7/29/26: Case on the third point for acceptance and uncharged California wobblers

 In United States v. Yates, --- F.4th ---, No. 25-2987 (9th Cir. 2026), the Court affirmed the sentence imposed on Darneko Yates following a stipulated-facts bench trial at which the district court found Yates guilty of possessing a firearm as a felon in violation of 18 U.S.C. § 922(g)(1).

On appeal, Yates challenged the government's decision not to move for an extra one-point reduction under U.S.S.G. § 3E1.1(b) because the government had to allocate its resources to preparing for a bench trial.  He also argued that the district court erred in applying a four-level enhancement under U.S.S.G. § 2K2.1(b)(7)(B) for possessing the firearm in connection with another felony offense—child endangerment under California Penal Code § 273a(a).

The Ninth Circuit rejected both claims. 

The government was within its discretion to decline to move for an additional one-level reduction under § 3E1.1(b). Because we have yet to substantively interpret § 3E1.1(b) following amendments to that guideline in 2013 and 2023, we do so here and affirm that our precedent remains good law.

Section 3E1.1(a) provides that if a “defendant clearly demonstrates acceptance of responsibility for his offense,” the district court should “decrease the offense level by 2 levels.” Further, “upon motion of the government stating that the defendant has assisted authorities in the investigation or prosecution of his own misconduct by timely notifying authorities of his intention to enter a plea of guilty, thereby permitting the government to avoid preparing for trial and permitting the government and the court to allocate their resources efficiently,” the court should “decrease the offense level by 1 additional level.”

“With authority over the third acceptance-ofresponsibility point now firmly in the government’s hands, courts were faced with questions about the scope of the prosecutor’s discretion.” Orona, 118 F.4th at 865. We held in Espinoza-Cano that the government “cannot refuse to file a motion on the basis of an unconstitutional motive (e.g., racial discrimination) or arbitrarily.”

Espinoza-Cano reaffirmed that “there is nothing improper about the government providing an incentive to plea bargain.” Id. at 1138. “Even if [this court] were to equate a stipulated bench trial with an outright guilty plea, the government still would not be compelled to file” a § 3E1.1(b) motion because there are “rational reasons for the government to prefer an outright guilty plea to a stipulated bench trial”—including not “expending resources[,] [and] anticipating, and ultimately defending, a complete appeal.” Id. (cleaned up). In short, “the PROTECT Act makes the third level reduction subject to the discretion of the government, rather than mandatory, upon a timely plea of guilty.”

A few years later, we held in Johnson that “the allocation and expenditure of prosecutorial resources for the purposes of defending an appeal is a rational basis for declining to move for the third reduction point.” 581 F.3d at 1002. Johnson entered a conditional guilty plea, seeking to preserve his right to appeal the denial of a motion to suppress. Id. at 998. We agreed with the government that Johnson had not “accepted responsibility in a way that ensures the certainty of his just punishment in a timely manner.” Id. at 1003 (quoting U.S.S.G. § 3E1.1, cmt. backg’d.). The government was within its discretion not to move for the extra reduction because “[a]voiding the expenditure of additional resources in anticipation of and defending against an appeal is a legitimate governmental interest.” Id. (citation omitted). 

In United States v. Sahagun-Gallegos, we vacated and remanded—considering Amendment 775—so the district court could assess whether the defendant should receive a third point for acceptance of responsibility. 782 F.3d 1094, 1097 (9th Cir. 2015). The PSR stated that the government “would have moved for a third-level reduction if SahagunGallegos had waived his appeal rights.” Id. And, because the amended application note stated that the motion should not be withheld “based on interests not identified in § 3E1.1, such as whether the defendant agrees to waive his or her right to appeal,” we remanded. Id. (quoting § 3E1.1, cmt. n.6 (2014)). We also stated in a section heading that “[r]emand is required because the Government improperly withheld a motion for a third-level reduction for acceptance of responsibility.” Id.

Amendment 775 is no longer in effect. The Commission adopted Amendment 820 to § 3E1.1 in 2023. U.S.S.G. supp. to app. C, amend. 820 (Nov. 1, 2023) (codified at § 3E1.1). Amendment 820 deleted the statement that Amendment 775 added to Application Note 6.

[A]s we noted in Espinoza-Cano, the PROTECT Act’s language providing for a downward departure “upon motion of the government,” PROTECT Act, § 401(g)(1)(A), 117 Stat. at 671, vests the government “with broad discretion to determine when the adjustment is appropriate,” and this discretion may be leveraged to provide a defendant with an incentive to waive his right to an appeal. 456 F.3d at 1135, 1137–38. Those two features do not depend on the meaning of “preparing for trial,” and thus Amendment 820 supplies no basis to depart from these holdings. Espinoza-Cano and Johnson are still good law. Those decisions—and the plain language and statutory history of the PROTECT Act—confirm that Congress intentionally gave the government extremely broad discretion to decide whether to move for the extra reduction under § 3E1.1(b).

[Here,] [t]he government properly exercised its discretion not to file a motion under § 3E1.1(b). The government declined to move for the extra point because, “although the trial was limited in scope, the government was not able to avoid preparing for trial and still had to allocate its resources to that trial, including by having a second attorney learn the case to be able to try it.”

The district court properly applied the four-level enhancement under § 2K2.1(b)(7)(B) because Yates possessed a firearm in connection with another felony offense. Despite being a wobbler offense, child endangerment under California Penal Code § 273a(a) qualifies as “another felony offense” under § 2K2.1(b)(7)(B) here because the district court properly exercised its discretion to decide that Yates’s conduct was felonious.

Denton’s rationale applies with equal force to determine whether uncharged wobbler conduct constitutes a “felony offense” under § 2K2.1(b)(7)(B). When a defendant uses or possesses a firearm in connection with “a wobbler offense, but the offense is uncharged, no presumption applies as to whether the offense is punishable as a felony (more than one year’s imprisonment) or a misdemeanor (less than one year’s imprisonment).” Denton, 611 F.3d at 652. Consistent with Denton, “the district court must exercise its discretion to  decide whether the offense was punishable by more than one year’s imprisonment.” Id. But the district court need not say any magic words or take an overly formulaic approach. It simply must determine whether a state trial court likely would have punished the offense as a misdemeanor or a felony. See id. To do so, the district court should examine the Alvarez factors, including the nature and circumstances of the offense, the defendant’s attitude towards the offense, the defendant’s character, and the objectives of sentencing. 

The court cannot know for certain whether a wobbler offense will be punishable by imprisonment for a term exceeding one year until the defendant is arrested and charged. If the defendant were arrested and charged with a misdemeanor offense from the outset, the offense would not be punishable by over a year in prison and therefore would not qualify as “another felony offense” under § 2K2.1(b)(7)(B). In other words, with uncharged wobbler conduct, the district court cannot say whether the offense is a misdemeanor or felony without exercising its discretion. Thus, the Denton approach applies equally here. The district court appropriately exercised its discretion under Denton.  

7/28/26: Important categorical analysis decision

In United States v. Lopez, --- F.4th ---, No. 24-3268 (9th Cir. 2026), the Court vacated the sentence imposed on Jonathan Lopez following his guilty plea to possessing a firearm as a prohibited person, in violation of 18 U.S.C. § 922(g)(1), and remanded for resentencing. 

Jonathan Lopez pleaded guilty to possessing a firearm as a prohibited person in violation of 18 U.S.C. § 922(g)(1). The district court enhanced his sentence, under U.S.S.G § 2K2.1(a)(2), after determining that Lopez had previously been convicted of two felony “crimes of violence,” under California Penal Code § 273.5,1 a domestic violence offense that punishes a person who “willfully inflicts corporal injury resulting in a traumatic condition.” 

Lopez appeals his sentence, arguing that a conviction under Cal. Penal Code § 273.5 is not a categorical crime of violence. We have previously held that § 273.5 is a categorical crime of violence, but these decisions are clearly irreconcilable with the Supreme Court’s ruling in Borden v. United States, 593 U.S. 420 (2021), and our recent en banc decision in United States v. Gomez, 165 F.4th 1199 (9th Cir. 2026) (en banc), cert. denied, No. 25-7193, 2026 WL 1377030 (U.S. May 18, 2026). Together, Borden and Gomez make clear that to qualify as a categorical crime of violence an offense requires a mens rea greater than recklessness as to the use of force directed at another person. Thus, in light of Borden and Gomez, we hold that convictions under Cal. Penal Code § 273.5 do not categorically qualify as crimes of violence. 

Our prior precedent holding that § 273.5 is categorically a crime of violence must be overruled under Miller v. Gammie, 335 F.3d 889, 893 (9th Cir. 2003) (en banc), because it is clearly irreconcilable with the Supreme Court’s opinion in Borden, and our en banc decision in Gomez. Thus, we hold that the district court incorrectly applied the sentencing enhancement for two felony crimes of violence in this case.

In Borden, a plurality of the Supreme Court2 held that the elements clause does not include an offense that “requires only a mens rea of recklessness.” 593 U.S. at 423. Notably, the Court also defined what it meant by recklessness for purposes of the violent felony definition: reckless conduct requires that an individual “‘consciously disregard[] a substantial and unjustifiable risk’ attached to his conduct, in ‘gross deviation’ from accepted standards.” Id. at 427 (citing Model Penal Code § 2.02(2)(c)). Thus, as compared to the mental states of purpose and knowledge, “[r]ecklessness” is “less culpable” because it “instead involve[s] insufficient concern with a risk of injury.” Id.

Before Borden, our circuit had also held that “crimes of recklessness cannot be crimes of violence.” Fernandez-Ruiz v. Gonzales, 466 F.3d 1121, 1130 (9th Cir. 2006) (en banc). But we did not distinguish which element of a crime required a mens rea greater than recklessness to qualify as a crime of violence. This distinction is important because crimes with multiple elements may require a different mens rea with respect to each. For example, a battery statute may require that “the force involved” is “volitional,” meaning that an “involuntary motion” would not qualify. Voisine v. United States, 579 U.S. 686, 692–93 (2016). But even if the initial act of force is intentional, the defendant may have acted without intending that the force be directed at anyone else; a defendant may intentionally throw an object, but not know, or intend, that the object would hit another person. Even then, however, the defendant may be guilty of battery if the thrown object hits another person. In such a scenario, the mens rea with respect to the initial act (the throw) is intent, though the statute is “indifferent as to whether the actor has the mental state of intention, knowledge, or recklessness with respect to the harmful consequences of his volitional conduct.”

Borden makes clear that, to qualify as a crime of violence, an offense requires a mens rea greater than recklessness (e.g., purpose or knowledge)4 as to the use of force directed at another person—rather than merely any volitional act involved with the crime. 

Accordingly, the elements clause includes only those crimes that require a mens rea greater than recklessness as to the use of violent force directed at another person—meaning that when the defendant used force, he intended or knew that the force would touch another. 5 Borden explains that “violent felon[ies]” are “best understood to involve not only a substantial degree of force, but also a purposeful or knowing mental state—a deliberate choice of wreaking harm on another, rather than mere indifference to risk.”

In sum, to qualify as a categorical “crime of violence,” an offense must require a mens rea greater than recklessness both as to the use of force sufficient to cause pain or injury and to the fact that the use of force is directed at another person.

While the text of Cal. Penal Code § 273.5 may appear to suggest that the statute requires that the defendant intentionally injure the victim, California appellate courts have consistently explained that § 273.5 “requires only the mens rea of intending to do the assaultive act” that is likely to result in bodily harm and does not require a separate intent to cause injury.

Cal. Penal Code § 273.5(a) is not a categorical crime of violence because the offense “requires only the mens rea of intending to do the assaultive act,” and does not require intent, or knowledge, that the use of force is directed at another.

As a three-judge panel, we are generally bound by our circuit’s precedent. Kohler v. Presidio Int’l, Inc., 782 F.3d 1064, 1070 (9th Cir. 2015). However, we “should reject the prior circuit” precedent when it is “clearly irreconcilable with the reasoning or theory of intervening higher authority.” Miller, 335 F.3d at 893. Our previous caselaw holding that § 273.5 is categorically a crime of violence is clearly irreconcilable with Borden and Gomez.

Convictions under California Penal Code § 273.5 are not crimes of violence. We overrule our prior precedent to the contrary under Miller v. Gammie. See 335 F.3d at 889. The district court, while properly applying our prior precedent, should not have applied U.S.S.G. § 2K2.1(a)(2) to enhance to Lopez’s sentence. We REMAND with instructions to the district court to resentence Lopez consistently with this opinion. 

7/27/26: Two decisions today: ACCA and True Threats

 First, in United States v. Patterson, --- F.4th ---, No. 22-30088 (9th Cir. 2026), the Court affirmed the ACCA sentence imposed on James Patterson following his guilty plea to being a felon in possession of a firearm, in violation of 18 U.S.C. § 922(g)(1).

James Patterson pleaded guilty to being a felon in possession of a firearm, in violation of 18 U.S.C. § 922(g)(1). Finding that Patterson had previously committed three violent felonies on different occasions, the district court imposed a mandatory minimum sentence under the Armed Career Criminal Act of 1984 (ACCA), 18 U.S.C. § 924(e)(1). Patterson appeals, arguing first that his Fifth and Sixth Amendment rights under Apprendi v. New Jersey, 530 U.S. 466 (2000), were violated because the judge, rather than a jury, determined that his prior offenses occurred on different occasions, and second that the court erred in holding that his prior Oregon convictions—two for third-degree assault and one for strangulation—are violent felonies. In light of the Supreme Court’s decision in Erlinger v. United States, 602 U.S. 821 (2024), we conclude that the district court violated Apprendi, but that the error was harmless. We also conclude that Patterson’s Oregon state felony convictions are violent felonies. We therefore affirm. 

The district court’s determination that Patterson committed his prior offenses “on occasions different from one another” made him subject to the ACCA, thereby increasing both the statutory minimum and the statutory maximum sentence for his violation of section 922(g). Patterson argues that the district court violated Apprendi because the indictment did not allege that the offenses were committed on different occasions and because the court found that fact itself instead of submitting the question to a jury.

A sentencing court has no “authority [to] . . . assume for itself the responsibility of deciding whether [a defendant’s] past offenses differed enough in time, location, character, and purpose to have transpired on different occasions . . . with an eye toward increasing his punishment.”

Applying Erlinger, we agree with both parties that the district court violated Apprendi by determining whether Patterson committed his prior offenses on different occasions. Making that determination required the court to ascertain facts beyond “a prior conviction and the thenexisting elements of that offense,” and Patterson did not admit those factsin his guilty plea. Erlinger, 602 U.S. at 839; see Mathis, 579 U.S. at 511. Our prior decision in Grisel is “clearly irreconcilable” with Erlinger and has “been effectively overruled” by it.

Based on the record, we are confident that no reasonable jury could determine that offenses committed three years apart were committed on the same occasion. The Apprendi error was therefore harmless beyond a reasonable doubt.

Patterson also disputes the district court’s determination that his Oregon assault and strangulation convictions are violent felonies constituting ACCA predicates.

To determine whether an offense constitutes an ACCA predicate, we employ a “categorical approach” under which “the facts of a given case are irrelevant” and “[t]he focus is instead on whether the elements of the statute of conviction meet the federal standard.”

We first consider Patterson’s third-degree assault conviction. In Oregon, third-degree assault, which carries a punishment of up to five years of imprisonment, is committed by anyone who, “[w]hile being aided by another person actually present, intentionally or knowingly causes physical injury to another.” Or. Rev. Stat. § 163.165(1)(e) (1995); see id. § 163.165(2) (defining third-degree assault as a class C felony); id. § 161.605 (specifying the punishment for class C felonies). Patterson’s conviction is an ACCA predicate because the Oregon statute requires that a defendant intentionally or knowingly cause “physical injury,” so it has “as an element the use . . . of physical force.” 

Finally, we reject Patterson’s argument that Oregon’s strangulation offense is not categorically violent. 

Of note, Judge Berzon has a thorough concurrence regarding the application of the harmless-error standard in cases where no jury trial took place.  She calls for en banc review, so we should be doing the same. 

I write separately to express my continued concerns regarding this court’s application of the harmless error review standard established in Neder v. United States, 527 U.S. 1 (1999), to cases, like this one, in which no jury trial took place.

“If the defendant did not admit an essential fact during his plea colloquy and evidence concerning the fact was not put to a jury, it violates Apprendi for a court to allow the government, postconviction, to introduce new evidence, find that evidence of the fact would have been introduced in a hypothetical jury trial, and then determine that the essential fact would have been found by the hypothetical jury had thenewly produced evidence been before it.” Guerrero-Jasso, 752 F.3d at 1204 (Berzon, J., concurring). This court’s precedents—now extended into the Erlinger context— permit the government to do just that. So I renew my call for this court to reconsider en banc “our Apprendi harmless error cases, particularly Zepeda-Martinez.” Id.

Next, in United States v. Anten, --- F.4th ---, No. 24-7261 (9th Cir. 2026), the Court affirmed Mark William Anten’s jury conviction for transmitting threats to injure another in interstate commerce, in violation of 18 U.S.C. § 875(c).

In an escalating barrage of emails to the Federal Bureau of Investigation (FBI), Mark William Anten eventually declared himself the “UNABOMBER” and stated, “I WILL UNABOMB THE LOS ANGELES FBI HQ.” A jury convicted Anten of transmitting threats to injure another in interstate commerce, in violation of 18 U.S.C. § 875(c). Anten now appeals, arguing that the district court’s jury instructions omitted an objective element required by the statute. We decide that the statute requires an objective element, but the jury instructions covered it. Anten’s other challenge, that the district court invaded the province of the jury, lacks merit too. We have jurisdiction under 28 U.S.C. § 1291, and we affirm.

We now confirm what we have previously assumed—that § 875(c) requires an objective threat element as well as a subjective mental state element.

To parse out the elements of 18 U.S.C. § 875(c), we begin with the text. Section 875(c) states: “Whoever transmits in interstate or foreign commerce any communication containing any threat to kidnap any person or any threat to injure the person of another, shall be fined under this title or imprisoned not more than five years, or both.” The text contains at least two elements: (1) “a communication be transmitted” and (2) “the communication contain a threat.” 

The statute specifies no “mental state with respect to these elements.” Id. In Elonis, the Supreme Court held that a mental state element must be read into § 875(c). 

In Counterman, the Court addressed the constitutional constraints around the mental state element in true threats cases generally, holding that “a mental state of recklessness is sufficient” to satisfy the First Amendment.

A threat is a threat, whatever the speaker’s mental state. To secure a conviction under § 875(c), the government must prove that a defendant both made an objective “threat,” and that he did so with the necessary subjective mental state. 

We next examine what constitutes an objective “threat.” “What is a threat must be distinguished from what is constitutionally protected speech.” Watts v. United States, 394 U.S. 705, 707 (1969) (per curiam). Thus, the statement at issue must be a “true” threat. Id. at 708. “The ‘true’ in that term distinguishes what is at issue from jests, ‘hyperbole,’ or other statements that when taken in context do not convey a real possibility that violence will follow (say, ‘I am going to kill you for showing up late’).”

The Supreme Court has provided less guidance on whose perspective must be considered to determine whether a statement is a “true threat”: the speaker’s, the victim’s, or a hypothetical reasonable person’s. Process of elimination reveals the answer. 

The answer cannot be the defendant speaker’s perspective, or the “threat” element would be no different than the defendant’s subjective mental state requirement. See id. (“Whether the speaker is aware of . . . the threatening aspect of the message is not part of what makes a statement a threat . . . .”). The answer also cannot be the victim’s perspective. The victim of a threat could be a particularly sensitive or paranoid individual, unable to determine whether the statement “taken in context” “convey[s] a real possibility that violence will follow.” See id. (citing Watts, 394 U.S. at 708) (italics added). Thus, whether a statement is a “threat” must be judged through the eyes of a reasonable person. Indeed, without a reasonable-person standard, a defendant could be convicted without ever having uttered a true threat—both the threat’s author and the threat’s target could have made unreasonable assessments.

Having determined that § 875(c) requires both an objective and a subjective element, we now turn to the jury instructions. Anten is correct that the district court’s instructions could have been clearer. In describing what constitutes a “threat,” the instructions did not mention the “reasonable person” standard and did not define a “threat” as a “serious expression.” Nonetheless, we disagree with Anten that the district court omitted the statute’s objective threat element.

We hold that § 875(c) requires both an objective threat element and a subjective mental state element. The district court did not err in its jury instructions, which included both elements. The district court also did not err in its response to the jury’s note. We therefore affirm. 

7/23/26: Fourth Amendment parolee exception case

In United States v. Pereda, --- F.4th ---, No. 25-4461 (9th Cir. 2026), a divided Court reversed the district court’s order granting Binisio Nicolas Pereda’s motion to suppress evidence found in a trailer and a pickup truck and remanded for further proceedings. 

One morning, police officers found Binisio Nicolas Pereda—a state early-release supervisee subject to warrantless search conditions—in a trailer located near his girlfriend’s parents’ home. Officers then searched the trailer and a nearby pickup truck thought to be used by Pereda. They found ammunition in the trailer and drugs in the truck. We must decide whether officers had probable cause to believe that Pereda (1) resided in or controlled the trailer, and (2) owned or controlled the pickup truck. Because the officers had probable cause for both searches, the searches were consistent with the Fourth Amendment. We thus reverse the district court’s order granting Pereda’s motion to suppress and remand for further proceedings.

The Fourth Amendment to the United States Constitution provides that “[t]he right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures, shall not be violated[.]” U.S. Const. amend. IV. Warrantless searches are generally “per se unreasonable.” United States v. Estrella, 69 F.4th 958, 964 (9th Cir. 2023) (simplified). But there are exceptions to this rule. United States v. Brown, 996 F.3d 998, 1004 (9th Cir. 2021). One of the recognized exceptions to the warrant requirement is the so-called parolee exception. Estrella, 69 F.4th at 964. Under the parolee exception, “[a] search of a parolee that complies with the terms of a valid search condition will usually be deemed reasonable under the Fourth Amendment.”

Pereda’s post-release community supervision search conditions authorized a search of his “residence.” When a search condition permits a warrantless search of a parolee’s “residence,” “law enforcement officers must have probable cause to believe that the parolee is a resident of the house to be searched.” United States v. Grandberry, 730 F.3d 968, 973 (9th Cir. 2013) (simplified). The probable cause needed to search a parolee’s residence is the same well-established probable-cause standard used throughout criminal law. It “is not a high bar[.]”

Considering the whole picture, the officers had ample probable cause to believe that the trailer was Pereda’s residence.

Applying that fact-intensive standard, we conclude that there was a fair probability that Pereda was residing in the trailer. So the officers had probable cause to search the trailer according to Pereda’s parole conditions. The search was thus consistent with the Fourth Amendment.

The officers also reasonably searched the truck parked next to the trailer. As noted above, Pereda’s post-release community supervision terms permitted a warrantless search of any “property under [Pereda’s] control.” That provision applies to the search of a parolee’s vehicle. See id. at 981. So Pereda’s supervised release conditions permitted a search of his vehicle if the officers had probable cause to believe that he owned or controlled the vehicle searched. Dixon, 984 F.3d at 822. Once again, probable cause is “not a high bar” and “requires only the kind of fair probability on which reasonable and prudent people, not legal technicians, act.” Kaley, 571 U.S. at 338 (simplified).

Based on the totality of the circumstances, the facts available to the officers established a fair probability that Pereda owned or controlled the white pickup truck. Armed with probable cause, the officers validly searched the truck according to Pereda’s supervised release terms.  

Judge Paez dissented in a thorough opinion that would have affirmed the district court.

7/21/26: Actual juror bias

In United States v. Collar, --- F.4th ---, No. 23-2939 (9th Cir. 2026), the Court vacated Attila Colar’s conviction on 44 counts arising from an alleged scheme to defraud the Paycheck Protection Program and efforts to obstruct the investigation, reversed the district court’s judgment, and remanded for a new trial based on the district court's failure to adequately address a juror bias.

The Sixth Amendment guarantees criminal defendants the right to trial by an impartial jury. The participation of even a single biased juror in rendering a guilty verdict is structural error requiring reversal. To safeguard this right, a district court must address any signs of bias during jury selection and throughout trial. The court has broad discretion over how to investigate a juror’s impartiality. That discretion, however, governs the form of the inquiry, not whether to conduct one. Once a colorable issue of juror bias arises, the court must investigate before the juror continues to serve, and the inquiry must match the seriousness of the claim. The court cannot defer that obligation simply because the juror is an alternate. 

A jury convicted Attila Colar on 44 federal counts arising from an alleged scheme to defraud the Paycheck Protection Program and efforts to obstruct the investigation. At voir dire, Colar advanced a defense theory that he maintained throughout his trial: law enforcement targeted him based on his race and religion. After the jury was impaneled, an alternate juror expressed actual bias against Colar. He told the court that he had a strong negative reaction to Colar’s defense and that he had concluded that the defense lacked substance and credibility. 

The district court abused its discretion by deferring any meaningful investigation into the juror’s bias until he was needed on the regular panel. And its belated inquiry did not confirm that the juror could set aside the bias he previously expressed or that continued exposure to the defense theory would not revive it. The juror’s resulting participation in the verdict was structural error. We reverse and remand for a new trial.

A district court’s duty to protect against juror bias begins, but does not end, at voir dire. Voir dire allows the parties and the court to identify and investigate prejudice at the outset. Rosales-Lopez v. United States, 451 U.S. 182, 188 (1981). But the court’s obligation does not end when the jury is impaneled; trial courts must remain “ever watchful to prevent prejudicial occurrences and to determine the effect of such occurrences when they happen.” Smith v. Phillips, 455 U.S. 209, 217 (1982). When information casting doubt on a sitting juror’s impartiality comes to the court’s attention during trial, its continuing obligation to ensure an impartial jury can be described in two steps.

First the court must assess whether the information before it amounts to a “colorable claim of juror bias.” Dyer, 151 F.3d at 974. A colorable claim is not “wholly insubstantial and frivolous.” See Keniston v. Roberts, 717 F.2d 1295, 1298 (9th Cir. 1983). If the information does not rise to a colorable claim, the inquiry ends. But a court presented with a colorable claim of juror bias must take a second step and inquire further. It “must undertake an investigation of the relevant facts and circumstances” and “erect[], and employ[], a suitable framework for investigating the allegation of bias and gauging its effects.” 

The court has considerable discretion in how it investigates juror bias. It may attempt a rehabilitative colloquy,see United States v. Alexander, 48 F.3d 1477, 1484 (9th Cir. 1995), hold an evidentiary hearing, see Hendrix, 549 F.2d at 1227⁠–⁠29, or excuse the juror outright, see Gonzalez, 214 F.3d at 1114. Still, this discretion is “not unbounded.” United States v. Eubanks, 591 F.2d 513, 516 n.3 (9th Cir. 1979). Which framework fits depends on “the content of the allegations, including the seriousness of the alleged misconduct or bias, and the credibility of the source.” Hendrix, 549 F.2d at 1227⁠–⁠28; Dyer, 151 F.3d at 975 (noting that an in camera hearing to investigate potential bias must “be reasonably calculated to resolve the doubts” about a juror’s impartiality). The more serious the claim, the more in-depth the inquiry must be. See Hendrix, 549 F.2d at 1229. In conducting this inquiry, the court “must find facts, not make assumptions,” Dyer, 151 F.3d at 976, resolving any doubts about the juror’s impartiality against the juror, Gonzalez, 214 F.3d at 1114. And the court should promptly investigate such claims, rather than “leaving the matter in doubt” and the juror in place. See Dyer, 151 F.3d at 976. 

Applying this framework, we hold that the district court abused its discretion. We appreciate the difficulty of addressing a pro se defendant’s challenge to a vocal alternate juror. But Juror 48’s statements raised a colorable issue of whether he was biased. That demanded a searching inquiry before he continued to serve on the jury as an alternate. Instead, the court deferred an investigation until Juror 48 was needed for the regular jury panel. This was legal error. See Hinkson, 585 F.3d at 1261⁠–⁠62. And the court’s belated investigation, after Juror 48 had heard evidence, failed to establish that he set aside the bias he previously expressed or that continued exposure to the defense theory would not revive it. Its finding that Juror 48 was not actually biased thus lacked sufficient support in the record. Id. at 1262.  

7/2/26: Case on Sentencing Guidelines for money laundering offenses

In United States v. Shi, et al., --- F.4th ---, No. 24-2054 (9th Cir. 2026), the Court affirmed in part and vacated in part the sentences imposed on Defendants Blade Bai, Bowen Hu, and Tairan Shi following their convictions for offenses arising from their participation in a sophisticated scheme to launder Target gift cards purchased by telephone-scam victims; and remanded for a limited resentencing. 

This case is about how to properly calculate the Guidelines for these types of offenses. 

Over the course of a year, Defendants Blade Bai, Bowen Hu, and Tairan Shi (Defendants) participated in a sophisticated scheme to launder Target gift cards purchased by telephone-scam victims. The jury convicted Defendants of conspiracy to commit money laundering in violation of 18 U.S.C. § 1956(h) (Count 1). The jury also convicted Bai of conspiracy to commit money laundering while under release in violation of 18 U.S.C. §§ 1956(h) and 3147 (Count 2).1 Defendants appeal their sentences, challenging the district court’s calculation of the loss amount, application of the sophisticated laundering enhancement, and application of the aggravated and minor role adjustments. For the reasons below, we affirm as to the calculation of the loss amount and role adjustments but reverse as to the sophisticated laundering enhancement.

We first consider Defendants’ claims regarding the district court’s calculation of loss amount for the base offense level determination. Defendants contend that (1) as a matter of law the intended loss amount may not be included in the value of laundered funds, (2) the district court improperly calculated the amount of laundered funds, and (3) if the loss amount calculation was not error, the district court should have applied a three-level reduction under U.S.S.G. § 2X1.1(b)(2). 2 For the reasons discussed below, we reject these arguments and conclude that the district court properly applied and calculated the loss amount.

To determine the offense level, the district court applied § 2X1.1, which is the starting point for sentencing calculations of offenses involving attempts, solicitations, and conspiracies not already covered by specific offense guidelines. Under § 2X1.1(a), courts calculate the base offense level by using “[t]he base offense level from the guideline for the substantive offense, plus any adjustments from such guideline for any intended offense conduct that can be established with reasonable certainty.” (Emphases added). The “substantive offense” “means the offense that the defendant was convicted of soliciting, attempting, or conspiring to commit.” Id. cmt. n.2. 

Accordingly, the district court turned to § 2S1.1, the money laundering guideline. Section 2S1.1(a)(2) provides for a base offense level of “8 plus the number of offense levels from the table in § 2B1.1 (Theft, Property Destruction, and Fraud) corresponding to the value of the laundered funds.” The district court found that the value of the laundered funds conspired to be laundered exceeded $1.5 million but was less than $3.5 million. It thus increased Defendants’ base offense level by 16 pursuant to § 2B1.1(b)(1)(I) for a total level of 24. 

Defendants contend this analysis was error, arguing that the term “adjustments” as used in § 2X1.1(a) refers only to specific offense characteristics under § 2S1.1(b), and does not apply to alter the base offense level. Thus, under Defendants’ theory, the district court only considers intended loss when looking at specific offense characteristics.

[W]e conclude that the district court did not err by considering the intended loss as part of the “value of the laundered funds” under § 2S1.1(a)(2). 3 Defendants’ “intended conduct is a proper basis for the enhancements the district court applied,” including an “enhancement[] for loss amount . . . based on conduct that [they] intended but did not carry out.”

Defendants also argue that even if the district court did not err by considering the intended loss in its base offense level determination, it nevertheless improperly calculated the amount of funds laundered. At sentencing, the evidence showed, and the district court agreed, that the total value of laundered funds came to $2.48 million. Defendants argue that the district court wrongfully assumed that every single card number was valued at $500, constituted proceeds of a specified unlawful activity, and was successfully laundered. 

We see no basis to disturb the district court’s findings on review. The Government presented evidence extracted from Defendants’ WeChat messages showing that over the course of 12 months, Defendants received 5,256 unique gift card numbers from Magic Lamp. After further automated extraction and manual review, the Government calculated that the average value of each card was $472. Thus, the total value came to $2.48 million.

Finally, Defendants alternatively assert that if the district court properly calculated their base offense level, it should have also granted them a three-level reduction under § 2X1.1(b)(2).

The district court did not plainly err by not applying this three-level reduction. The evidence showed that Defendants engaged in a money laundering scheme for over a year. There is no indication that they intended to stop their unlawful activities until Bai was arrested in November 2020. And the jury found in Count 2 that even the first arrest did not stop Bai from continuing to offend. Further, Magic Lamp continually supplied Defendants with cards for the duration of the conspiracy, supporting the inference that the cards they transacted were successfully laundered. In sum, we hold that the district court did not err with respect to its calculation of loss amount.

Defendants next challenge the two-level enhancement for sophisticated laundering under § 2S1.1(b)(3).

The district court determined that Defendants were engaged in the business of laundering funds and so applied subsection (b)(2)(C) to increase Defendants’ offense level by four. It also applied subsection (b)(3) on the basis that Defendants engaged in sophisticated laundering, which increased the offense level by another two levels. Defendants thus received a six-level total increase under § 2S1.1(b). The issue we must decide is whether subsection (b)(3) requires the district court to have actually imposed the two-level increase under subsection (b)(2)(B), or if it is enough simply that the defendant was convicted under 18 U.S.C. § 1956.

As an initial matter, the parties dispute the applicable standard of review.

Defendants respond that “it is claims that are deemed waived or forfeited, not arguments.” United States v. Pallares-Galan, 359 F.3d 1088, 1095 (9th Cir. 2004). They assert that de novo review applies because though their argument now includes condition (A) as a basis for why subsection (b)(3) does not apply, their general claim that subsection (b)(3) does not apply was raised to the district court. Defendants have the better argument.

[T]he plain language and structure of § 2S1.1 leads us to conclude that a district court must actually apply (b)(2)(B) before a defendant can be subject to (b)(3). Because the district court here applied (b)(2)(C), and not (b)(2)(B), it improperly imposed the two-level enhancement under (b)(3) for sophisticated laundering. This error requires that we vacate and remand for the limited purpose of adjusting the guideline computation and resentencing the Defendants. The district court is not required to conduct a plenary resentencing.

Finally, we consider the role adjustments applied to Hu’s and Shi’s sentences. For the reasons discussed below, we affirm the district court’s application of the aggravated role adjustment to Hu and Shi and its decision not to grant Shi a minor role adjustment.

6/29/26: SCOTUS on the Fourth Amendment and the Ninth on return of property

First, in Chatrie v. United States, 609 U.S. ---, No. 25-112 (2026), the Supreme Court majority held that Police officers conducted a Fourth Amendment search when they acquired Chatrie’s location data from Google because an individual has a reasonable expectation of privacy in his cell-phone location information.  In other words, the Court held that geo-fence warrants implicate the Fourth Amendment.  The majority's decision, however, is a bit unusual because it does not consider the validity of the warrant obtained, but only whether a warrant was necessary at all. 

In recent years, law enforcement officers have employed so-called geofence warrants to obtain information that technology companies collect about their users’ cell-phone locations. Suppose that investigators know a crime was committed at a particular place and time, but do not have a suspect. They may draw a “geofence”—a virtual perimeter—around the crime scene and get a warrant compelling a company to hand over data about the cell phones located in that area near the time of the crime. Following a process specified in the warrant, the company will turn over the cell-phone data and eventually identify by name one or more of the users thus disclosed. 

The geofence warrant at issue here was directed to Google, and used to solve a bank robbery. Hundreds of millions of Google users have activated a service called Location History, which records the location of a user’s cell phone every two minutes or so. Through a geofence warrant, police officers required Google to turn over Location History data revealing cell phones within the vicinity of a bank at around the time it was robbed. At the end of the multi-step process described in the warrant, Google gave the police three names. The Federal Government soon charged one of the individuals thus identified, petitioner Okello Chatrie, with committing the crime. 

Today, we consider how the Fourth Amendment applies to that use of a geofence warrant. Answering that question in full would mean deciding whether the police conducted a Fourth Amendment “search” when they acquired the cellphone data leading to Chatrie’s arrest and, if so, whether that search was reasonable given the features of the warrant they employed. We decide the first part of that inquiry today, concluding that the police conducted a search when they gained access to Location History data. An individual has a reasonable expectation of privacy in records about his cell phone’s location, and police intrude on that constitutionally protected interest when they demand the information—even though for only a limited time, and from a third-party tech company. We leave to the Court of Appeals the further question whether, given the warrant issued, the search here was reasonable, meaning that each of its steps was properly described with particularity and found to be supported by probable cause.

We granted certiorari solely on the question whether the police violated the Fourth Amendment in obtaining Chatrie’s location data, thus declining to consider the exclusionary rule issue. See 607 U. S. 1148 (2026). The disputed Fourth Amendment question divides into two parts. First, did law enforcement officials conduct a search under the Fourth Amendment when they acquired Chatrie’s location data from Google? We hold that they did because an individual has a legitimate expectation of privacy in his cell-phone location data. Second, did the multi-step geofence warrant issued here make that search reasonable? We leave that question—which requires deciding whether the warrant satisfied the Fourth Amendment’s probable cause and particularity requirements at each stage of the search process—to the Court of Appeals to address in the first instance.

Whatever the form of an attempted incursion, the Fourth Amendment protects Americans’ long-held conviction that no government official should have free access to the most closely kept aspects of their lives.

Contrary to the Government’s view, an individual has a legitimate expectation of privacy in the information Location History collects about his cell phone’s—meaning his own—movements. The police invade that expectation, and thus conduct a search, when they acquire that information, even though for only a limited period of time and even though via a third-party tech company.

Location History resembles other private materials—think of emails, documents, photographs, or calendars—that even if stored on Google’s servers, a user reasonably views as his own. And as a result, that he reasonably expects to be shielded from the “inquisitive eyes” of the government. 

The Government replies with an odd argument. It thinks that “tracking [someone] into a private residence”—yes, even for two hours—would “probably” be a search, but tells us not to worry because Chatrie did not go home. Tr. of Oral Arg. 98, 134. That approach, however, is foreign to the way the Fourth Amendment works. Whether something is a search does not depend on what it finds. See Di Re, 332 U. S., at 595 (“[A] search is not to be made legal by what it turns up. In law it is good or bad when it starts”). An officer, after all, cannot know the fruits of a given surveillance in advance. The surveillance must be either a search or not regardless. The Government’s concession thus gives away its argument that, for purposes of the Fourth Amendment, two hours of cell-phone location data is not enough.

Location History is even more “revealing” than CSLI, because it provides a yet more precise record of an individual’s movements. See supra, at 16–17. Access to that record enables officials to undertake nearly perfect, retrospective surveillance of countless persons and places. See supra, at 17–18. And for Location History, that surveillance is based on information that a user reasonably understands as his own, even though stored on Google’s servers—much like his emails, photos, and calendar entries. See supra, at 18. Likewise, the information is “not truly shared,” in the normal sense of wanting a third party to see or use it. Carpenter, 585 U. S., at 315. The exposure of that information to Google is merely what happens when a user avails himself of one of the services on his cell phone. Or said a bit differently, it is the automatic price of conventional cellphone usage—which, just as Carpenter noted, is a “pervasive and insistent part of daily life.” Ibid. So just as the third-party doctrine did not apply in Carpenter, it does not apply here.

For all those reasons, we hold that police officers invade a cell-phone user’s reasonable expectation of privacy when they access his Location History. It does not matter if the time period scrutinized was only two hours. Nor does it matter that the materials obtained were handed over by a third-party tech company. When the government “accesses historical cell phone” location information—Location History as much as CSLI—it “conducts a search under the Fourth Amendment.”

That conclusion does not resolve this case, because the Fourth Amendment prohibits only searches that are “unreasonable.” When law enforcement officials undertake a search to discover evidence of a crime, the reasonableness standard generally requires that they seek a warrant from “a neutral and detached magistrate.” Johnson v. United States, 333 U. S. 10, 14 (1948); see Vernonia School Dist. 47J v. Acton, 515 U. S. 646, 653 (1995).11 That requirement subjects the officials’ assessment of a search’s propriety to the “deliberate, impartial judgment of a judicial officer.” United States v. Grubbs, 547 U. S. 90, 99 (2006). The magistrate, in turn, may issue a warrant only when “probable cause is properly established and the scope of the authorized search is set out with particularity.” Kentucky v. King, 563 U. S. 452, 459 (2011). 

“[P]robable cause is a fluid concept—turning on the assessment of probabilities in particular factual contexts—not readily, or even usefully, reduced to a neat set of legal rules.” Illinois v. Gates, 462 U. S. 213, 232 (1983). But a magistrate must always determine that there is a “fair probability that contraband or evidence of a crime will be found” in the place searched. Id., at 238. That means determining, to the requisite “fair probability,” both that the place searched will have the materials sought and that those materials will contain evidence “aid[ing]” in a criminal’s “apprehension or conviction.” Messerschmidt v. Millender, 565 U. S. 535, 551, 552, n. 7 (2012); see Zurcher v. Stanford Daily, 436 U. S. 547, 556 (1978) (“The critical element” is whether there is the requisite “cause to believe that the specific ‘things’ to be searched for and seized are located” in the targeted place). The particularity requirement, for its part, ensures that the search will be of an appropriate scope—that it is “carefully tailored to its justifications, and will not take on the character of the wide-ranging exploratory searches the Framers intended to prohibit.” Maryland v. Garrison, 480 U. S. 79, 84 (1987). That requirement typically looks to such matters as the geographic and durational expanse of the search. See id., at 84–85; Karo, 468 U. S., at 718. And it too must take account of “particular factual contexts,” including in surveillance cases the nature of the technology to be used. 

In his famed and vindicated dissent, Justice Brandeis explained why a wiretap was a search, subject to Fourth Amendment requirements. See Olmstead v. United States, 277 U. S. 438, 471 (1928). Those who drafted the Amendment could not have imagined such a technology. But they understood, Justice Brandeis wrote, a matter of more transcendent importance: that Americans had “as against the Government, the right to be let alone” and that the Fourth Amendment must protect against “every unjustifiable intrusion by the Government upon the privacy of the individual, whatever the means employed.” Id., at 478. 

Far more recently, this Court in Carpenter invoked Justice Brandeis’s opinion in explaining why law enforcement officials could not have “unrestricted access to a wireless carrier’s database of physical location information.” 585 U. S., at 320. Said Carpenter: “[T]he Court is obligated—as ‘[s]ubtler and more far-reaching means of invading privacy have become available to the Government’—to ensure that the ‘progress of science’ does not erode Fourth Amendment protections.” Ibid. (quoting 277 U. S., at 473–474 (dissenting opinion)). For new technological tools, the Court continued, may “risk[] Government encroachment of the sort the Framers, after consulting the lessons of history, drafted the Fourth Amendment to prevent.” 585 U. S., at 320. 

Today’s decision follows from the same judicial obligation, to guard against the same risk of undue encroachment. The Fourth Amendment applies, too, when officials tap into Google’s “database of physical location information.” Ibid. That database is new, but the principle covering it is not: That principle is instead the one our history has given. The Fourth Amendment must, as ever, protect against unjustified governmental intrusion on the privacy of the individual.

Next, in Cannon v. United States, --- F.4th ---, No. 24-1317 (9th Cir. 2026), the Court reversed the district court’s grant of summary judgment for the federal government in Lionel Cannon’s motion under Federal Rule of Criminal Procedure 41(g) seeking the return of his cash seized by the government.

The FBI was investigating Cannon on federal drug trafficking charges. During a lawful search, agents seized $585,000 in cash. FBI Special Agent Scott Bowmann pocketed $218,200. After pleading guilty to drug charges, Cannon moved under Rule 41(g) for the return of the $218,000 that Agent Bowman had stolen and had never been forfeited.

Federal Rule of Criminal Procedure 41(g) provides that a person who has been deprived of his property by the government may move for the property’s return. In Ordonez v. United States, 680 F.3d 1135 (9th Cir. 2012), we held that when property seized by the government has been lost or destroyed, “an award of money damages against the government under Rule 41(g) is barred by sovereign immunity.” Id. at 1140. The issue in this case is whether the same rule applies to cash that the government seized, lost, but then recovered in restitution proceedings. We hold that when the government has recovered money traceable to the funds it lost, the person from whom it was seized may seek its return under Rule 41(g). When its return is sought, the government bears the burden of demonstrating that it has a legitimate reason to retain the money. The government did not meet its burden here. We thus reverse the district court’s grant of summary judgment for the government.

Cannon seeks the return of the very thing to which he is entitled, namely “the property” under Rule 41(g). In this case, that is the cash the government seized but did not pursue in forfeiture proceedings. As in Bowen, Cannon’s claim is not for money damages; rather, it is “a suit seeking to enforce the statutory mandate itself, which happens to be one for the payment of money.” 487 U.S. at 900. The fact that Cannon seeks the return of cash does not make his claim one for money damages. We have long treated Rule 41(g) motions—even those for return of cash seizures—as motions in equity, not law, and equity is directed to ownership of the property itself, not to compensation, which is a legal remedy. 

But that the money the United States is recouping from Agent Bowman is not the same “physical currency”—that is, the exact same bills that were taken from Cannon’s safe—does not alter the result. Such a hyperformalistic approach misconstrues the nature of money. Money differs from other property in that its principal value lies in the unit of account; its exchange value, not its intrinsic value. 

The relevant question then is not whether the bills Agent Bowman is remitting to the government have the same serial numbers as the ones from Cannon’s safe, but whether he is remitting the value or amount traceable to the money taken from Cannon’s safe. We think it clear that he is. 

[W]e must begin with a presumption that Cannon is in lawful possession of the property taken from him; it is the government’s burden to show that he did not lawfully possess it. The government has not satisfied its burden.

Nowhere in the interview memorandum does Cannon admit that the money was drug proceeds or that he was otherwise not entitled to the money. The government’s claim to the contrary is simply an inference from silence. The bare fact that Cannon had money in a safe—admittedly, a lot of money—is not proof that the money was obtained from entirely illegitimate and unlawful sources, and Cannon’s admission in his plea deal that some of the money was drug proceeds is not an admission that all of it was.

The district court stated that Cannon’s plea deal to forfeit $366,800.00 from the safe “support[s] an inference that he agreed to forfeit all property seized from the safe in his mother’s house.” But this inference impermissibly shifts the burden of proof and misstates the parties’ bargaining position. During the plea negotiations, both Cannon and the United States understood that the government had seized far more than $366,800 from his safe that day. The $366,800 figure on the forfeiture order was no clerical error—it was the amount the parties had negotiated.

The government—not Cannon—bears the burden of demonstrating a legitimate basis for refusing to return the money; if it cannot, it must return Cannon’s property. Wright, 49 F.4th at 1225. Thus, the government bore the burden of showing that the entire amount seized was unlawful proceeds. Cannon provided evidence supporting his claim that lawful funds were intermingled with the drug proceeds subject to forfeiture. The government had the opportunity to dispute this. On this record, it has not done so, and summary judgment was in error. We return the case to the district court for additional proceedings.