5/29/25: Case on statements in made in plea agreements under Fed. R. Evid. 410.

In United States v. Puig Valdes, --- F.4th ---, No. 23-3214 (9th Cir. 2025), on interlocutory appeal by the Government, the Court affirmed the district court’s ruling that the factual basis of a pre-indictment plea agreement signed by Yasiel Puig Valdes would be excluded at trial.

In July 2022, Defendant Yasiel Puig Valdes (“Puig”) signed a pre-indictment plea agreement with the Government, under which he would plead guilty to one count of making false statements to federal officers in violation of 18 U.S.C. § 1001, and in exchange, the Government would recommend a reduced sentence and decline to bring an additional charge of obstruction of justice under 18 U.S.C. § 1503. When Puig later declined to plead guilty, the Government declared that Puig was in breach of his plea agreement, and as a remedy it sought to enforce a provision of the agreement waiving all evidentiary objections to the admission of the plea agreement’s factual basis at trial. This waiver expressly included any objections based on Rule 410 of the Federal Rules of Evidence, which generally bars the admission, against a defendant, of any statements made during plea negotiations. The district court ultimately held that Rule 410 remained applicable here, and it therefore ruled that the factual basis of Puig’s plea agreement would be excluded at trial. Pursuant to 18 U.S.C. § 3731, the Government brought this interlocutory appeal challenging that pretrial ruling. Although our reasoning differs somewhat from the district court’s, we agree that Rule 410 remains applicable here, and we therefore affirm.

On appeal, the Government contends that, even though the district court had not yet accepted either Puig’s plea agreement or any guilty plea from him, that agreement remained binding and enforceable, including its waiver of the rule of exclusion contained in Federal Rule of Evidence 410.

[T]his written statement attributed to Puig plainly constitutes a “statement made during plea discussions with an attorney for the prosecuting authority,” and it is therefore “not admissible against the defendant who . . . participated in the plea discussions” where, as here, “the discussions did not result in a guilty plea.” 

However, because Evidence Rule 410 and Criminal Rule 11(f) were “enacted against a background presumption that legal rights generally, and evidentiary provisions specifically, are subject to waiver by voluntary agreement of the parties,” the Supreme Court has held that, “absent some affirmative indication that the agreement was entered into unknowingly or involuntarily, an agreement to waive the exclusionary provisions of [these] Rules is valid and enforceable.”

Here, of course, the plea agreement was never accepted or approved by the district court before Puig disavowed it, and under this line of authority, the agreement would generally be deemed not to be enforceable by the court. 

Viewing the language of Puig’s plea agreement against the backdrop of this case law, we conclude that Puig’s Rule 410 waiver was not triggered here. As we have explained, Puig’s waiver of the protections of Rule 410 (and Rule 11(f)) was expressly contingent on the district “[c]ourt’s finding” that there was a “breach of this agreement.” The terms of that waiver are most naturally understood as requiring that there be an “agreement” that, under our case law, was enforceable by the “[c]ourt[]” and as to which the court could therefore make the requisite “finding” of a “breach.” And because the plea agreement was a Type A agreement requiring the district court’s approval, and because that approval never occurred, the agreement was not enforceable by the court under our precedent. The waiver, by its own terms, therefore did not apply. Consequently, Rule 410 remains applicable with full force here, and the factual basis of Puig’s plea agreement is “not admissible against” Puig. FED. R. EVID. 410(a).

Also in FN 3, the Court explains that the factual basis is not admissible even for impeachment: "Although the Government argued in its opening brief that the factual basis of Puig’s plea agreement should at least be admissible at trial for impeachment purposes, the Government clarified in its reply brief that it was not contending that, even if Rule 410 applies, the Government is nonetheless entitled to a carve-out from that rule’s prohibitions if the evidence is used only for impeachment. Rather, the Government has clarified that its only argument on this score is that, if this court concludes “that public policy prohibits the admission of the factual basis in the government’s case-in-chief, [the court] should at least permit admission of the [factual basis] for impeachment and rebuttal.” Because we do not rely on “public policy,” but on the plain text of Rule 410, we have no authority to create exceptions to that rule’s terms. And because the Government concedes that Rule 410’s terms bar admission of covered statements for any purpose, including impeachment, no impeachment exception is applicable here."

5/23/25: Two Ninth Circuit decisions today

In United States v. Greene, --- F.4th ---, No. 23-4097 (9th Cir. 2025), the Court reversed the district court’s use of a “pseudo-count” of carjacking to calculate Namir Malik Ali Greene’s offense level at sentencing in a case in which Greene pleaded guilty to one count of Hobbs Act robbery.  Here are some key points from the opinion. 

Namir Malik Ali Greene challenges his 120-month sentence for interference with commerce by robbery (Hobbs Act robbery). 18 U.S.C. § 1951(a). He argues that the district court erred by basing his Sentencing Guidelines range on a “pseudo-count” of carjacking, 18 U.S.C. § 2119, because the elements of federal carjacking were not specifically established by his plea agreement. U.S. Sent’g Guidelines Manual (USSG) § 1B1.2(c) (U.S. Sent’g Comm’n 2023). We agree. Applying plain error review, we reverse and remand for resentencing on an open record.

Greene argues on appeal that the district court erred when it used the carjacking pseudo-count to calculate his offense level because the plea agreement did not establish all elements of federal carjacking.

The first two prongs of the plain error test are satisfied because the plea agreement did not specifically establish all elements of federal carjacking; in particular, the stipulated facts do not specifically establish that Greene acted with the “intent to cause death or serious bodily harm,” which is the mens rea required for federal carjacking. 18 U.S.C. § 2119; USSG § 1B1.2(c). This error was plain under Supreme Court precedent. Braxton v. United States, 500 U.S. 344, 350–51 (1991).

A sentencing court “must consider only conduct included in the plea agreement” when selecting a Guideline under § 1B1.2(a). That conduct must “specifically establish[]” all elements of the base offense. USSG § 1B1.2(c). To determine whether a plea agreement “specifically establishes” an offense, the Supreme Court has explained that we are to interpret a plea agreement the same way we would interpret a contract.

Braxton instructs that a plea agreement must evince a defendant’s stipulation to facts that unambiguously establish the required elements of a separate offense in order for the separate offense to be treated as “specifically established” for purposes of sentencing. As the Court explained, “a stipulation that at best supports two reasonable readings” does not specifically establish an additional offense.

Greene’s plea agreement stipulated only that “using the BB-gun to intimidate a vehicle owner, [he] stole a vehicle, namely, a 2010 Honda Accord.” This stipulation does not specifically establish the mens rea for § 2119 carjacking because it does not unambiguously support the conclusion that Greene had the specific “intent to cause death or serious bodily harm.”

Because the stipulated facts fell short of establishing all elements of a § 2119 carjacking, this offense should not have served as the base offense for the calculation of Greene’s Guidelines range.

Our case law does not permit district courts to infer a defendant’s mental state from ambiguous facts in a plea agreement, even if the inference is “reasonable,” as the government argues. And without facts specifically establishing the requisite mens rea, it was error to treat § 2119 carjacking as a pseudo-offense for the calculation of Greene’s offense level. After Braxton, this error was plain.

Turning to the third prong of the plain error test, we are persuaded that the error in the Guidelines calculation affected Greene’s substantial rights. By relying on a base offense of § 2119 carjacking, the court calculated Greene’s total offense level as 30, two levels higher than the total offense level would have been if based on the Hobbs Act robbery conviction. 

[T]he broad authority of the Commerce Clause does not permit the federal government to “convert congressional authority under the Commerce Clause to a general police power of the sort retained by the States.”

Turning to the fourth prong of plain error review, we exercise our discretion to correct the error because, on the facts of this case, the miscalculated base offense level “seriously affect[s] the fairness, integrity, or public reputation of judicial proceedings.”

Next, in United States v. Watson, --- F.4th ---, No. 24-1865 (9th Cir. 2025), the Court affirmed the district court’s denial of Tyler Jay Watson’s motion to suppress incriminating statements he made to police officers and the fruits thereof in a case in which Watson entered a conditional guilty plea to one count of possession with intent to distribute fentanyl.  Here are some key points from the opinion. 

Based on information received from a reliable confidential informant (“CI”), a police task force in Nampa, Idaho, began investigating Defendant-Appellant Tyler Jay Watson for drug distribution. When law enforcement learned that Watson was on parole, they coordinated with Probation and Parole (“P&P”) officers to conduct a compliance search of Watson’s vehicle and residence. The CI had advised officers that Watson transported drugs in magnetized containers under his vehicle. After officers found methamphetamine attached to the vehicle’s undercarriage, they drove to Watson’s residence and conducted a search. While the search was ongoing, Watson remained detained in Officer Scott’s patrol vehicle parked down the street. A police officer, Detective Coronado, approached Watson in the back of the patrol car and read Watson his Miranda rights. Watson acknowledged his rights and stated his willingness to cooperate. Watson then admitted he was holding more of “the product” at his grandmother’s home. Following Watson’s confession, officers drove to Watson’s grandmother’s house and obtained her consent to search her garage. Officers discovered and seized fentanyl, methamphetamine, and cash.

On appeal, Watson argues that his Agreement of Supervision’s (“Agreement”) condition requiring that he “cooperate with the requests of [his] probation/parole officer,” where “[c]ooperation includes being truthful,” created a “classic penalty situation” in which Watson was compelled—under threat of parole revocation—to make incriminating statements to law enforcement. We disagree. Watson’s statements were made to a police officer, not his P&P officer, after an adequate Miranda warning.

Because Watson’s Agreement required cooperation and truthfulness with his parole officer, but not all law enforcement officers, we cannot conclude that a Mirandized interrogation by police in the course of investigating a new, separate offense was involuntarily compelled. Accordingly, we affirm the district court’s denial of Watson’s motion to suppress and hold that Watson was not subject to a penalty situation under these circumstances.

The Fifth Amendment’s privilege against self-incrimination generally applies only to those who “claim it.” However, this general rule does not apply when an individual is “denied the free choice to admit, to deny, or to refuse to answer.” Id. This can occur when the government creates a situation where “an individual’s refusal to answer incriminating questions subjects him to a penalty.” Id. In a “penalty situation,” the Fifth Amendment becomes self-executing. In other words, “if the state, either expressly or by implication, asserts that invocation of the privilege would lead to revocation of probation . . . the failure to assert the privilege would be excused, and the probationer’s answer would be deemed compelled and inadmissible in a criminal prosecution.” 

Watson was not subject to a penalty situation because nothing in Watson’s Agreement required that he speak, cooperate, or be truthful with law enforcement; he was properly Mirandized immediately prior to his incriminating statements, stated he understood his rights, and agreed to cooperate; and he was never told that refusing to answer officers’ questions would result in the revocation of his parole or any other penalty. Because Watson’s statements were not involuntarily compelled under penalty of parole revocation, the district court properly denied Watson’s motion to suppress.

5/22/25: SCOTUS decision on fraudulent-inducement theory of wire fraud under 18 U. S. C. §1343.

In Kousisis v. United States, 605 U.S. ___, No. 23-909 (2025), the Court affirmed the defendant's convictions for wire fraud based on a fraudulent-inducement theory.  Here are some key parts of the majority opinion. 

Stamatios Kousisis and the industrial-painting company he helped manage, Alpha Painting and Construction Co., secured two government contracts for painting projects in Philadelphia. Both contracts required the participation of a disadvantaged business—and in its bids for the projects, Alpha represented to the Pennsylvania Department of Transportation (PennDOT) that it would obtain its materials from a qualifying supplier. See 49 CFR §§26.21(a), 26.5 (2024). This promise turned out to be an empty one: In addition to using the supplier solely as a pass-through entity, Alpha and Kousisis submitted multiple false certifications to cover up their scheme. So although Alpha’s paint work met expectations, its adherence to the disadvantaged business requirement did not. 

The Government charged Alpha and Kousisis with wire fraud, asserting that they had fraudulently induced PennDOT to award them the painting contracts. See 18 U. S. C. §1343. Under the fraudulent-inducement theory, a defendant commits federal fraud whenever he uses a material misstatement to trick a victim into a contract that requires handing over her money or property—regardless of whether the fraudster, who often provides something in return, seeks to cause the victim net pecuniary loss. We must decide whether this theory is consistent with §1343, which reaches only those schemes that target traditional money or property interests. It is, so we affirm.

To convict Alpha and Kousisis, the Government needed to prove that they used the wires to execute a “scheme or artifice to defraud, or for obtaining money or property by means of false or fraudulent pretenses, representations, or promises.” 18 U. S. C. §1343. Despite the use of the disjunctive “or,” we have declined to interpret §1343 as establishing alternative pathways to a conviction. Instead, reading the two clauses together, we have held that “the moneyor-property requirement of the latter phrase” operates as a limitation on the former.   A defendant commits federal wire fraud, in other words, only if he both “‘engaged in deception’” and had “‘money or property’” as “‘an object’” of his fraud. 

The fraudulent-inducement theory is consistent with both the text of the wire fraud statute and our precedent interpreting it. We therefore reject petitioners’ proposed economic-loss requirement.

In short, the wire fraud statute is agnostic about economic loss. The statute does not so much as mention loss, let alone require it. Instead, a defendant violates §1343 by scheming to “obtain” the victim’s “money or property,” regardless of whether he seeks to leave the victim economically worse off. A conviction premised on a fraudulent inducement thus comports with §1343. 

To summarize, [] common-law courts did not uniformly condition an action sounding in fraud on the plaintiff ’s ability to prove economic loss. More specifically, if the action was one for rescission or a prosecution for false pretenses, the plaintiff ’s required “injury” ordinarily need not be financial. That sounds the death knell for Alpha and Kousisis’s reliance on the common law. The old-soil principle does not apply in the absence of a well-settled rule. Kemp, 596 U. S., at 539. In Pasquantino v. United States, for example, we refused to read “the wire fraud statute to except frauds directed at evading foreign taxes” because the relevant common-law rule did not “clearly ba[r] such a prosecution.” So too here: The common law did not establish a generally applicable rule that all fraud plaintiffs must plead and prove economic loss, so we will not read such a requirement into the wire fraud statute.

[T]he common law has long embraced a different standard—namely, materiality—as the principled basis for distinguishing everyday misstatements from actionable fraud. Whether in tort or contract law, “materiality look[s] to the effect on the likely or actual behavior of the recipient of the alleged misrepresentation.” Resembling a but-for standard, materiality asks whether the misrepresentation “constitut[ed] an inducement or motive” to enter into a transaction. Or, as we explained in Universal Health Services, a misrepresentation is material if a reasonable person would attach importance to it in deciding how to proceed, or if the defendant knew (or should have known) that the recipient would likely deem it important.

For now, it is enough to reiterate “that materiality of falsehood is an element of ”—and thus a limit on—the federal fraud statutes.  A conviction premised on the fraudulent-inducement theory cannot be sustained without it.

The “demanding” materiality requirement substantially narrows the universe of actionable misrepresentations. And the boundaries of the fraudulent-inducement theory are not so imprecise as to risk encroachment on States’ authority or to “create traps” for the “unwary.” Snyder v. United States, 603 U. S. 1, 15 (2024). Rather, the theory criminalizes a particular species of fraud: intentionally lying to induce a victim into a transaction that will cost her money or property. As Judge Learned Hand put it, “[a] man is none the less cheated out of his property, when he is induced to part with it by fraud, because he gets a quid pro quo of equal value.” 

5/15/25: Case on loss = gain under the Guidelines

In United States v. Yafa, --- F.4th ---, No. 23-4108 (9th Cir. 2025), the panel affirmed the district court’s reliance on Application Note 3(B) in the commentary to United States Sentencing Guidelines § 2B1.1, which, at the time the Yafa brothers were sentenced, instructed courts to use the gain that resulted from the defendant’s offense as an alternative measure for calculating loss where loss cannot reasonably be determined.

The case arises from an alleged “pump-and-dump” stock manipulation scheme.  The brothers were convicted of securities fraud and conspiracy to commit securities fraud. 

At sentencing, the district court relied on Application Note 3(B) in the commentary to United States Sentencing Guidelines (“U.S.S.G.”) § 2B1.1, which instructs courts to use the gain that resulted from the defendant’s offense as an alternative measure for calculating loss where loss cannot reasonably be determined. U.S.S.G. § 2B1.1 cmt. n.3(B). 1 On appeal, the Yafas contend that it was legal error to defer to § 2B1.1’s commentary because the term “loss” is not genuinely ambiguous. Because we hold that deference is appropriate, we conclude that the district court did not err when it relied on the commentary and used “gain” as an alternative measure for the “loss” attributable to the Yafas. 

Because the Yafas were sentenced before Application Note 3(B) was moved from the commentary to the Guidelines, we must determine whether the district court’s reliance on the commentary was appropriate. To determine whether it is appropriate to defer to the commentary’s interpretation of a Guideline, we apply the analysis set forth in Kisor v. Wilkie, 588 U.S. 558 (2019). Castillo, 69 F.4th at 655–56.4 Kisor instructs that courts owe deference to an agency’s interpretation of its own rules where (1) the regulation is “genuinely ambiguous,” (2) the interpretation is “reasonable,” and (3) the interpretation is entitled to “controlling weight.” 588 U.S. at 574–79. Section 2B1.1’s commentary instructing courts to use gain as an alternative measure of loss satisfies these requirements.

First, the term “loss” is genuinely ambiguous. A court may only conclude that a regulation is genuinely ambiguous after “exhaust[ing] all the traditional tools of construction.” Id. at 575 (internal quotation marks omitted). In other words, it must consider “the text, structure, history, and purpose of a regulation, in all the ways it would if it had no agency to fall back on.” Id. Although “[w]e have not previously held that the term ‘loss’ under § 2B1.1 is genuinely ambiguous,” United States v. Hackett, 123 F.4th 1005, 1012 (9th Cir. 2024), several of our sister circuits have reached this conclusion applying Kisor. See, e.g., United States v. You, 74 F.4th 378, 397–98 (6th Cir. 2023) (holding that “loss” is genuinely ambiguous); United States v. Boler, 115 F.4th 316, 328–29 (4th Cir. 2024) (same). But see United States v. Banks, 55 F.4th 246, 257–58 (3d Cir. 2022) (holding that the ordinary meaning of “loss” controls). 

Second, Application Note 3(B)’s instruction to use gain is a reasonable interpretation of “loss.” An agency’s interpretation is reasonable if it “come[s] within the zone of ambiguity the court has identified after employing all its interpretive tools.” Kisor, 588 U.S. at 575–76. As the foregoing analysis demonstrates, the zone of ambiguity for the term “loss” stretches, at a minimum, along a spectrum from the actual, calculable loss experienced by victims of an economic crime to the far broader harms involved in and arising out of a defendant’s criminal conduct. Application Note 3(B) falls within this range. Because federal theft and fraud statutes can “cover a broad range of conduct,” U.S.S.G. § 2B1.1, cmt. (background), relying on the amount of gain enables courts to “maintain[] sufficient flexibility” and issue sentences that “avoid[] unwarranted sentencing disparities among defendants . . . who have been found guilty of similar criminal conduct,” even where the actual loss resulting from a defendant’s conduct is difficult to calculate, 28 U.S.C. § 991(b)(1)(B). 

Finally, the character and context of Application Note 3(B) “entitles it to controlling weight.” 

Application Note 3(B)’s interpretation of “loss” warrants deference under Kisor. Therefore, the district court did not err when it used the gain that resulted from the Yafas’s offenses as an alternative measure for loss. 

Of note:  "our conclusion is limited to Application Note 3(B) and says nothing about the reasonableness of Application Note 3(A)’s instruction defining loss as the 'greater of actual loss or intended loss.'"

5/9/25: En banc felon-in-possession decision

In United States v. Duarte, ---. F4th ---, No. 22-50048 (9th Cir. 2025), the en banc Court affirmed Steven Duarte’s conviction for felon in possession of a firearm in violation of 18 U.S.C. § 922(g)(1). 

The opinion is long with several concurrences and a partial dissent.  But the essence of the decision is as follows:

Duarte argues that § 922(g)(1) is unconstitutional as applied to non-violent felons like him under Bruen’s analytical framework. While this is an issue of first impression for our court, we do not write on a blank slate, as Courts of Appeals across the nation have been wrestling with fresh challenges to the viability of § 922(g)(1) in the wake of Bruen. Four circuits have upheld the categorical application of § 922(g)(1) to all felons. See United States v. Hunt, 123 F.4th 697, 707–08 (4th Cir. 2024) (rejecting an as applied challenge on a categorical basis); United States v. Jackson, 110 F.4th 1120, 1129 (8th Cir. 2024) (same); Vincent v. Bondi, 127 F.4th 1263, 1265–66 (10th Cir. 2025) (rejecting an as-applied challenge because neither Bruen nor United States v. Rahimi, 602 U.S. 680 (2024), abrogated circuit precedent foreclosing such a challenge); United States v. Dubois, 94 F.4th 1284, 1293 (11th Cir. 2024), cert. granted, judgment vacated, No. 24-5744, 2025 WL 76413 (U.S. Jan. 13, 2025) (holding that Bruen did not abrogate circuit precedent foreclosing such challenges).

Other circuits have rejected as-applied challenges, but have left open the possibility that § 922(g)(1) might be unconstitutional as applied to at least some felons. See United States v. Diaz, 116 F.4th 458, 471 (5th Cir. 2024) (rejecting an as-applied challenge because the defendant’s underlying felony was sufficiently similar to a death-eligible felony at the founding); United States v. Williams, 113 F.4th 637, 661–62 (6th Cir. 2024) (rejecting an as-applied challenge because the defendant’s criminal record sufficiently showed that he was dangerous enough to warrant disarmament). By contrast, the Third Circuit has held that § 922(g)(1) is unconstitutional as applied to a felon who was convicted of making a false statement to secure food stamps. See Range v. Att’y Gen., 124 F.4th 218, 222–23 (3d Cir. 2024) (en banc). And, as of the date of this writing, the First and Second Circuits have declined to address constitutional challenges to § 922(g)(1) on the merits, while the Seventh Circuit has yet to definitively resolve an as-applied challenge. See United States v. Langston, 110 F.4th 408, 419–20 (1st Cir. 2024) (rejecting an as-applied challenge because there was no “plain” error); United States v. Caves, No. 23-6176-CR, 2024 WL 5220649, at *1 (2d Cir. Dec. 26, 2024) (same); United States v. Gay, 98 F.4th 843, 846–47 (7th Cir. 2024) (assuming for the sake of argument that there is some room for an as-applied challenge, but rejecting the defendant’s specific as-applied challenge because his prior felonies included aggravated battery of a peace officer and possession of a weapon while in prison). 

Today, we align ourselves with the Fourth, Eighth, Tenth and Eleventh Circuits and hold that § 922(g)(1) is not unconstitutional as applied to non-violent felons like Steven Duarte.  

In sum, these [historical] laws demonstrate that § 922(g)(1)’s permanent and categorical disarmament of felons is consistent with this Nation’s historical tradition of firearm regulations. Legislatures have historically retained the discretion to punish those who commit the most severe crimes with permanent deprivations of liberty, and legislatures could disarm on a categorical basis those who present a “special danger of misuse” of firearms. Rahimi, 602 U.S. at 698. We agree with the Fourth and Eighth Circuits that either historical tradition is sufficient to uphold the application of § 922(g)(1) to all felons. See Jackson, 110 F.4th at 1127–28; Hunt, 123 F.4th at 706. 

Section 922(g)(1) “is by no means identical to these [historical laws], but it does not need to be.” Rahimi, 602 U.S. at 698. History does not require “felony-by-felony litigation” to support the application of § 922(g)(1). Jackson, 110 F.4th at 1125; Hunt, 123 F.4th at 700. Instead, consistent with our historical tradition, the government is “empowered to regulate guns through categorical restrictions.” Atkinson, 70 F.4th at 1038 (Wood, J., dissenting).

Finally, we recognize that these historical principles “may allow greater regulation than would an approach that employs means-end scrutiny with respect to each individual person who is regulated.” Jackson, 110 F.4th at 1129. However, these are the fruits of Bruen’s constitutional test. See id.; see also Heller v. District of Columbia, 670 F.3d 1244, 1274 (D.C. Cir. 2011) (Kavanaugh, J., dissenting) (“[G]overnments appear to have more flexibility and power to impose gun regulations under a test based on text, history, and tradition than they would under strict scrutiny.” (emphasis omitted)). 

5/8/25: Case on implied breach

In United States v. Plancarte, --- F.4th ---, No. 24-327 (9th Cir. 2025), the Court affirmed the district court's finding that the government did not breach the plea agreement, enforced the appellate waiver, and dismissed Erika Marie Plancarte’s appeal in a case in which she pleaded guilty to conspiracy to transport an alien into the United States.

Erika Marie Plancarte pleaded guilty to one count of conspiracy to transport an alien into the United States. The plea agreement bound the government to recommend a sentence of 90 days of imprisonment. We hold that the government did not implicitly breach the plea agreement by referencing Plancarte’s criminal history, expressing concern about Plancarte’s conduct and recidivism, clarifying an ambiguity in the presentence report, and declining to present mitigating evidence.

Criminal plea agreements “are essentially contracts between the government and a defendant.” Id. at 28. As such, either party can breach the agreement by violating its terms. United States v. Myers, 32 F.3d 411, 413 (9th Cir. 1994) (per curiam). For example, the government cannot agree “to recommend a sentence at the low end of the applicable guideline range,” but make “no recommendation” at all. Id. at 412, 413. Doing so violates “the terms of the plea agreement” and requires reversal and remand for resentencing. Id. at 413. In addition to complying with the literal terms of the contract, Farias-Contreras, 104 F.4th at 28, the parties must also comply with the “spirit of the plea agreement,” id. at 31. That means the parties’ arguments “must be made in good faith and advance the objectives of the plea agreement.” Id. We have compared this to contract law, which “implies a covenant of good faith and fair dealing in every contract.” Id. (citing Appling v. State Farm Mut. Auto. Ins. Co., 340 F.3d 769, 779 (9th Cir. 2003)). “This is a fact-specific inquiry based on contract principles.” Id. In conducting this inquiry, courts must consider the “totality of circumstances.” Id.

The government may implicitly breach a plea agreement by making arguments to the district court that undermine the parties’ agreement. But our court, sitting en banc, expressly disclaimed finding a “per se” implicit breach of a plea agreement when the government “present[s] any information already known and contained in the” PSR. Id. The government is “entitled” to do so when “respond[ing] to arguments” raised by the defense. Id. 

The government may undermine the plea agreement if it introduces “information that serves no purpose but to influence the court to give a higher sentence,” id. at 28 (quoting United States v. Whitney, 673 F.3d 965, 971 (9th Cir. 2012) (cleaned up)), or if it “purports to make the promised recommendation while ‘winking’ at the district court to impliedly request a different outcome,” id. (quoting United States v. Heredia, 768 F.3d 1220, 1231 (9th Cir. 2014)). One such example of bad faith behavior includes making inflammatory or pejorative comments about a defendant’s past offenses, such as analogizing drug dealers to vampires. See id. at 28–29. Or, the government may act in bad faith by inviting “the district court’s skepticism as to its recommendation” by noting a difference of opinion in the prosecutor’s office as to the correct sentence. Id. at 29. A court may consider such conduct to weigh in favor of finding that the government implicitly breached the plea agreement. 

On the other hand, so long as the government’s arguments are not precluded by the literal text of the plea agreement, id. at 30–31, the government may make arguments in support of its sentencing recommendation, including by highlighting aspects of the defendant’s offense or criminal history, see id. at 29; cf. United States v. Minasyan, 4 F.4th 770, 780–81 (9th Cir. 2021) (holding that the government did not implicitly breach the plea agreement by reiterating facts from the PSR that were relevant to the defendant’s sentencing). And when the defendant argues for a below-Guidelines sentence, the government may oppose the defendant’s argument and supplement the facts with relevant information, including by repeating facts in the PSR. Farias-Contreras, 104 F.4th at 30. However, the “government does not have carte blanche to use inflammatory rhetoric,” id. at 29, and the “government’s response [to a defendant’s request for a sentence lower than what the government recommends] must be tethered to its obligations under the plea agreement,” id. at 31.

The government did not implicitly breach its plea agreement with Plancarte. Instead, the government stuck to the letter of the agreement, and it did not make statements in bad faith to undermine the parties’ bargain. The district court properly reached its own conclusion as to Plancarte’s sentence, as anticipated by the plea agreement. Therefore, we enforce the appellate waiver in the plea agreement, and we dismiss the appeal.

5/6/25: Taking money from inmate trust accounts to cover restitution

In United States v. Myers, --- F.4th ---, No. 23-1034 (9th Cir. 2025), a divided court affirmed the district court’s order granting the government’s motion to turn over certain funds in Ronald Myers’s inmate trust account and apply them to Myers’s restitution obligation.

This decision could have a significant impact on any clients that receive, over time, substantial trust account deposits.  

The Mandatory Victims Restitution Act requires an inmate who “receives substantial resources from any source, including inheritance, settlement, or other judgment” to put such resources toward unpaid restitution. 18 U.S.C. § 3664(n). The question here is whether this provision applies to the gradual accumulation of cash deposits from family and friends in an inmate’s trust account. We hold that it does. Because § 3664(n) authorizes a district court to turn over periodic deposits that substantially accrue in an inmate’s account, we affirm.

Myers, like other federal inmates, has a trust account maintained by the Federal Bureau of Prisons (BOP). See 28 C.F.R. §§ 506.1, 506.2. Since 2013, over $30,500 has been deposited in Myers’s account. Most deposits ($27,872) were from family and friends. The rest ($2,747) were prison wages. Myers claims he saved some of the money, but records show he spent most of it. Over nine years, Myers donated $1,580 to charity, sent $1,334 to other individuals, and spent about $128 on subscriptions. He spent the remainder at the prison commissary. As of late 2022, Myers’s account contained $1,622.

Myers still owes his victims over $35,000 in restitution. So when the government discovered the activity on Myers’s trust account, it asked the district court to direct BOP to turn over most of the remaining funds and apply them to Myers’s obligation. The government disclaimed any efforts to target Myers’s prison wages. But as to the accumulated deposits from family and friends, the government invoked the Mandatory Victims Restitution Act (MVRA), which requires an inmate who “receives substantial resources from any source, including inheritance, settlement, or other judgment, . . . to apply the value of such resources to any restitution or fine still owed.” Pub. L. No. 104-132, § 206(a), 110 Stat. 1227, 1235–36 (1996) (codified at 18 U.S.C. § 3664(n)). 

The district court granted the turnover motion.

On appeal, the majority held:  "Section 3664(n) provides that if an inmate “receives substantial resources from any source, including inheritance, settlement, or other judgment,” he “shall be required to apply the value of such resources to any restitution or fine still owed.” 18 U.S.C. § 3664(n). In interpreting this provision, we face two related questions. Is § 3664(n) restricted to payments from a single source? And does the statute’s reference to “inheritance, settlement, or other judgment” limit its application to sudden financial windfalls? Addressing each question in turn, we hold that § 3664(n) applies not just to one-time financial windfalls, but also to substantial aggregated sums from multiple sources—like family and friends—that gradually accrue in an inmate’s trust account."

Judge McKeown wrote a compelling dissent: 

Under 18 U.S.C. § 3664(n), a defendant who “receives substantial resources from any source, including inheritance, settlement, or other judgment, during a period of incarceration” is required to “apply the value of such resources to any restitution or fine still owed.” The most logical interpretation of this statute is that it applies only to resources that are substantial at the time of receipt.

The text of § 3664(n), though brief, contains key terms— “receives,” “substantial resources,” “from any source,” “including inheritance, settlement, or other judgment.” Although the statute applies during a defendant’s incarceration, it must be read in conjunction with 18 U.S.C. § 3664(k), which requires both incarcerated and released defendants to notify the court and the Attorney General “of any material change in the defendant’s economic circumstances that might affect the defendant’s ability to pay restitution.” Both subsections serve the common policy goal of ensuring victims receive restitution.

Section 3664(n) is susceptible to competing interpretations. It might apply only to windfalls, in line with Myers’s position and the view taken by the First, Fifth, and Sixth Circuits. See United States v. Saemisch, 70 F.4th 1, 6 (1st Cir. 2023) (“[T]he MVRA requires certain defendants to apply to their restitution obligation any sudden windfalls they receive.”); United States v. Hughes, 914 F.3d 947, 951 (5th Cir. 2019) (“[W]e think [§ 3664(n)] refers to windfalls or sudden financial injections.”); United States v. Carson, 55 F.4th 1053, 1056 (6th Cir. 2022) (describing § 3664(n) as a “windfall provision”). This approach interprets the phrase “including inheritance, settlement, or other judgment” to cabin the nature of the source and limit the provision to singular, large, and unanticipated gains. Alternatively, as the majority urges, § 3664(n) could apply to any substantial aggregated amount, including amounts originating from multiple distinct sources over any period. No appellate court has adopted this view. Or § 3664(n) could apply to a substantial amount, from any source, so long as the amount is substantial at the time it is received. 

I think the last interpretation—that substantiality is measured at the time of receipt—is the most straightforward and logical reading of the text based on § 3664(n)’s use of the present tense “receives.” The parties’ and the majority’s arguments in favor of the two other alternatives—that § 3664(n) applies only to windfalls, or that it applies to aggregated amounts—do not persuade me otherwise.