apply the most analogous offense guideline.

The Sentencing Commission should amend the Statutory Index. Likely the easiest change, the Sentencing Commission should periodically identify sufficiently recurring federal offenses that uniformly analogize to the same guideline and should amend the Statutory Index accordingly. Currently, the maritime drug[257] and bankruptcy-fraud crimes[258] discussed above fit these criteria, and the statutory index could simply refer those offenses to §2D1.1 (Drug Offenses) and §2B1.1 (Larceny), respectively. This modest action would address 126 of the 434 §2X5.1 cases from 2016–2021, almost a third of the cases.[259] The Sentencing Commission could thereafter monitor §2X5.1 offenses for other candidates in a similar manner to how it accounts for new federal laws.

Conclusion

There appears to be unfairness and disparity in sentencing with and without guidelines in §2X5.1 cases. But that blame does not fall solely, or even primarily, on the judges or the parties (who often consent to these sentences); they are simply working within the unusual system thrust upon them. Some blame might be attributed to Joseph Story for envisioning a regime of borrowing state offenses for conviction in federal courts.

The entity that can have the greatest impact in resolving unfairness in §2X5.1 sentences is the Sentencing Commission. Congress gave the Sentencing Commission the power[260] but also the duty[261] to collect and report[262] federal sentencing data, in furtherance of the SRA’s goals of “avoiding unwarranted sentencing disparities.”[263] These cases have fallen through the cracks, and the Sentencing Commission should consider the modest recommendations described above to ensure that hundreds of offenders each year are not so readily exposed to unwarranted disparities, with no data to judge the reasonableness of the sentences.[264]

At least for the most common assimilated state offenses, judges appear to be harsher and more inconsistent when they sentence without guidelines. It is possible that there exists something unique about assimilated crimes that makes judges impose harsher sentences. If true, I would attribute this to (1) a lack of access to data in comparator cases (a problem that does not exist for other federal offenses); (2) the relative unfamiliarity with the types of cases (particularly with child abuse[265] and impaired driving); and (3) a stronger emotional response to child victims and drunk drivers.[266] Regardless, transparency and data should at a minimum help prevent the most extreme sentences on either end.

As Judge Jon O. Newman explained, “we can and should find a place for discretion between the two poles of unfettered discretion and no discretion.”[267] Section 2X5.1 cases demonstrate the soundness of this sentiment as regards the former. The Sentencing Guidelines are commonly regarded as unnecessarily harsh for many offenses.[268] Judges[269] and prosecutors[270] seem to agree. But §2X5.1 cases reveal the possibility that if judges had no guidelines and no data on comparable sentences (a situation that exists in many state systems), sentences could be harsher and much more disparate.

Appendix

U.S.S.G §2X5.1 Child Abuse Cases by State

State

#NAG

#AG

Guideline (GL)

Occurrence per GL

Arizona[271]

2

12

Ag. Assault[272]

Assault[273]

Involuntary Manslaughter[274]

7

3

2

Michigan[275]

0

3

Assault[276]

3

Minnesota[277]

1

0

N/A

N/A

Montana[278]

0

2

Ag. Assault[279]

2

Nebraska[280]

1

2

Kidnapping[281]

2

New Mexico[282]

21

3

Ag. Assault[283]

3

North Carolina[284]

0

1

Ag. Assault[285]

1

North Dakota[286]

0

8

Assault[287]

8

Oklahoma[288]

5

1

Ag. Assault[289]

1

South Dakota[290]

16

7

Ag. Assault[291]

Assault[292]

5

2

Texas[293]

0

4

Ag. Assault[294]

Assault[295]

1

3

Washington[296]

0

2

Assault[297]

Sex crimes[298]

1

1

Wisconsin[299]

1

0

Ag. Assault[300]

Assault[301]

Invol. Mansl.[302]

7

3

2

  1. * I completed this Article while I was a Visiting Assistant Professor at the University of Oklahoma College of Law. I compiled the data and drafted the Article while working as a Supreme Court Fellow in 2022–2023 with a placement at the United States Sentencing Commission, but the views expressed are in my personal capacity. This Article benefitted substantially by the generous feedback I received at presentations to the Criminal Law Committee of the Judicial Conference, the Tribal Issues Advisory Group at the United States Sentencing Commission, the Supreme Court Fellows Program Annual Workshop, and the Central States Law Schools Association Annual Scholarship Conference. For comments on drafts, structure, and data, I am indebted to countless staff members at the United States Sentencing Commission, Professors Melissa Mortazavi, Roger Michalski, Joseph Schremmer, and Jon Lee, as well as Brad Baranowski, Ph.D., Anya Ku, John Rockenbach, Celia Goetzl, and other practitioners and judges who have experience with these issues and contributed important practical considerations. No individual or organization acknowledged in this Article necessarily supports the Article’s interpretations or conclusions and I am solely responsible for any mistakes or misjudgments. ↑

  2. . United States v. Walker, 74 F.4th 1163, 1174–75 (10th Cir. 2023), cert. denied, 144 S. Ct. 611 (2024), cert. denied, 144 S. Ct. 1079 (2024). ↑

  3. . Id. ↑

  4. . Id. at 1175–76. ↑

  5. . Id. Mr. Morrison was charged with child abuse and Ms. Walker was charged with enabling it. Id. at 1176. ↑

  6. . Id. at 1177. ↑

  7. . Id. at 1181. ↑

  8. . Id. at 1206. ↑

  9. . Id. at 1206 (quoting United States v. Garcia, 946 F.3d 1191, 1215 (10th Cir. 2020)). ↑

  10. . See infra Section II.C.1.a.ii. ↑

  11. . See infra Section II.C.1. This does not include murder. See infra Section II.C.1.a.ii. ↑

  12. . Nora V. Demleitner, Douglas A. Berman, Marc L. Miller & Ronald F. Wright, Sentencing Law and Policy: Cases, Statutes, and Guidelines 84 (5th ed. 2022). ↑

  13. . 28 U.S.C. § 991(b)(1)(B) (stating that the Sentencing Commission should establish policies that avoid “unwarranted sentencing disparities among defendants with similar records who have been found guilty of similar criminal conduct”); Pub. L. No. 98-473, 98 Stat. 1837, 1987 (1984) (codified at Titles 18 and 28 of the United States Code). ↑

  14. . See 28 U.S.C. § 994(b)(2). ↑

  15. . As shown in Part II, §2X5.1 cases regularly exceed this cap (often dramatically), meaning these sentences defy Congress’s mandate. ↑

  16. . See, e.g., Letter from Hon. Ralph R. Erickson, Chair, Tribal Issues Advisory Grp., to Hon. Carlton W. Reeves, Chair, U.S. Sent’g Comm’n 1 (Sept. 7, 2022), https://www.ussc.gov/sites/default
    /files/pdf/amendment-process/public-comment/20221017/tiag.pdf [https://perma.cc/C9VD-PMHR%5D (stating that U.S.S.G. §2X5.1 “leaves sentencing judges in a position of great uncertainty as to the appropriate base offense level” in these cases); see also infra Part IV. ↑

  17. . U.S. Sent’g Guidelines Manual §2X5.1 (U.S. Sent’g Comm’n 2023). ↑

  18. . See generally Kate Stith & José A. Cabranes, Fear of Judging: Sentencing Guidelines in the Federal Courts (1998); Demleitner et al., supra note 11, at 171. On the unexpected complexity of the guidelines, see Jon O. Newman, The Federal Sentencing Guidelines: A Good Idea Badly Implemented, 46 Hofstra L. Rev. 805, 807 (2018) (“Those of us who supported the SRA expected the Sentencing Guidelines to be a fairly brief document, similar to the parole guidelines . . . [that were] just three pages, plus some explanatory material.”). See also Jon O. Newman, Simplified Guidelines—Now or Never, 35 Fed. Sent’g Rep. 5, 5 (2022) (“The first draft of the Federal Sentencing Guidelines . . . running to nearly 300 pages, appalled us by its complexity. . . . The guidelines remain as complex as they were when the first version was issued.”). ↑

  19. . See, e.g., Stith & Cabranes, supra note 17, at 62–63. ↑

  20. . See, e.g., Kevin Clancy, John Bartolomeo, David Richardson & Charles Wellford, Sentence Decisionmaking: The Logic of Sentence Decisions and the Extent and Sources of Sentence Disparity, 72 J. Crim. L. & Criminology 524, 526 (1981) (studying “possible sentences for sixteen hypothetical cases”); Anthony Partridge & William B. Eldridge, Fed. Jud. Ctr., The Second Circuit Sentencing Study: A Report to the Judges of the Second Circuit 5–7 (1974) (asking different judges to sentence hypothetical cases with real, fake, or modified sentencing documents). ↑

  21. . See Stith & Cabranes, supra note 17, at 108–10 (criticizing the use of hypothetical sentences to make conclusions about actual sentences and citing one judge admitting that at least some participants treated such exercises as a “joke”). ↑

  22. . This Article leaves for another day what the “right” sentence length is for a given offense and instead focuses on the variability of sentences. See Daniel Kahneman, Olivier Sibony & Cass R. Sunstein, Noise: A Flaw in Human Judgment 27 (2021) (explaining that all one needs to know is that the intent was to hit a target (here, a just sentence); with that piece of information, the reviewer need not know precisely where the target was “to measure the scatter . . . and to realize that the variability [is] a problem”). ↑

  23. . Id. at 30. ↑

  24. . Based on the current way that the Sentencing Commission processes the data in these cases, the cases are not searchable on the Sentencing Commission’s Interactive Data Analyzer (IDA) or the Sentencing Commission’s Judiciary Sentencing Information (JSIN) platform, nor is the information readily accessible if a judge requests it from the Sentencing Commission. See infra Part IV. ↑

  25. . 591 U.S. 894 (2020). ↑

  26. . For an analysis of sentences in Capitol Breach cases, see Sam J. Merchant, The Relative Severity of Criminal Sentences in the January 6, 2021, Capitol Breach Cases, 17 Drexel L. Rev. 171 (2024). ↑

  27. . The categorial approach is generally used to determine whether an enhanced punishment is appropriate based on an offender’s prior conviction for a violent crime or a serious drug offense. See, e.g., Taylor v. United States, 495 U.S. 575 (1990) (establishing the categorical approach); see also U.S. Sent’g Comm’n, Primer: Categorical Approach (2024), https://www.ussc.gov/sites/default/files
    /pdf/training/primers/2024_Primer_Categorical_Approach.pdf [https://perma.cc/6SA4-353Q]. ↑

  28. . See, e.g., Simplification Draft Paper, U.S. Sent’g Comm’n, https://www.ussc.gov/research
    /research-and-publications/simplification-draft-paper-1 [perma.cc/VQ3Q-2YWR] (consolidating or eliminating guidelines that are never or rarely applied “would allow courts to exercise discretion in choosing an analogous guideline or, if one cannot be found, sentencing according to the general precepts of the Sentencing Reform Act”). ↑

  29. . The level of detail in Part II has been provided so that even if some of the key recommendations in Part III are not enacted, relevant stakeholders will have this Article as a resource to assess the reasonableness of future sentences. ↑

  30. . Demleitner et al., supra note 11, at 84 (emphasis added). ↑

  31. . Id. In the federal system, the executive agencies were the United States Parole Commission or the Federal Bureau of Prisons. ↑

  32. . Id. (“An executive branch agency . . . ultimately determines the actual sentence each defendant serves.”); Brent E. Newton & Dawinder S. Sidhu, The History of the Original United States Sentencing Commission, 1985–1987, 45 Hofstra L. Rev. 1167, 1170 (2017) (“As the result of parole and good time allowances, many federal prisoners were released well before the expiration of the sentences of imprisonment imposed by federal district courts.”); Stith & Cabranes, supra note 17, at 19 (“[T]he introduction of parole [in 1910] significantly reduced federal judicial authority over the duration of prison sentences because parole authorities, not judges, would determine each federal prisoner’s actual release date.”); see also Kevin R. Reitz, Don’t Blame Determinacy: U.S. Incarceration Growth Has Been Driven by Other Forces, 84 Tex. L. Rev. 1787, 1795 (2006) (“‘Determinate’ sentencing systems, by formal definition, are those that have abolished parole-release discretion.”). ↑

  33. . Newton & Sidhu, supra note 31, at 1173–74 (citing, among other sources, comments from Commissioner and eventually Supreme Court Justice Stephen G. Breyer). ↑

  34. . See id. at 1174 n.43. ↑

  35. . Bruce M. Selya & John C. Massaro, The Illustrative Role of Substantial Assistance Departures in Combatting Ultra-Uniformity, 35 B.C. L. Rev. 799, 800 (1994); see also Michael M. O’Hear, Beyond Rehabilitation: A New Theory of Indeterminate Sentencing, 48 Am. Crim. L. Rev. 1247, 1247 (2011) (indeterminate sentencing “had been closely associated with the rehabilitative paradigm in criminal law, which also fell from favor in the 1970s”). On the history of the lack of constitutional protections at sentencing and on appeal, see Kevin R. Reitz, The New Sentencing Conundrum: Policy and Constitutional Law at Cross-Purposes, 105 Colum. L. Rev. 1082, 1083 (2005) (describing the historical “hands-off approach to criminal sentencing” as a constitutional matter). ↑

  36. . Stith & Cabranes, supra note 17, at 31–37; Demleitner et al., supra note 11, at 132. But see Stith & Cabranes, supra note 17, at 106–12 (calling the studies “flawed” and unworthy of reliance). ↑

  37. . Newton & Sidhu, supra note 31, at 1180. ↑

  38. . Id. at 1167; see also Stith & Cabranes, supra note 17, at 1 (The Federal Sentencing Guidelines brought two centuries of sentencing in federal courts “to an abrupt end”). ↑

  39. . See Jeffrey S. Parker & Michael K. Block, The Limits of Federal Criminal Sentencing Policy; or, Confessions of Two Reformed Reformers, 9 Geo. Mason L. Rev. 1001, 1002 (2001) (“[T]he Sentencing Reform Act is unusual in directing an agency of government both to create a new body of law and to develop new knowledge about the efficacy of criminal punishment—a subject of central concern to society that has been debated for centuries.”). ↑

  40. . Sentencing Reform Act of 1984, Pub. L. No. 98-473, § 217(a), 98 Stat. 1837, 2018 (codified at 28 U.S.C. § 991(b)(1)(b)). ↑

  41. . Id. § 212(a)(2), 98 Stat. at 1989 (codified at 18 U.S.C. § 3553(a)). ↑

  42. . See, e.g., U.S. Sent’g Guidelines Manual ch. 1, pt. A(2) (U.S. Sent’g Comm’n 2023) (explaining that the Guidelines Manual “took effect on November 1, 1987”). ↑

  43. . U.S. Sent’g Guidelines Manual §1B1.1 (U.S. Sent’g Comm’n 2023). ↑

  44. . The Sentencing Table is located on the back cover of the Sentencing Guidelines Manual, at Chapter 5 of the Sentencing Guidelines, or online at Annotated 2021 Chapter 5, U.S. Sent’g Comm’n, https://www.ussc.gov/guidelines/2021-guidelines-manual/annotated-2021-chapter-5 [https://perma.cc
    /S8JQ-M477]. ↑

  45. . United States v. Booker, 543 U.S. 220 (2005). ↑

  46. . Gall v. United States, 552 U.S. 38, 51 (2007). ↑

  47. . See infra note 144 and accompanying text. ↑

  48. . Rita v. United States, 551 U.S. 338, 347 (2007). Additionally, judges are required to record significantly less information and provide fewer reasons for sentences that are within guideline ranges. See Sam J. Merchant, Plea Agreements and Suspending Disbelief, 37 Fed. Sent’g Rep. 15 (2025) (identifying problems with these minimal reporting requirements). ↑

  49. . 28 U.S.C. § 994(b)(2). ↑

  50. . See Newton & Sidhu, supra note 31, at 1237–40. ↑

  51. . Parker & Block, supra note 38, at 1007; see also Newton & Sidhu, supra note 31, at 1175–76 (describing the “crazy quilt” and “haphazard nature of federal penal laws in the mid-1980s”). ↑

  52. . See infra Part II. ↑

  53. . See United States v. Hudson, 11 U.S. (7 Cranch) 32, 34 (1812). ↑

  54. . Joseph Story, Manuscript (1816), in Life and Letters of Joseph Story 293, 297 (Boston, Charles C. Little & James Brown, William W. Story ed., 1851). ↑

  55. . Id. at 297–98 (“The only question is, whether this is to be done by passing laws in detail respecting every crime in every possible shape, or shall give the Courts general jurisdiction to punish wherever the authority of the United States is violated. . . . In my judgment, the former course is utterly impracticable. Crimes are so various in their nature and character, and so infinitely diversified in their circumstances, that it is almost impossible to enumerate and define them with the requisite certainty. An ingenious rogue will almost always escape from the text of the statute book.”). ↑

  56. . Act of Mar. 3, 1825, ch. 65, 4 Stat. 115; United States v. Press Publ’g Co., 219 U.S. 1, 12 (1911) (noting that section 11 of Justice Story’s draft became section 3 of the Assimilative Crimes Act of 1825). ↑

  57. . Story, supra note 53, at 297. ↑

  58. . Federal enclaves can include tribal lands, national parks, federal buildings, airports, aboard air or space craft, and areas within the special maritime jurisdiction of the United States. See 18 U.S.C. § 7. ↑

  59. . See 18 U.S.C. § 13(a). ↑

  60. . Lewis v. United States, 523 U.S. 155, 160 (1998) (citing Williams v. United States, 327 U.S. 711, 718–19 (1946)). The ACA encompasses federal regulations as well as statutes. E.g., United States v. Yates, 211 F. App’x 925, 927 (11th Cir. 2006) (“Every court to have decided the issue has concluded that federal regulations are ‘enactment[s] of Congress’ within the meaning of the Act.”) (collecting cases). ↑

  61. . 18 U.S.C. § 13(a). It is worth clarifying that the federal courts are not enforcing state laws; federal prosecutors are enforcing federal law (through the Assimilative Crimes Act), “the details of which, instead of being recited, are adopted by reference.” Puerto Rico v. Shell Co. (P.R.), Ltd., 302 U.S. 253, 266 (1937) (citing Press Publ’g Co., 219 U.S. at 1). ↑

  62. . 18 U.S.C. § 13(a); see also Barbara J. Van Arsdale et al., Which State Laws Are Assimilated by Assimilative Crimes Act, in 8A Fed. Proc., Law.’s Edition § 22:20 (“[F]ederal courts are not required to follow specific provisions of state law that go beyond establishing the elements of an offense and the range of punishment.”). ↑

  63. . See, e.g., United States v. Christie, 717 F.3d 1156, 1171 (10th Cir. 2013) (“This circuit and others have unanimously interpreted the ACA’s ‘like punishment’ mandate to require federal sentencing courts to abide any maximum and minimum prison terms proscribed by state law for an assimilated crime.”). ↑

  64. . 18 U.S.C. § 1152. ↑

  65. . Id. § 1153. ↑

  66. . The GCA and ACA use the term “Indian.” As applied to tribal members, this term originated based on Christopher Columbus’s mistaken belief that he had reached India, not America. See 14 Am. Jur. Pleadings & Practice Forms Annotated: Indians, American Introductory comments § 1, Westlaw (database updated Sept. 2024) (“The name ‘Indian’ was given to the inhabitants of this continent, on its discovery by Columbus, under the erroneous supposition that the new land was part of India. . . .”). This Article generally uses “tribal.” ↑

  67. . Thomas W. Hutchison, Sigmund G. Popko, Michael P. O’Connor & Celia M. Rumann, Federal Sentencing Law and Practice, § 2X5.1 cmt. 2 (2023 ed.). ↑

  68. . See U.S. Sent’g Guidelines Manual §2X5.1 (U.S. Sent’g Comm’n 2023). ↑

  69. . For example, 30 U.S.C. § 820 (false statements by operator of a mine) or 12 U.S.C. § 1715z–19 (equity skimming on a mortgaged home). ↑

  70. . For example, the Sentencing Commission has not promulgated a guideline for contempt because of the unusual variety of forms contempt can take, and associated impropriety of attempting to capture all such conduct in one guideline. U.S. Sent’g Guidelines Manual §2J1.1, Application Note 1 (U.S. Sent’g Comm’n 2023) (“Because misconduct constituting contempt varies significantly and the nature of the contemptuous conduct, the circumstances under which the contempt was committed, the effect the misconduct had on the administration of justice, and the need to vindicate the authority of the court are highly context-dependent . . . .”). ↑

  71. . N.D. Cent. Code §§ 14-09-22(1), 12.1-32-01 (2023). ↑

  72. . U.S. Sent’g Guidelines Manual §3A1.1(b)(1), §2A2.3(b)(1)(B) (U.S. Sent’g Comm’n 2023). ↑

  73. . Id. §3E1.1(a)–(b). Offenders who accept responsibility are eligible for a 2- or 3-point reduction in their offense level. See id. §3E1.1. Because the mother in this hypothetical has an offense level of below 16, she is only eligible for a 2-point reduction. See id. §3E1.1(b). ↑

  74. . See Sentencing Table (in Months of Imprisonment), U.S. Sent’g Comm’n (2021), https://
    http://www.ussc.gov/sites/default/files/pdf/guidelines-manual/2021/Sentencing_Table.pdf [https://perma.cc
    /7EY3-35B3]. ↑

  75. . See S.D. Codified Laws §§ 26-10-1, 22-6-1 (2024). ↑

  76. . Mich. Comp. Laws § 750.110a(5) (2024). ↑

  77. . La. Stat. Ann. § 14:98.4(A)(1) (2024). ↑

  78. . Okla. Stat. tit. 21, § 843.5(C) (2024). ↑

  79. . U.S. Sent’g Guidelines Manual §2X5.1 (U.S. Sent’g Comm’n 2023) (“If the offense is a felony for which no guideline expressly has been promulgated, apply the most analogous offense guideline.”). ↑

  80. . United States v. Allard, 164 F.3d 1146, 1150 (8th Cir. 1999) (citing United States v. Osborne, 164 F.3d 434 (8th Cir. 1999)). ↑

  81. . United States v. Jackson, 862 F.3d 365, 387 (3d Cir. 2017). ↑

  82. . United States v. Leahy, 169 F.3d 433, 441 (7th Cir. 1999). ↑

  83. . See, e.g., Jackson, 862 F.3d at 404 (McKee, J., dissenting) (“I think the Majority’s application of that test confuses the two steps of the analysis.”). The Tenth Circuit, likely with the benefit of hindsight, appears to have sufficiently parsed the guideline and provided the clearest test in United States v. Clark, 981 F.3d 1154 (10th Cir. 2020). ↑

  84. . See Hutchison et al., supra note 66, §2X5.1 cmt. 4. ↑

  85. . See infra note 113. ↑

  86. . Approaches in the “Other” category lack a clearly defined test. Most appear to holistically consider whether an analogy was “reasonable” based on a review of the defendant’s conduct (typically limited to only conduct alleged in the indictment) together with guidelines that appear related to the conduct. ↑

  87. . United States v. Osborne, 164 F.3d 434 (8th Cir. 1999). ↑

  88. . To identify prospect statutes, one can (1) rely on experience and party submissions, and (2) review the titles of guidelines in the Table of Contents to find guidelines covering similar conduct and then refer to the Statutory Provisions cited after those guidelines. See, e.g., United States v. Montanez, 36 F.4th 824, 825–26 (8th Cir. 2022); United States v. Neilson, 721 F.3d 1185, 1188 (10th Cir. 2013). ↑

  89. . United States v. Jackson, 862 F.3d 365, 382 (3d Cir. 2017) (“[I]t is only to be expected that the offense of conviction may include more expansive elements than the federal offense or additional elements missing from the federal counterpart. . . . ‘[A]nalogy does not mean identity.’” (quoting United States v. Langley, 919 F.2d 926, 931 (5th Cir. 1990))). ↑

  90. . United States v. Clark, 981 F.3d 1154, 1163 (10th Cir. 2020) (quoting Jackson, 862 F.3d at 375). For assimilated crimes, if the elements perfectly match, the government should have charged the federal offense. See Lewis v. United States, 523 U.S. 155, 165 (1998). ↑

  91. . Jackson, 862 F.3d at 388 (explaining that if “‘some plausible analog[y]’ exists between the elements of the defendant’s crime and the elements of federal offenses covered by the existing offense guideline,” it can be considered sufficiently analogous) (alteration in original) (quoting United States v. Rakes, 510 F.3d 1280, 1288 (10th Cir. 2007)). This Article observes no meaningful difference between how courts analyze “within the ballpark” and “plausibl[y] analog[ous]” for these purposes. See id. If a choice is forced, the latter is preferred, given courts’ greater familiarity with plausibility determinations and lack of precision in a “within the ballpark” standard. See id. ↑

  92. . Rakes, 510 F.3d at 1287. ↑

  93. . United States v. Montanez, 36 F.4th 824, 825 (8th Cir. 2022). ↑

  94. . State Bar of Ariz., Revised Arizona Jury Instructions (Criminal) § 36.23B (2022). ↑

  95. . Comm. on Model Jury Instructions, Manual of Model Criminal Jury Instructions for the Ninth Circuit § 8.6 (2022). ↑

  96. . Id. § 8.7. ↑

  97. . The same might not be true for cases of child endangerment or neglect, which would not require an affirmative act. E.g., United States v. Clark, 981 F.3d 1154, 1165 (10th Cir. 2020) (explaining that unlike assault, “child neglect does not require an affirmative act on the part of the defendant, but rather only willful or malicious neglect”). ↑

  98. . This is because only certain variables are currently coded when the Sentencing Commission receives case files from sentencing courts. This protocol is sufficient to capture the necessary data for federal offenses that have guidelines (most federal offenses), but special coding or modifications to the existing protocol would be needed for the more unusual offenses that do not have direct guidelines. Important variables that would make these sentences accessible are coded as “other,” “N/A,” or similar, and therefore do not appear to be captured when typical reports are run. I identified these cases using alternative methods, described in Section II.A. ↑

  99. . See U.S. Sent’g Comm’n, Variable Codebook for Individual Offenders: Standardized Research Data Documentation for Fiscal Years 1999 – 2022 (2023), https://www.ussc.gov/sites/default/files/pdf/research-and-publications/datafiles/USSC_Public_Release
    _Codebook_FY99_FY22.pdf [https://perma.cc/VW46-VHMB]. ↑

  100. . See, e.g., U.S. Sent’g Comm’n, 2021 Annual Report and Sourcebook of Federal Sentencing Statistics 71 (2021) (reporting zero cases sentenced with §2X5.1 as the primary guideline in FY2021). ↑

  101. . The fiscal year begins October 1 and ends September 30. E.g., Glenn R. Schmitt & Lindsey Jeralds, U.S. Sent’g Comm’n, Fiscal Year 2021 Overview of Federal Cases 24 n.2 (2022), https://www.ussc.gov/sites/default/files/pdf/research-and-publications/research-publications
    /2022/FY21_Overview_Federal_Criminal_Cases.pdf [https://perma.cc/TZF3-V7GA]. ↑

  102. . See 28 U.S.C. § 994(w)(1). ↑

  103. . This included variables “STA1_1-STA1” and “NWSTAT1-NWSTAT” for and offender’s first count of conviction and associated variables for any subsequent counts, if applicable. See U.S. Sent’g Comm’n, supra note 98. ↑

  104. . See U.S. Sent’g Guidelines Manual §2X5.2 (U.S. Sent’g Comm’n 2023). ↑

  105. . See U.S. Sent’g Comm’n, supra note 98. ↑

  106. . This included “GDLINE1-GDLINEX,” “GDLINEHI,” and “OFFGUIDE” which presents the “primary type of crime for the case generated mainly from the primary guideline the court applied at sentencing.” U.S. Sent’g Comm’n, supra note 98, at 7–8, 43. ↑

  107. . For example, a federal offense that does not have a specific guideline. ↑

  108. . The dataset includes five cases that were not included in the initial results but were identified through a quality review (for offenders who had co-defendants, ensuring that the co-defendant’s case was also in the dataset). ↑

  109. . I have endeavored to capture appeals as of the time of drafting, but as with all sentencing data, it is possible that other sentences were modified on appeal or through other post-sentencing remedies. ↑

  110. . See Lewis v. United States, 523 U.S. 155, 165 (1998). ↑

  111. . Eleventh Circuit (N=68); Fifth Circuit (N=10); Ninth Circuit (N=9); Second Circuit (N=8); First Circuit (N=8). ↑

  112. . This Article ultimately recommends amending the Statutory Index to sentence 46 U.S.C. § 70503 cases under §2D.1.1. See infra Section IV.C. ↑

  113. . Much of the quantitative data used in this analysis is available to future researchers.
    See Commission Datafiles, U.S. Sent’g Comm’n, https://www.ussc.gov/research/datafiles/commission
    -datafiles [https://perma.cc/SRP2-XNJW]; Integrated Database (IBD), Fed. Jud. Ctr., https://www.fjc
    .gov/research/idb [https://perma.cc/E322-9SGE]; Public Access to Court Electronic Records, U.S. Cts., https://pacer.uscourts.gov/ [https://perma.cc/9K6Z-XLZT]. However, some of the qualitative analysis (requiring review of confidential PSRs, SORs, or plea agreements) is likely only possible by the Sentencing Commission staff or with specific authorization. I had access to this information as a Supreme Court Fellow with a detail at the Sentencing Commission. ↑

  114. . Notably, each of these circuits applies the elements-based approach to §2X5.1. See supra Table 1. Far more §2X5.1 cases occur within elements-based-approach jurisdictions (N=346) compared to courts that have not adopted that approach (N=54). ↑

  115. . For states with more than one federal district, the districts have been combined. ↑

  116. . As set forth in Section II.E, Oklahoma and the District of Columbia saw a sharp increase in cases in 2022 that is not reflected in this table (2016–2021). ↑

  117. . Although jurisdictions refer to this offense differently (e.g., “DWI”), for simplicity this Article will refer to all such offenses as “DUI.” ↑

  118. . 49 U.S.C. § 46306(b)(7). ↑

  119. . 18 U.S.C. §§ 951, 371; 22 U.S.C. §§ 612, 618(a)(1). ↑

  120. . 18 U.S.C. § 401(3). ↑

  121. . Id. § 1924. ↑

  122. . 49 U.S.C. § 46306(b)(7). ↑

  123. . 18 U.S.C. § 157. ↑

  124. . Id. § 231(a)(3). ↑

  125. . Id. § 2101(a). ↑

  126. . 236 of 331 cases. ↑

  127. . Letter from Hon. Ralph R. Erickson, supra note 15, at 1 (the current §2X5.1 guidance “leaves sentencing judges in a position of great uncertainty as to the appropriate base offense level” in these cases). ↑

  128. . See supra note 97. ↑

  129. . See supra Table 1. ↑

  130. . United States v. Allard, 164 F.3d 1146, 1150 (8th Cir. 1999) (citing United States v. Osborne, 164 F.3d 434 (8th Cir. 1999)). ↑

  131. . Neb. Rev. Stat. § 28-707(1)(a) (2024). ↑

  132. . Ariz. Rev. Stat. Ann. § 13-3623(A)–(B) (2024). ↑

  133. . Compare id., with Neb. Rev. Stat. § 28-707(1)(a) (2024). ↑

  134. . S.D. Codified Laws § 26-10-1 (2024) (definitions at § 26-8A-2). ↑

  135. . N.D. Cent. Code § 14-09-22 (2024). ↑

  136. . Id. § 14-09-22.1. ↑

  137. . The single child-endangerment-category case during the review period (hit and run with children in the car, intoxicated offender but not to the point of DUI) was charged under the North Dakota child-neglect statute. ↑

  138. . Cf. Dissenting View of Commissioner Paul H. Robinson on the Promulgation of Sentencing Guidelines by the United States Sentencing Commission, 52 Fed. Reg. 18121, 18126 n.55 (May 13, 1987) (“[L]ess-frequent offenses may be the offenses for which judges have the greatest need for sentencing guidance. The less frequently an offense occurs, the less likely it is that the judge will be familiar with any special sentencing considerations of the offense, and the less likely it is that the judge will know what sentence other judges commonly give. The potential for the disparity may increase dramatically with decrease in frequency.”). ↑

  139. . In a few cases, key conduct was recited in the plea agreement but not in the PSR, and conduct from both was comprehensively considered. ↑

  140. . Other cases of endangerment involved using drugs around a child or while breastfeeding; discharging a firearm near a child; parental knowledge of abuse or unsafe conditions and refraining from action; or mothers becoming intoxicated, falling asleep next to a baby, and the baby being dead when the mother awakes. ↑

  141. . If all courts applied the elements-based approach, we can conclude that where the NAG conduct mirrored the AG conduct, courts were probably incorrect in determining that there were no analogous guidelines. Thirteen of the 14 NAG cases of physical child abuse occurred within circuits that apply the elements-based approach to the application of §2X5.1. Elements-based jurisdictions: Arizona (N=1); Minnesota (N=1); New Mexico (N=6); Oklahoma (N=1); South Dakota (N=4). Other: Wisconsin (N=1). ↑

  142. . The conduct in AG and NAG cases within the physical abuse category was similar. I adjusted for severity in subsequent sections and still found an analogous-guideline gap. ↑

  143. . See U.S. Sent’g Comm’n, Demographic Differences in Federal Sentencing 22 (2023), https://www.ussc.gov/sites/default/files/pdf/research-and-publications/research-publications
    /2023/20231114_Demographic-Differences.pdf [https://perma.cc/6C6Y-ZRTE] (calling the bottom end of the guidelines the “presumptive sentence”); U.S. Sent’g Comm’n, Demographic Differences in Federal Sentencing Practices: An Update of the Booker Report’s Multivariate Regression Analysis app. A, at A-4 to A-5 tbl.A (2010) [hereinafter Demographic Differences in
    Federal Sentencing Practices Update], https://www.ussc.gov/sites/default/files/pdf/research-and
    -publications/research-publications/2010/20100311_Multivariate_Regression_Analysis_Report.pdf [https://perma.cc/QD4P-ZX3C] (most federal sentences are at or below the bottom end of the guidelines). ↑

  144. . Supra note 142. ↑

  145. . U.S. Sent’g Comm’n, Intra-District Differences in Federal Sentencing Practices: Sentencing Practices Across Districts from 2005 – 2017, at 18 (2020), https://www.ussc.gov
    /sites/default/files/pdf/research-and-publications/research-publications/2020/20200122_Inter-District
    -Report.pdf [https://perma.cc/EQ6K-RG6F]; see also U.S. Sent’g Comm’n, supra note 99, at 95 (44.1% of all sentences in FY2021 were at the guideline minimum). At 44.1%, sentences at the guideline minimum were the most common guideline sentence, over twice as common as the next category (lower half of guideline range, 20.3%). Id. ↑

  146. . Demographic Differences in Federal Sentencing Practices Update, supra note 142, at A-1 n.74; see also Amy Farrell, Geoff Ward & Danielle Rousseau, Intersections of Gender and Race in Federal Sentencing: Examining Court Contexts and the Effects of Representative Court Authorities, 14 J. Gender Race & Just. 85, 104 (2010) (applying presumptive sentence). ↑

  147. . Of course, it is possible that a judge in one of the NAG cases might have analogized to a guideline other than assault or aggravated assault, but that would be an outlier. See supra Section II.C.1 (100% of physical-abuse AG cases were analogized to assault or aggravated assault). ↑

  148. . These figures assume a total offense level of 9: 7 base offense level for physical contact, U.S. Sent’g Guidelines Manual §2A2.3(a)(1) (U.S. Sent’g Comm’n 2023), +4 for the victim being under the age of 16, id. §2A2.3(b)(1)(B), and -2 for acceptance of responsibility, id. §3E1.1(a). This was a common calculation in physical-abuse cases when courts analogized to assault. ↑

  149. . These figures assume a total offense level of 16: 14 base offense level, id. § 2A2.2(a), +3 for bodily injury, id. § 2A2.2(b)(3)(A), +2 for a vulnerable victim, id. § 3A1.1(b)(1), and -3 for acceptance of responsibility, id. § 3E1.1. This was a common calculation in physical-abuse cases when courts analogized to aggravated assault. ↑

  150. . See supra note 142 and accompanying text. ↑

  151. . One reason for the level of detail in this Article is to account for the possibility that its recommended transparency changes are not adopted. In that case, interested parties will at least have five years of data in this Article from which to make meaningful comparisons in future cases. ↑

  152. . Of course, the dataset excludes murder or similar convictions (which have guidelines and are not sentenced under §2X5.1). See, e.g., U.S. Sent’g Guidelines Manual §§2A1.1–1.4 (U.S. Sent’g Comm’n 2023). ↑

  153. . The highest sentence in physical-abuse AG cases actually did not involve the death of the child. One mother was sentenced to 180 months for locking two children in a room, burning her malnourished child with scalding water, and failing to subsequently seek medical attention for the injuries. She was high on methamphetamine, did not have hot water, boiled water for the child’s bath, and, apparently not knowing that the water was scalding hot, placed a child in the water. The court analogized the case to aggravated assault, id. §2A2.2, the guideline the parties stipulated to in the plea agreement. Additionally, the 180-month sentence was lower than her first sentence (240 months), which had been reversed for failing to adequately explain the sentence. See United States v. Crow Eagle, 729 F. App’x 503, 503–04 (8th Cir. 2018). Her guideline range was 57–71 months, but the statutory minimum was 120 months. ↑

  154. . It could be argued that the father’s case fits within the child endangerment category, not the physical abuse category. But due to the overlapping facts, insinuations of the father’s abuse in the PSR, and because physical abuse caused the child’s death, the case was kept in the physical abuse category. ↑

  155. . For example, the base offense level for aggravated assault is 14. U.S. Sent’g Guidelines Manual §2A2.2(a) (U.S. Sent’g Comm’n 2023). Five levels might be added for the victim sustaining a serious bodily injury, id. §2A2.2(b)(3)(B), two levels are added for a vulnerable victim, id. §3A1.1(b)(1), and three levels are removed for acceptance of responsibility, id. §3E1.1. The father in the AG case received an additional four-level enhancement for using a dangerous weapon, id. §2A2.2(b)(2)(B), and there was no evidence of the same in the NAG case. ↑

  156. . Id. §2A1.3. ↑

  157. . Id. §3E1.1. ↑

  158. . See id. §2A1.2 (base offense level for second degree murder is 38). After three levels are removed for acceptance of responsibility, id. §3E1.1, to arrive at 35, this results in a guideline range of 186–210 months. ↑

  159. . See, e.g., Sonja B. Starr & M. Marit Rehavi, Mandatory Sentencing and Racial Disparity: Assessing the Role of Prosecutors and the Effects of Booker, 123 Yale L.J. 2, 17 (2013) (“Studies in systems without guidelines . . . control for conviction severity.”). To analyze disparity in the system writ large, Starr and Rehavi favor controlling for possible disparity in pre-sentence stages. They therefore include in their analysis arrest data, prosecution files, and court records in addition to sentencing information. Id. at 24. However, the purpose of this Article is narrower; it does not render observations and conclusions on the system as a whole, it focuses only on judicial discretion and the sentencing guidelines. ↑

  160. . See, e.g., Shawn D. Bushway & Anne Morrison Piehl, Judging Judicial Discretion: Legal Factors and Racial Discrimination in Sentencing, 35 Law & Soc’y Rev. 733, 738 (2001) (“The study of unwarranted disparity in sentence outcomes was first implemented in jurisdictions with indeterminate sentences and later in jurisdictions that had adopted determinate sentencing structures, primarily sentencing guidelines. . . . [T]hese models always include measures of offense severity and criminal history of the defendant . . . .”); Darrell Steffensmeier, John Kramer & Cathy Streifel, Gender and Imprisonment Decisions, 31 Criminology 411, 416 (1993) (This article criticized two prior approaches that failed to offer “a rigorous control for offense gravity because the offense categories typically are graded statutorily into two or more levels of offense seriousness. This failure to differentiate grades of severity within an offense category or grouping may easily confound gender-based comparisons in sentencing outcomes because female offenders tend to commit the less serious forms of crime within the broad category.”). ↑

  161. . See Starr & Rehavi, supra note 158, at 15 (noting limited importance of statutory minimums and maximums in disparity analyses as a result of prosecutorial power in charge bargaining). ↑

  162. . See Demographic Differences in Federal Sentencing Practices Update, supra note 142, at A-4 to A-5 tbl.A. ↑

  163. . For example, +7 if there was permanent or life-threatening bodily injury, +5 if a firearm was discharged, +5 if there was serious bodily injury, and +3 if there was a bodily injury. See U.S. Sent’g Guidelines Manual §2A2.2(b) (U.S. Sent’g Comm’n 2023). ↑

  164. . See supra note 159. ↑

  165. . See Minn. Sent’g Guidelines Comm’n, Minnesota Sentencing Guidelines and Commentary section II.A.05 (Aug. 1, 1985), https://mn.gov/sentencing-guidelines/assets/1985
    -Sentencing%20Guidelines_tcm30-31771.pdf?sourcePage=/sentencing-guidelines/guidelines/archive
    .jsp%3Fid=30-32040 [https://perma.cc/6D9Q-LGAV] (explaining that “judges should exercise their discretion by assigning an offense a severity level which they believe to be appropriate”). This is the 1985 version of this guideline commentary, but the 2023 version still instructs the sentencing judge to “assign an appropriate severity level for the offense and specify on the record why that particular level was assigned.” Minn. Sent’g Guidelines Comm’n, Minnesota Sentencing Guidelines and Commentary section 2.A.4 (Aug. 1, 2023), https://mn.gov/sentencing-guidelines/assets/1August2023
    MinnSentencingGuidelinesCommentary_tcm30-586295.pdf [https://perma.cc/H7UR-UXH6]. ↑

  166. . See Newton & Sidhu, supra note 31, at 1197 (“Before drafting sentencing guidelines, the original [Sentencing] Commission engaged in an extensive study of the existing state guidelines (in particular, the guidelines in Minnesota, Pennsylvania, and Washington), the federal parole guidelines, as well as a large body of criminological literature relevant to the issues that the SRA required the [Sentencing] Commission to address in creating guidelines (such as offender and offense characteristics.” (footnotes omitted)). ↑

  167. . Marvin E. Wolfgang, Robert M. Figlio, Paul E. Tracy & Simon I. Singer, U.S. Dep’t of Justice, The National Survey of Crime Severity (1985) (asking sixty thousand adults to rank the severity of offenses ranging from playing hooky from school to planting a bomb that kills twenty people). ↑

  168. . In a few cases, the offender admitted to the elements of the offense but not to specifically identified conduct, and none of the sentencing documents described the conduct. ↑

  169. . For additional reference, Wisconsin has some of the most detailed child-abuse statutes. Chapter 948 of the Wisconsin Code provides different subsections for varying conduct, including intentional physical abuse, reckless causation of bodily harm, failing to prevent bodily harm, neglect, child sex crimes, and repeated instances of these and other forms child abuse. Wis. Stat. § 948 (2024). ↑

  170. . Stipulated probation in a Rule 11(c)(1)(C) plea agreement. Fed. R. Crim. P. 11(c)(1)(C). ↑

  171. . Supra note 169. ↑

  172. . See 28 U.S.C. § 994(b)(2). ↑

  173. . Id. ↑

  174. . See Fed. R. Crim. P. 11(c)(3) (judicial consideration of plea agreement). ↑

  175. . See Merchant, supra note 47 (analyzing use of plea agreements to potentially circumvent statutory and Guideline ranges). ↑

  176. . See Part IV. ↑

  177. . The original indictment in the case charged first-degree murder and aggravated child battery along with child abuse, with murder and child battery dismissed as a part of the plea process. ↑

  178. . In two of the AG cases, a child suffocated, but there was no evidence of the cause. Both parents received probation. ↑

  179. . A surprising number of the child-death-NAG cases (4 of 5) involved a similar, negligence-based fact pattern: an intoxicated mother fell asleep next to her child and the child died overnight. Sentences in these cases (all offenders were CHC-1) were 5 months (credit for 3 months’ time served), 8 months (times served), 12 months, and 14 months. ↑

  180. . This category includes two cases in which the defendant’s intoxication level was not proven, but facts about the defendants’ intoxicated driving played prominently in the PSRs (and therefore likely played a role in sentencing). ↑

  181. . For example, one offender coached a 16-year-old boy on how to sexually abuse his 10-year-old sister, and different offender (apparently suffering a psychotic episode) wandered the highway naked and ultimately walked into a lake with her baby (both were fortunately saved by passersby). ↑

  182. . A final note on child-abuse cases, it should be noted that many of the offenses were committed on tribal reservations, and prior tribal convictions are excluded from an offender’s criminal history calculation. U.S. Sent’g Guidelines Manual §4A1.2(i) (U.S. Sent’g Comm’n 2023) (“Sentences resulting from tribal court convictions are not counted . . . .”). Thus, one might argue that sentences in child abuse cases occurring on tribal lands might be longer due to this exclusion, and due to judges considering these prior convictions when imposing longer sentences. However, the guidelines expressly provide for an upward departure based on the inadequacy of criminal history category based on the exclusion of tribal sentences. Id. §4A1.3(a)(1)–(2)(A); id. Application Note (C). No case in the dataset cited this reason for a departure. If this is occurring, it is unstated (and should be explicit in the SOR). ↑

  183. . In the DUI NAG case that involved injury to a victim, a son was intoxicated, hit his father with a lead pipe during a fight, and then entered the car in the garage. Officers convinced him to exit the car and he was arrested. Through the plea process, he pleaded guilty to felony DUI. Although this case is qualitatively different from the other DUI NAG cases, because the defendant pleaded guilty to DUI, it was included in the analysis. The defendant (CHC-3) received a sentence of 6 months. ↑

  184. . Interestingly, in both DUI AG and child-endangerment-impaired-driving AG cases, judges (or parties in plea agreements) seemed to analogize the offenses to a guideline if a victim was injured, but not if there was no injury. But under the elements-based approach, the decision turns on the elements of the offenses, not whether a victim was injured (and injury is not an element of these offenses). ↑

  185. . This figure excludes one sentence in a NAG case of 20 days’ home confinement, which is considered “probation” for the purpose of this section. ↑

  186. . See supra Section II.C.1. ↑

  187. . The number of DUIs was derived from the number that triggered the offense of conviction. In rare cases, DUIs existed on the defendant’s criminal history (in the PSR) but were not considered a prior conviction, perhaps based on an applicable state law. This table only includes prior DUIs that were considered for purposes of the statute of conviction (even though a judge might consider other DUIs in when fashioning an appropriate sentence under 18 U.S.C. § 3553).

    Unfortunately, in approximately half of the DUI NAG cases (10 of 23, mostly in the Western District of Texas), the PSR did not calculate the criminal history categories, perhaps viewing the exercise as futile since the PSR considered the offenses to have no guidelines and therefore had no total offense level. As discussed in Section IV.B, probation officers should still calculate the criminal history categories even in NAG cases. ↑

  188. . Unlike other assimilated crimes that involve mostly tribal defendants, these DUI cases occurred within other federal enclaves and had more racial diversity among defendants. ↑

  189. . These offenders included a Caucasian male and female and a Hispanic male. ↑

  190. . This offender’s CHC was not calculated in the PSR. See infra note 247 and accompanying text. However, a review of the criminal history in the PSR reveals that he likely would have been CHC-3. ↑

  191. . Again, the probation officer did not calculate the CHC in this case, but it likely would have been CHC-3. ↑

  192. . It is worth noting that this offender’s criminal history notes a prior state sentence of 12 years for a DUI, and he appears to have served around half of that sentence. Although it is speculation, perhaps the sentencing judge wanted to impose a sentence that was longer than the offender’s prior DUI sentences. But other comparable DUI offenders had similarly long prior sentences. ↑

  193. . See, e.g., N.M. Stat. Ann. § 66-8-102 (2016) (deferred sentences availability, requirement of substance-abuse-treatment program, and requirement of ignition-interlock device); id. § 31-16A (New Mexico’s “Preprosecution Diversion” program). ↑

  194. . State alternatives cannot currently be assimilated. See United States v. Sharpnack, 355 U.S. 286, 292–93 (1958) (only substantive state law is assimilated, and state law is not assimilated if it conflicts with federal law); United States v. Pate, 321 F.3d 1373, 1376 (11th Cir. 2003) (“[T]he ‘like punishment’ clause of the ACA does not require a federal court to implement state policies regarding eligibility for early release and alternative forms of confinement . . . .”); United States v. Pinto, 755 F.2d 150 (10th Cir. 1985) (state parole requirements not assimilated). ↑

  195. . See supra notes 117–24. ↑

  196. . 18 U.S.C. § 157. ↑

  197. . In two of these cases, the primary statute of conviction was for tax evasion, with accompanying convictions for bankruptcy fraud. ↑

  198. . E.g., 18 U.S.C. § 951; 22 U.S.C. § 612. ↑

  199. . 49 U.S.C. § 46306. ↑

  200. . Prosecutors in the January 6, 2021, “Capitol Breach” cases sometimes obtained convictions for civil disorder, but typically brought federal charges under statutes that have guidelines, like 18 U.S.C. § 1752(a)(l) (Knowingly Entering or Remaining in any Restricted Building or Grounds Without Lawful Authority), 18 U.S.C. § l752(a)(2) (Knowingly Engaging in Disorderly or Disruptive Conduct in Restricted Building), and 40 U.S.C. § 5104(e)(2) (Violent Entry and Disorderly Conduct on Capitol Grounds). See Sentences Imposed in Cases Arising Out of the Events of January 6, 2021, U.S. Att’y’s Off. D.C. (Jan. 6, 2025), https://www.justice.gov/usao-dc/capitol-breach-cases [https://perma.cc/BSE8-788W]. All have base offense levels between 4 and 10. U.S. Sent’g Guidelines Manual app. A (U.S. Sent’g Comm’n 2023).

    However, the most severe of these convictions, seditious conspiracy (18 U.S.C. § 2384), has no analogous guideline and therefore is sentenced under §2X5.1. See id. The Sentencing Commission likely declined to assign a guideline to seditious conspiracy in the Statutory Index because of the wide variety of conduct that can support a conviction in these cases. (Similar to contempt, see supra note 69.) The few courts to have addressed it have analogized seditious conspiracy to §2M1.1 (Treason), putting the offender in the highest offense level in the guidelines—43. See United States v. Rahman, 189 F.3d 88, 151–52 (2d Cir. 1999); see also United States v. Ford, 216 F. App’x 652 (9th Cir. 2007) (unpublished). ↑

  201. . Three of the four of the NAG sentences were probationary. ↑

  202. . But see infra Section II.E (describing the increasing number of civil disorder convictions after the January 6, 2021, Capitol Breach). ↑

  203. . For example, for offenders convicted of similar conduct during George Floyd protests, some received probation and others received years in prison. ↑

  204. . Most of the remaining Fed. Stat. offenses are rare or unusual crimes. E.g., 18 U.S.C. § 2339B (providing material support or resources to designated foreign terrorist organizations, analogized to §2M5.3 (Unlawful Activity with Nuclear Material)); 30 U.S.C. § 820 (false statements by operator of a mine, analogized to §2B1.1 (Larceny)); 12 U.S.C. § 1715z–19 (equity skimming, analogized to §2B1.1 (Larceny)). ↑

  205. . See, e.g., Judicial Business 2021, U.S. Cts., https://www.uscourts.gov/statistics-reports
    /judicial-business-2021 [https://perma.cc/UF2T-MXV7] (“In the 12-month period ending September 30, 2021, the workload in most areas of the federal Judiciary was affected by the [COVID-19] pandemic.”). ↑

  206. . For example, 300, 240, and 120-month sentences when the calculated guideline ranges were 84–105, 51–63, and 46–57 months, respectively; and probation or time served when the guideline ranges were 33–41 and 15–21 months. ↑

  207. . For example, some offenders received probation and others over a year in prison for similar spankings with a belt. And one sentence eclipsed all others, including all of those in the primary dataset. That case, described in this Article’s introduction, involved a guilty jury verdict for harsh physical beatings, though not near as severe as some of the other conduct from the dataset. The offender was sentenced to 300 months after the judge granted the government’s novel motion to depart or vary upward and treat the conduct as he would treat a conviction under 18 U.S.C § 3559(f)(2), which carries a mandatory minimum sentence of 25 years for convictions under 18 U.S.C. § 114 (throwing acid or intentionally torturing, maiming, or disfiguring by cutting, biting or slitting an eye, lip, tongue, or limb). The sentence was affirmed on appeal. United States v. Walker, 74 F.4th 1163 (10th Cir. 2023); cf. supra note 109 and accompanying text (if conduct substantially overlaps with conduct punished by a federal statute, the government should charge the federal offense, not the state offense).

    Consistent with the severity scale set forth in Section II.C.1.a.ii, this offense would likely be at a level 4 or 5, and AG and NAG sentences within those levels were between probation and 180 months. Even on the higher end of this range, 300 months is 10 years longer than the longest previous sentence for similar conduct. The mandatory minimum in 18 U.S.C § 3559(f)(2) does not appear to accurately reflect the severity of these offenses. ↑

  208. . 591 U.S. 894 (2020). In McGirt, the Supreme Court determined that the state of Oklahoma did not have the authority to prosecute violent crimes by or against Native Americans that occurred on certain tribal lands, and that only the federal and tribal governments had jurisdiction to prosecute those offenses. Id. In Oklahoma v. Castro-Huerta, the Court walked back its holding in McGirt to some extent, carving out non-tribal defendants. 597 U.S. 629 (2022). After Castro-Huerta, Oklahoma has concurrent jurisdiction with federal and tribal governments to prosecute non-tribal offenders even if the victim is a tribal member, and even on tribal lands. See id. Setting aside distinctions between major and minor crimes, the general effect of this is that federal and tribal governments have sole jurisdiction to prosecute offenses by tribal offenders (regardless of the status of the victim), and the state has concurrent jurisdiction to prosecute offenses committed by non-tribal offenders (regardless of the status of the victim). ↑

  209. . Most of the tribal land in Oklahoma that is impacted by McGirt and Castro-Huerta is in the eastern half of the state, meaning the decisions directly impact the Eastern and Northern Districts of Oklahoma more than the Western District of Oklahoma. See McGirt, 591 U.S. at 928; Castro-Huerta, 597 U.S. at 634. ↑

  210. . See U.S. District Courts–Combined Civil and Criminal Federal Court Management Statistics, U.S. Cts. (June 30, 2022), https://www.uscourts.gov/file/45252/download [https://perma.cc
    /U4D3-URPR]. ↑

  211. . Castro-Huerta, 597 U.S. 629 (effectively reducing the number for McGirt prosecutions in Oklahoma by modifying state and federal jurisdiction over crimes committed on tribal lands). ↑

  212. . See supra note 209. Based on the data referenced above, 69% of the 2021 and 2022 Oklahoma cases (27 of 39) were child abuse or endangerment cases. On courts being at capacity, see Curtis Killman, Feds Decline More than 5,800 Criminal Cases Since McGirt Ruling, Tulsa World (June 5, 2022), https://tulsaworld.com/news/state-and-regional/crime-and-courts/feds-decline-more
    -than-5-800-criminal-cases-since-mcgirt-ruling/article_0cf8aa3e-dd0a-11ec-ab20-737a4fd2f591.html [https://perma.cc/3RVQ-TPRK] (notably, this Tulsa World article counts misdemeanor offenses). ↑

  213. . In 2021, there were 39 assimilated §2X5.1 cases, 19 from Oklahoma. In 2022, there were 45 total cases, 20 from Oklahoma. ↑

  214. . A vast majority of Capitol Breach cases charged offenders with offenses that have guidelines. Supra note 199. The 2022 §2X5.1 Capitol Breach sentences were generally for convictions for several offenses, and for which the conviction for civil disorder (18 U.S.C. § 231(a)(3)) resulted in a higher offense level and therefore drove the sentence. President Donald Trump pardoned or commuted all Capitol Breach offenders after his reelection, but many of the sentences are analyzed in Merchant, supra note 25. ↑

  215. . Excluding the most extreme outlier, the average AG sentence was lower, 8.11 months (15 months P.E.), and the median sentence was 3 months (8 P.E.). ↑

  216. . The other types of offenses were not compared to guideline minimums because there were no consistent analogous guidelines that would cover those offenses. (For example, there is likely no analogous guideline for a non-injury felony DUI offense.) ↑

  217. . See, e.g., Sentencing Guidelines for United States Courts, 88 Fed. Reg. 7180, 7211 (proposed Feb. 2, 2023) (“The term ‘controlled substance offense’ means . . . [a]ny offense under state law (or the offense of aiding or abetting, or conspiring, soliciting, or attempting to commit any such offense), punishable by imprisonment for a term exceeding one year, for which the most appropriate guideline would have been one of the Chapter Two guidelines listed in paragraph (2) had the defendant been sentenced under the guidelines in federal court (as determined under subsection (c)).” (emphasis added)). ↑

  218. . See supra note 27. ↑

  219. . See Kahneman et al., supra note 21, at 49–50. ↑

  220. . Gall v. United States, 552 U.S. 38, 51 (2007). ↑

  221. . See id. at 49; United States v. Leahy, 169 F.3d 433, 441 (7th Cir. 1999) (correct selection of “most ‘analogous’ guideline still serves an important function” by providing “the starting or reference point within the structure of the Guidelines” and serving as a reference point for departures or variances). ↑

  222. . See U.S. Sent’g Guidelines Manual §3E1.1 (U.S. Sent’g Comm’n 2023). ↑

  223. . See Merchant, supra note 47 (arguing that the justice system might rely on several embedded fictions, including that a guideline range accurately matches the actual offense conduct); Thea Johnson, Fictional Pleas, 94 Ind. L.J. 855, 857 (2019) (“[A] fictional plea [is] . . . a plea bargain agreement in which the defendant pleads guilty to a crime he did not commit, with the consent and knowledge of multiple actors in the criminal justice system—to avoid the profound collateral consequences that would flow from a conviction on his initial charge.”); Ilene H. Nagel & Stephen J. Schulhofer, A Tale of Three Cities: An Empirical Study of Charging and Bargaining Practices Under the Federal Sentencing Guidelines, 66 S. Cal. L. Rev. 501 (1992) (analyzing circumvention of guidelines in plea bargaining); see also Jeffrey Bellin & Jenia I. Turner, Sentencing in an Era of Plea Bargains, 102 N.C. L. Rev. 179, 215–19 (2023) (empirically examining sentences imposed based on different types of plea agreements). ↑

  224. . John Gleeson, Sentence Bargaining Under the Guidelines, 8 Fed. Sent’g Rep. 314, 314 (1996) (“But beyond that obvious point is an equally important one: If a set of legal rules causes otherwise law-abiding attorneys to engage in systemic deception in order to achieve what they believe is a just result, attention ought to be paid to the rules themselves, and to whether they should be modified.”). ↑

  225. . 28 U.S.C. § 994(o). ↑

  226. . Id. § 995(a)(12)(A). ↑

  227. . See id. § 994(w)(1). ↑

  228. . Final Priorities for Amendment Cycle, 87 Fed. Reg. 67756, 67756 (Nov. 9, 2022) (“In light of Commission studies, consideration of possible amendments to the Guidelines Manual relating to criminal history to address (A) the impact of ‘status’ points under subsection (d) of section 4A1.1 (Criminal History Category); (B) the treatment of defendants with zero criminal history points; and (C) the impact of simple possession of marihuana offenses.”); see also Amendments to the Sentencing Guidelines, U.S. Sent’g Comm’n (2023), https://www.ussc.gov/sites/default/files/pdf/amendment
    -process/official-text-amendments/202305_Amendments.pdf [https://perma.cc/734S-326F%5D. ↑

  229. . 2023 Criminal History Data Briefing, U.S. Sent’g Comm’n, at 22:00–23:37 (Jan. 23, 2023), https://www.ussc.gov/education/videos/criminal-history-data-briefing [https://perma.cc/AW63-8D7H]. ↑

  230. . Id. at 22:35–23:37. ↑

  231. . United States v. Leahy, 169 F.3d 433, 441 (7th Cir. 1999). ↑

  232. . United States v. Allard, 164 F.3d 1146, 1150 (8th Cir. 1999) (citing United States v. Osborne, 164 F.3d 434 (8th Cir. 1999)). ↑

  233. . See Stith & Cabranes, supra note 17, at 169 (“The federal Sentencing Guidelines of today are based on a fear of judging; they attempt to repress the exercise of informed discretion by judges.”). Of course, Fear of Judging was written when the guidelines were mandatory, not advisory. See United States v. Booker, 543 U.S. 220 (2005). Post-Booker, “[t]he abject fear of judging has abated considerably.” Kate Stith, The Arc of the Pendulum: Judges, Prosecutors, and the Exercise of Discretion, 117 Yale L.J. 1420, 1495 (2008). ↑

  234. . Some judges have suggested that federal courts should be able to consider similar sentences in state cases when a federal judge sentences assimilated offenses, with the goal of avoiding federal/state disparities in sentencing. E.g., United States v. Lasley, 832 F.3d 910, 920 (8th Cir. 2016) (Bright, J., dissenting) (“Federal courts must take into account sentences and actual time served for similar state-law crimes when sentencing under the Major Crimes Act to ensure Indians in Indian Country do not receive drastically increased penalties for crimes traditionally handled in state or tribal court.”); United States v. Clark, 434 F.3d 684, 688 (4th Cir. 2006) (Motz, J., concurring) (there are “some cases in which consideration of state sentences will not conflict with § 3553(a)(6)’s mandate to ‘avoid unwarranted sentence disparities’ and may in fact help courts to apply correctly the other factors set forth in § 3553(a). One example is when a federal criminal statute incorporates state law. See, e.g., 18 U.S.C.[] § 13(a).”). Consideration of state sentencing policy is generally prohibited in federal sentencing. Courts have noted the tension between the SRA’s goal of federal sentencing uniformity, and the ACA’s goal of intrastate sentencing uniformity. See, e.g., United States v. Garcia, 893 F.2d 250, 253 (10th Cir. 1989), superseded by statute, 18 U.S.C. § 3551(a), as recognized in United States v. Martinez, 1 F.4th 788 (10th Cir. 2021); Hutchison et al., supra note 66, § 1B1.1 n.317. Assimilated offenses could be an exception to the general rule, though that subject is better suited for a future article. ↑

  235. . E.g., Joshua M. Divine, Booker Disparity and Data-Driven Sentencing, 69 Hastings L.J. 771, 795 (2018). ↑

  236. . See 28 U.S.C. § 994(b)(2). ↑

  237. . See Letter from Judge Myron H. Bright, Cir. J., 8th Cir., to Mythili Raman, Acting Assistant Att’y Gen., Dep’t of Just. (Feb. 27, 2014), in U.S. Sent’g Comm’n, Tribal Issues Advisory Grp., Report of the Tribal Issues Advisory Group, app. D, at 3 (2016), https://www.ussc.gov
    /sites/default/files/pdf/research-and-publications/research-publications/2016/20160606_TIAG-Report
    .pdf [https://perma.cc/T988-NK77] (“[S]entencing in Indian Country is simply one of the most difficult things that we are called to do as sentencing judges. A number of factors coalesce to render it nearly impossible to sentence in a just manner.” (quoting Judge Ralph R. Erickson)). ↑

  238. . See Ryan W. Scott, The Skeptic’s Guide to Information Sharing at Sentencing, 1 Utah L. Rev. 345, 396 (2013) (self-described information-sharing skeptic, nonetheless suggesting that “information sharing can contribute to inter-judge consistency and rationality under the right conditions”). ↑

  239. . Sometimes the prosecutor does not include this information in the charging document. In such cases, the judge can confirm the subsection on the record with the parties and then include it in the J&C. ↑

  240. . Stith & Cabranes, supra note 17, at 79. ↑

  241. . William H. Pryor Jr., The Integral Role of Federal Probation Officers in the Guidelines System, 81 Fed. Prob. 13, 15, 17 (2017) (quoting Douglas A. Berman, Is Fact Bargaining Undermining the Sentencing Guidelines?, 8 Fed. Sent’g Rep. 300, 301 (1996)). ↑

  242. . See U.S. Prob. and Pretrial Servs., Charter for Excellence (2002) (probation officers “[p]rovid[e] objective investigations and reports with verified information and recommendations to assist the court in making fair pretrial release, sentencing, and supervision decisions”). ↑

  243. . Pryor, supra note 240, at 15. The PSR also serves as the referral to the Federal Bureau of Prisons, and inaccuracies can have an impact on prison placement, processing, and programming. See Fed. R. Crim. P. 32(i)(3)(C) (court must make available to Bureau of Prisons its sentencing determinations and PSR). ↑

  244. . 18 U.S.C. § 3552(a). ↑

  245. . The probation officer should, of course, identify any stipulations, but those stipulations should not bind the officer’s independent conclusions. In some cases the officer might elect (and some judges in my experience might request), that the officer draft two calculations—one based on the parties’ stipulations and one based on the probation officer’s. The more relevant information the judge has in advance, the better. ↑

  246. . Counsel, too, should make efforts to identify sentences imposed on similar offenders for similar offenses. ↑

  247. . Cf. United States v. McDaniel, 59 F.4th 975 (8th Cir. 2023) (affirming use of the Sentencing Commission’s online JSIN database even though reliance was not previously disclosed to the parties). ↑

  248. . See supra notes 189–90 and accompanying text. ↑

  249. . See generally Nancy Gertner, Apprendi/Booker and Anemic Appellate Review, 99 N.C. L. Rev. 1369, 1372 (2021) (supporting a common law of sentencing, “a judge-made sentencing framework supplementing, and perhaps supplanting, the Guidelines”). ↑

  250. . Stith & Cabranes, supra note 17, at 78. ↑

  251. . Of note, JSIN requires a minimum threshold number of cases with the same criminal-history category and offense level during the prior five fiscal years or the database will not return information. Section 2X5.1 cases will not often meet these criteria. ↑

  252. . In this way, assimilated offenses can be isolated from convictions under a federal statute. ↑

  253. . This key change allows future inquiries into convictions for a specific state statute. But because certain states have a small number of convictions for a specific statute, coding the type of offense is also critical (e.g., DUI, child abuse and type, etc.) to allow comparisons across states. ↑

  254. . This kind of grouping is important so that analysts need not search for every relevant state statute in order to search for and compare these sentences, as is currently required. One can search all “DUI” cases, for example. ↑

  255. . Based on the incomplete coding, these cases are also not accurately tracked in the Sourcebook and other Sentencing Commission reports. See, e.g., U.S. Sent’g Comm’n, supra note 99, at 71 (reporting zero cases sentenced with §2X5.1 as the primary guideline in FY2021). ↑

  256. . The background commentary instructs courts to determine if there is a sufficiently analogous guideline and then determine whether there is a most analogous guideline. See U.S. Sent’g Guidelines Manual §2X5.1 cmt. background (U.S. Sent’g Comm’n 2023). For clarity, this commentary should probably be amended to clarify that the first step is to determine whether there are any analogous guidelines, and then if there are many, determine which is the most analogous. ↑

  257. . 46 U.S.C. § 70503. ↑

  258. . 18 U.S.C. § 157. ↑

  259. . Alternatively, there is precedent for providing the most analogous guideline in the application notes. See, e.g., U.S. Sent’g Guidelines Manual §2J1.1 cmt. 2 (U.S. Sent’g Comm’n 2023) (“For offenses involving the willful failure to pay court-ordered child support (violations of 18 U.S.C. § 228), the most analogous guideline is §2B1.1 (Theft, Property Destruction, and Fraud).”); see also United States v. Molak, 276 F.3d 45, 49 (1st Cir. 2002) (noting the same). ↑

  260. . 28 U.S.C. § 995(a)(8), (13)–(14) (The Commission has the power to “request such information, data, and reports from any Federal agency or judicial officer,” to “collect systematically the data obtained from studies, research, and the empirical experience of public and private agencies concerning the sentencing process,” and to “publish data concerning the sentencing process.”). ↑

  261. . Id. § 994(o) (“The Commission periodically shall review and revise, in consideration of comments and data coming to its attention, the guidelines promulgated pursuant to the provisions of this section.”). ↑

  262. . Id. § 997 (“The Commission shall report annually to the Judicial Conference of the United States, the Congress, and the President of the United States on the activities of the Commission.”). ↑

  263. . Id. § 991(b)(1)(B). ↑

  264. . There is reason to be optimistic that changes might be forthcoming. These findings were presented to the Tribal Issues Advisory Group, and the Group promptly acted on them. See Letter from Hon. Ralph R. Erickson, Chair, Tribal Issues Advisory Grp., to Hon. Carlton W. Reeves, Chair, U.S. Sent’g Comm’n (July 26, 2023), in U.S. Sent’g Comm’n, 2023-2024 Amendment Cycle: Public Comment on Proposed Priorities 88 FR 39907, at 107 (2023), https://www.ussc.gov/sites/default
    /files/pdf/amendment-process/public-comment/202308/88FR39907_public-comment_R.pdf [https://
    perma.cc/HR8E-52QV]. ↑

  265. . See U.S. Sent’g Comm’n, 2022 Annual Report and Sourcebook of Federal Sentencing Statistics 72–73 (2022) (in FY 2022, only 0.6% of cases (N=347) involved a vulnerable victim and 0.4% (N=223) involved use of a minor to commit a crime). ↑

  266. . See Adi Leibovitch, Punishing on a Curve, 111 Nw. U. L. Rev. 1205 (2017) (finding disparity based on the happenstance of judges’ varying dockets and familiarity with the subject matter). ↑

  267. . Conference Summary, Conference on the Federal Sentencing Guidelines, 101 Yale L.J. 2053, 2073 (1992) (comments of Judge Jon O. Newman, 2d Cir.). ↑

  268. . See, e.g., Mark Osler & Mark W. Bennett, A “Holocaust in Slow Motion?” America’s Mass Incarceration and the Role of Discretion, 7 DePaul J. for Soc. Just. 117, 139 (2014) (“[H]arshness in sentencing became the Sentencing Commission’s oxygen and mantra.”); Frank O. Bowman, III, The Failure of the Federal Sentencing Guidelines: A Structural Analysis, 105 Colum. L. Rev. 1315, 1328 (2005) (“[B]y any standard the severity and frequency of punishment imposed by the federal criminal process during the guidelines era is markedly greater than it had been before.”). ↑

  269. . See supra Section II.C.1 (judges sentence at or below the guideline minimum most of the time). ↑

  270. . See Amy Baron-Evans & Kate Stith, Booker Rules, 160 U. Pa. L. Rev. 1631, 1678–79 (2012) (noting that the government either moves for or does not oppose below-guideline sentences in a majority of cases). ↑

  271. . Ariz. Rev. Stat. Ann. §§ 13-3623(A)(2), (B)(3) (2024). ↑

  272. . U.S. Sent’g Guidelines Manual §2A2.2 (U.S. Sent’g Comm’n 2023). ↑

  273. . Id. §2A2.3. ↑

  274. . Id. §2A1.4. ↑

  275. . Mich. Comp. Laws § 750.136 (2024). ↑

  276. . U.S. Sent’g Guidelines Manual §2A2.3 (U.S. Sent’g Comm’n 2023). ↑

  277. . Minn. Stat. § 609.378(1)(b)(1) (2023). ↑

  278. . Mont. Code Ann. § 45-5-628 (2023). ↑

  279. . U.S. Sent’g Guidelines Manual §2A2.2 (U.S. Sent’g Comm’n 2023). ↑

  280. . Neb. Rev. Stat. § 28-707(1) (2024). ↑

  281. . U.S. Sent’g Guidelines Manual §2A4.1 (U.S. Sent’g Comm’n 2023). ↑

  282. . N.M. Stat. Ann. § 30-6-1 (2024). ↑

  283. . U.S. Sent’g Guidelines Manual §2A2.2 (U.S. Sent’g Comm’n 2023). ↑

  284. . N.C. Gen. Stat. § 14-318.4(a5) (2024). ↑

  285. . U.S. Sent’g Guidelines Manual §2A2.2 (U.S. Sent’g Comm’n 2023). ↑

  286. . N.D. Cent. Code §§ 14-09-22.1, 14-09-22.2 (2024). ↑

  287. . U.S. Sent’g Guidelines Manual §2A2.3 (U.S. Sent’g Comm’n 2023). ↑

  288. . Okla. Stat. tit. 21, §§ 843.5(C), 852.1(A)(4) (2024); Okla. Stat. tit. 47, § 11-902(A)(5) (2024). ↑

  289. . U.S. Sent’g Guidelines Manual §2A2.2 (U.S. Sent’g Comm’n 2023). ↑

  290. . S.D. Codified Laws § 26-10-1 (2024). ↑

  291. . U.S. Sent’g Guidelines Manual §2A2.2 (U.S. Sent’g Comm’n 2023). ↑

  292. . Id. §2A2.3. ↑

  293. . Tex. Penal Code Ann. §§ 22.041(c-1), (f); 12.35; 22.04(a) (2024). ↑

  294. . U.S. Sent’g Guidelines Manual §2A2.2 (U.S. Sent’g Comm’n 2023). ↑

  295. . Id. §2A2.3. ↑

  296. . Wash. Rev. Code § 9A.42.030 (2024). ↑

  297. . U.S. Sent’g Guidelines Manual §2A2.3 (U.S. Sent’g Comm’n 2023). ↑

  298. . Id. §2G1.3. ↑

  299. . Wis. Stat. §§ 948.21(2)(g), (3)(e)(1) (2024). ↑

  300. . U.S. Sent’g Guidelines Manual §2A2.2 (U.S. Sent’g Comm’n 2023). ↑

  301. . Id. §2A2.3. ↑

  302. . Id. §2A1.4. ↑

Exit mobile version