1. Introduction
In United States v. Sow, the Second Circuit affirmed a sentence arising from a substantial counterfeit-luxury-goods operation. The decision addresses four connected questions: whether Congress’s sentencing policy for certain first offenders requires a noncustodial sentence; what evidence supports a supervisory-role enhancement; how courts may estimate the value of counterfeit merchandise; and whether trademark owners may recover government-requested appraisal and hearing expenses under the Mandatory Victims Restitution Act (“MVRA”).
The defendant, Adama Sow, pleaded guilty without a plea agreement to trafficking in counterfeit goods under 18 U.S.C. § 2320(a)(1)–(2). Her operation used multiple units at Gotham Mini-Storage in Manhattan. Searches conducted in October 2023 uncovered more than 80,000 counterfeit items, including clothing, shoes, and accessories bearing luxury-brand marks.
After a two-day evidentiary sentencing hearing, the district court applied a two-level aggravating-role enhancement and a twenty-level enhancement based on the value of the counterfeit goods. A total offense level of 27 and Criminal History Category I produced an advisory imprisonment range of 70 to 87 months. The court nevertheless imposed a substantially lower sentence: 24 months’ imprisonment, three years of supervised release, and $111,419.32 in restitution to affected brands.
The appellate decision preserves individualized sentencing discretion while approving evidence-based estimates and compensation for necessary expenses incurred by victims in assisting the government’s investigation and prosecution.
3. Analysis
A. Appellate Review: Legal Questions and Factual Findings Receive Different Treatment
The court reviewed procedural sentencing reasonableness for abuse of discretion. Within that framework, it considered legal questions independently and reviewed factual findings only for clear error. The restitution award was subject to a comparable division between legal and factual review.
Precedents cited and their influence
- Gall v. United States: Supplied the overarching abuse-of-discretion standard for reviewing procedural sentencing reasonableness. The appellate court’s task was to identify legally significant sentencing errors, not to select a sentence afresh.
- United States v. Vinales: Explained that this standard incorporates de novo review of legal questions, including Guidelines interpretation, and clear-error review of factual questions. That distinction governed the different treatment of Sow’s statutory arguments and evidentiary objections.
- United States v. Reifler and United States v. Harris: Established the generally deferential abuse-of-discretion review applicable to MVRA restitution orders.
- United States v. Afriyie: Clarified that restitution orders nevertheless receive independent review on legal conclusions and clear-error review on factual findings. Thus, deference did not prevent the court from examining whether the claimed expenses were legally compensable.
This framework materially affected the outcome. Sow could obtain independent consideration of her statutory interpretations, but disagreement with the district judge’s assessment of witnesses or choice among reasonable valuation methods did not itself establish reversible error.
B. Section 994(j) Does Not Create an Individual Entitlement to Avoid Prison
The statutory distinction
Section 994(j) directs the Sentencing Commission to ensure that the Guidelines reflect the “general appropriateness” of a sentence other than imprisonment for a first offender whose conviction is neither violent nor otherwise serious. Sow argued that her Guidelines range conflicted with that policy and that the district court had deferred too heavily to the Guidelines.
The Second Circuit rejected the premise that the provision commands a particular sentence. Its addressee is the Sentencing Commission, and its language expresses a general policy. A district judge may consider that policy when assessing the persuasive force of a Guidelines provision, but must ultimately impose an individualized sentence under 18 U.S.C. § 3553(a).
Why the district court’s reasoning survived review
The sentencing record contradicted Sow’s claim of “blind deference.” The district court called the Guidelines range “draconian,” concluded that it was substantially longer than necessary, and imposed a sentence far below it. At the same time, the judge considered Sow’s sustained counterfeit business sufficiently serious to warrant imprisonment.
The district court also relied on Sow’s supervisory role. It observed that U.S.S.G. § 4C1.1 excludes defendants who receive an aggravating-role enhancement from the first-offender-related reduction otherwise available under that provision. The appellate court found no indication that the judge misunderstood her discretion to impose a lower sentence.
The holding should not be overstated. The opinion does not establish that every first offender with a leadership enhancement falls outside the statutory phrase “first offender,” nor does it comprehensively define an “otherwise serious offense.” It establishes that § 994(j) does not compel noncustodial punishment and that the district court permissibly exercised its sentencing discretion here.
Precedents cited and their influence
- United States v. Cavera: Provided the central individualized-sentencing principle. A court must consider all § 3553(a) factors and select a sentence sufficient, but not greater than necessary, to serve the purposes of punishment. It cannot presume that the Guidelines sentence is reasonable. The substantial downward variance and case-specific explanation demonstrated compliance with that obligation.
- United States v. Fernandez: Supported the presumption that a sentencing judge knows and correctly applies the law. The Second Circuit used that presumption to reject Sow’s assertion that the judge misunderstood her authority.
- United States v. Ayers: Appeared through the quotation in United States v. Fernandez expressing that same presumption of judicial knowledge. Its relevance was to the judge’s understanding of discretion, not to any presumption favoring a Guidelines sentence.
- Rita v. United States: Was identified as having abrogated United States v. Fernandez on other grounds. The opinion preserved that qualification while relying on the separate proposition concerning a judge’s presumed understanding of applicable law.
The distinction between these presumptions matters: a court may be presumed to understand its legal authority, but it may not presume that a Guidelines sentence is appropriate for the particular defendant.
The court left unresolved whether Sow’s statutory argument was subject to plain-error review because of an inadequate objection below. It found no error even under independent review.
C. Credited Evidence of Supervising One Participant Supports the Role Enhancement
The governing rule and its application
Section 3B1.1(c) provides a two-level increase when a defendant acted as an organizer, leader, manager, or supervisor in criminal activity. The enhancement does not require a large hierarchy: supervision of one other participant can suffice.
Two Gotham employees and paid government informants, Fernando and Thomas Velazquez, described Sow’s relationship with an assistant. Their testimony included observations that the assistant regularly worked with her, followed her instructions, moved boxes between her units and customers, and functioned as someone for whose conduct she could be approached. The evidence also included statements identifying the assistant as her employee and describing payment for his work.
The district court found the brothers generally credible and credited their accounts of Sow’s supervisory role. The Second Circuit concluded that these concrete descriptions of direction and work performed were sufficiently specific to support the enhancement.
Hearsay, reliability, and discrepancies
Sow argued that the testimony contained unreliable hearsay and inconsistencies. The court distinguished sentencing from a trial governed by stricter evidentiary limitations: hearsay may be considered at sentencing, but due process requires minimal indications of reliability.
Much of the evidence came from firsthand observations. Statements concerning the assistant’s status and compensation were therefore considered alongside observed conduct, not in isolation. The district court also explained why discrepancies did not undermine the essential account.
Precedents cited and their influence
- United States v. Hertular: Defined a managerial or supervisory role through control over others or a significant role in recruiting or supervising lower-level participants. It supplied the operative test applied to Sow’s relationship with her assistant.
- United States v. Garcia: Demonstrated that involvement with one subordinate participant can be enough. Its example of recruiting one drug courier defeated any suggestion that a broad supervisory structure was necessary.
- United States v. Norman: Explained why substantial deference is owed to credibility determinations made by the judge who heard the testimony. This principle supported both the role finding and, later, the attribution of storage units to Sow.
- United States v. Gomez and United States v. Sisti: Supported the admissibility of hearsay in sentencing determinations. United States v. Gomez quoted United States v. Sisti for that proposition.
- United States v. Martinez: Supplied the constitutional limitation: hearsay must carry at least minimal indicia of reliability. The overlapping observations and statements in this record satisfied that requirement.
- United States v. O'Connor: Established that inaccuracies, contradictions, or even untruthfulness in parts of testimony do not necessarily require rejection of its essential account. This addressed Sow’s reliance on discrepancies.
- Anderson v. Bessemer City: Supplied the broader clear-error principle that a factfinder’s choice between two permissible views of the evidence cannot be clearly erroneous. The appellate court therefore declined to substitute its own credibility assessment.
The decision does not make paid-informant testimony automatically reliable. It confirms that such testimony can support a sentencing finding when the judge evaluates credibility and the account contains adequate, specific indications of reliability.
D. Counterfeit Valuation May Rest on Reasonable, Defendant-Specific Estimates
Attributing the seized goods
Sow disputed responsibility for approximately 85,000 items recovered from sixteen storage units. Although the district court acknowledged that Gotham’s records were disorganized and its procedures imperfect, it credited Fernando’s testimony identifying the relevant units as hers. His account included personal observations of Sow in those units and testimony that the records had not been falsified.
The court therefore did not treat imperfect business records as conclusive against attribution. It evaluated them together with witness testimony connecting the goods to Sow.
Estimating the value of the seized inventory
Michael Rieger, an employee of Allegiance Protection Group (“APG”), appraised the counterfeit merchandise. He used product categories and midpoint street-sale values for average-quality counterfeits, producing an estimate of $7,486,562.50. The district court discounted that estimate by ten percent to account for sales in Canal Street or similar markets rather than higher-priced venues.
The Second Circuit approved the methodology as a reasonable estimate grounded in the quantity and market characteristics of the goods. Importantly, the analysis concerned the estimated market value of the counterfeit items, not a finding that the brands had lost an equivalent amount in genuine-product sales.
Financial records and earlier trafficking
The district court also considered $3.1 million moving through Sow’s bank accounts during eight months of 2021. It used those records to estimate earlier trafficking activity and reasoned that, even without extrapolating comparable activity into later years, adding that amount to the discounted seized-inventory value exceeded $9.5 million.
The Guidelines commentary expressly permits reasonable infringement estimates using relevant information, including financial records, when the number of infringing items cannot be determined. The appellate court found the combined evidence sufficient for the twenty-level increase.
Numerical qualification: The opinion describes the discounted inventory estimate as approximately $6.6 million, although a strict ten-percent reduction of $7,486,562.50 produces approximately $6.74 million. Either figure, combined with $3.1 million, exceeds the operative $9.5 million threshold. The affirmance turns on that threshold, not on precision to the last dollar.
Precedent cited and its influence
United States v. Archer supplied the principal evidentiary safeguard. A quantity-based enhancement requires specific evidence both of the quantity involved and of its connection to the defendant. Circumstantial evidence can satisfy that requirement. Here, testimony linked the units to Sow, the seizure established the inventory, expert analysis supported valuation, and bank records supported the estimate of prior activity.
The decision consequently permits estimation, not speculation. Nor does it establish that every bank deposit in a counterfeiting case automatically counts toward the infringement amount; it approved the district court’s inference on this record.
E. Government-Requested Expert Expenses Can Constitute Restitution
Two distinct statutory questions
The MVRA analysis required the court to separate:
- Applicability: Was there a qualifying offense and an identifiable victim suffering the required injury or pecuniary loss?
- Compensability: Were the particular expenses sought recoverable and sufficiently caused by the offense?
Sow argued that the brands had not proved monetary loss apart from APG’s fees and that unquantified harm to trademarks could not satisfy the statute. The court rejected both arguments.
Trademarks and intangible property
The district court identified ways counterfeit distribution harms brands: making branded-looking products ubiquitous can undermine exclusivity, while poor-quality imitations can damage consumers’ perceptions. Sow had also admitted that the counterfeit marks were likely to cause confusion, mistake, or deception.
The Second Circuit emphasized that property protection under the MVRA is not limited to tangible objects. Trademarks are property, and their intangible character does not exclude their owners from victim status. On this record, the court accepted the intangible harm as properly considered pecuniary loss and approved recovery of the associated, identifiable expenses.
The award was not a judicial estimate of lost prestige or hypothetical lost sales. It reimbursed $111,419.32 in concrete expenses incurred for APG’s services.
Necessity and causation
The government had specifically requested APG’s authentication and appraisal work. The expenses related exclusively to analyzing the number and value of goods attributable to Sow and participating in the evidentiary hearing. The district court found those services necessary.
Sow’s operation also had a sufficiently close connection to the investigation. As Gotham’s most prolific counterfeiter, she materially contributed to the circumstances supporting the search. Investigation of the counterfeit merchandise, including expert examination, was a foreseeable consequence of the conduct.
The opinion’s reference to Sow having requested the hearing does not make restitution a punishment for challenging sentencing evidence. The material findings were that the services were necessary, government-requested, and causally connected to the offense.
Precedents cited and their influence
- United States v. AvenaĴi: Supplied both the distinction between MVRA applicability and compensable loss and the requirement that recoverable loss be a direct and proximate result of the criminal conduct. This structured the restitution analysis.
- United States v. Bengis: Rejected a conception of protected property confined to tangible assets. It supported treating holders of intangible property rights as potential MVRA victims.
- United States v. Milstein: Specifically recognized trademarks as property and intellectual or intangible property as within criminal statutes protecting property. This provided the closest doctrinal connection between trademark injury and restitution eligibility.
- United States v. Afriyie: Also supported the substantive award through its approval of legal fees arising from the government’s criminal investigation and prosecution of insider trading. The analogy was that necessary expenses of assisting governmental proceedings can be recoverable even though they are not the underlying property deprivation itself.
- United States v. Diakhoumpa: Provided factually relevant support because it affirmed restitution for brands’ costs of investigating counterfeit sales. The opinion identified it as a summary order, distinguishing its status from a published precedential opinion.
- United States v. Goodrich: Explained the proximate-causation inquiry in terms of a sufficiently close connection and foreseeability. It supported the conclusion that investigation and appraisal expenses were foreseeable consequences of Sow’s substantial counterfeiting operation.
- United States v. Sow: The appellate opinion also cited the district court’s separate restitution decision, 2025 WL 2802547. That decision supplied the detailed findings about governmental requests, necessary services, and the connection between Sow’s operation and the investigation. It was the ruling under review, not an independent appellate precedent.
4. Potential Impact and Limits
Individualized sentencing remains central
First offenders may invoke § 994(j) as a policy argument, but cannot treat it as an exemption from imprisonment. Conversely, sentencing judges remain obligated to consider whether the Guidelines overstate the punishment necessary in the individual case. The decision validates neither automatic imprisonment nor automatic probation.
Sentencing records must connect evidence to the defendant
The role and valuation rulings demonstrate the practical importance of a developed evidentiary record. Specific observations, reasoned credibility findings, expert methodology, and financial records can overcome gaps in business documentation. Generalized assertions about the scale of an operation remain insufficient under the defendant-specific requirement emphasized in United States v. Archer.
Intellectual-property victims may recover qualifying participation costs
The restitution ruling is particularly significant for counterfeiting prosecutions, where brand representatives or specialists may be needed to authenticate and value seized products. Proven lost sales are not the only possible foundation for relief when trademark injury and necessary, causally connected expenses are established.
Nevertheless, the opinion does not authorize reimbursement of every private brand-protection expense. Its reasoning depends on government-requested work, necessity, attribution to Sow’s goods, and direct and proximate causation. It does not resolve the recoverability of unrelated monitoring expenses, speculative future costs, or investigations undertaken independently of government proceedings.
Guidelines valuation and restitution serve different purposes
The multimillion-dollar infringement estimate measured offense seriousness for sentencing. The much smaller restitution award compensated specified expenses. Future courts should not equate the Guidelines infringement amount with an automatically recoverable victim-loss figure.