Date: Apr 20, 2026

TVPRA Restitution in Multi-Perpetrator Trafficking Requires Defendant-Specific But-For Proof Before Any Apportionment

I. Introduction

In United States v. Cifuentes-Lopez (10th Cir. Apr. 14, 2026), the Tenth Circuit reversed substantial restitution awards entered under the Trafficking Victims Protection Reauthorization Act (TVPRA), 18 U.S.C. § 1593, following the defendant’s guilty plea to two counts of child sex trafficking (patronizing two minors, D.H. and S.H., for commercial sex acts) in violation of 18 U.S.C. §§ 2, 1591, and 1594.

The core issue was causation: where multiple perpetrators trafficked and abused the same victims, what must the government prove to obtain restitution from one defendant? The defendant argued the government failed to prove that any specific losses “would not have occurred but for” his conduct, as required by United States v. Anthony (Anthony I) and United States v. Anthony (Anthony II).

The government pursued restitution largely through expert reports that aggregated the victims’ losses across multiple offenders, and it urged either joint-and-several liability for the entire loss or, alternatively, apportionment of aggregated losses. The district court rejected joint-and-several liability but still imposed apportioned restitution. The Tenth Circuit reversed, holding the government’s proof did not satisfy the defendant-specific but-for causation requirement and that apportionment cannot substitute for that missing proof.

II. Summary of the Opinion

The court held that Anthony I and Anthony II controlled: restitution may be ordered only for losses “actually resulting from the offense of conviction,” requiring proof that the defendant’s conduct directly (but-for) and proximately caused the losses.
In multi-perpetrator child sex trafficking cases, the government must differentiate the harms attributable to the defendant from harms attributable to other abusers—i.e., show that had the victim never encountered this defendant, the victim would not have suffered the losses sought.
The government’s evidence—Dr. Missar’s psychology reports and Edith Wong’s accounting-based calculations—was aggregative and did not link any component of the claimed losses to Cifuentes-Lopez’s specific conduct.
The district court’s decision to apportion percentages of the victims’ losses to the defendant could not cure the lack of defendant-specific but-for proof; § 3664(h) apportionment is not a substitute for causation.
The panel rejected the government’s attempt to characterize the defendants as acting in a common enterprise warranting broader responsibility because the district court made contrary factual findings and the government did not challenge them as clearly erroneous (waiver).
The panel limited the remand to vacating the restitution portions of the judgment and denied de novo resentencing because the government had been repeatedly alerted to the evidentiary deficiency and did not cure it.

III. Analysis

A. Precedents Cited

1. United States v. Anthony (Anthony I), 942 F.3d 955 (10th Cir. 2019)

Anthony I is the foundation for the panel’s approach to TVPRA restitution in multi-abuser scenarios. It emphasizes that restitution under § 1593 is limited to losses that “actually result[] from the offense of conviction” and requires proof that the defendant “directly and proximately caused” the losses. Critically, in cases involving multiple perpetrators, Anthony I requires the government to “differentiate the harms” caused by the defendant from the harms caused by others.

In Cifuentes-Lopez, the panel treated this differentiation requirement as an evidentiary mandate: generalized proof that the victim was harmed by a broader trafficking environment is not enough; the record must support a counterfactual specific to the defendant—what losses would have been avoided if the victim never encountered him.

2. United States v. Anthony (Anthony II), 22 F.4th 943 (10th Cir. 2022)

Anthony II supplies the decisive template for rejecting aggregated expert reports. In that case, the court held that an expert report failed because it did not identify treatment or losses needed solely due to the defendant’s offenses and instead “lump[ed] together” harms caused by multiple traffickers.

The panel found Dr. Missar’s reports here to be materially the same type of proof rejected in Anthony II: they described serious trauma and projected costs, but did so as a combined product of abuse by the mother and multiple men, without defendant-specific isolation of losses. Wong’s accounting adjustments did not fix the defect because she expressly testified she could not “link” losses to Cifuentes-Lopez’s “direct actions.”

3. United States v. Anthony (Anthony III), 39 F.4th 1247 (10th Cir. 2022) (Hartz & Eid, JJ., dissenting from the denial of rehearing en banc)

The panel acknowledged that some judges have criticized the Anthony I/Anthony II standard as too rigorous for cases involving “numerous perpetrators,” suggesting instead a framework like Paroline v. United States. But the panel reiterated it was bound by existing circuit precedent absent en banc review or a contrary Supreme Court decision.

4. Paroline v. United States, 572 U.S. 434 (2014)

Paroline addressed restitution for child pornography victims harmed by many viewers and recognized that some injuries are “indivisible,” allowing a form of apportionment akin to civil damages allocation. The panel did not apply Paroline, referencing it only to explain that—whatever its appeal in multi-perpetrator sexual abuse contexts— Anthony I and Anthony II control in the Tenth Circuit’s TVPRA trafficking restitution cases.

5. In re Smith, 10 F.3d 723 (10th Cir. 1993) (per curiam)

Cited for the binding-force rule: a panel is bound by prior panel precedent unless overruled en banc or superseded by the Supreme Court. This doctrinal constraint is central to the outcome; the panel portrayed its reversal as compelled by Anthony I and Anthony II.

6. United States v. Coulter, 133 F.4th 1083 (10th Cir. 2025)

The government attempted to rely on Coulter to support restitution. The panel distinguished it on factual and evidentiary grounds: in Coulter, the record showed “no confusion as to who” were the but-for causes (the traffickers proven at trial), and the victims’ prior losses were not of a similar sexual-trafficking nature, supporting a causal link. Here, by contrast, the district court found the defendants were not acting together, and the proof did not isolate defendant-specific losses amidst multiple perpetrators (including uncharged men).

7. United States v. Montano, 109 F.4th 1275 (10th Cir. 2024) and United States v. Forsythe, 437 F.3d 960 (10th Cir. 2005)

These cases supplied the basis to limit remand. Where the government bears the burden, is alerted to evidentiary deficiencies, and fails to cure them, the appellate court may remand with instructions confined to the existing record. The panel used this principle to deny the government a second chance to build but-for evidence.

8. 18 U.S.C. § 3742(f)(1) and United States v. Anthony, 25 F.4th 792 (10th Cir. 2022)

The panel cited § 3742(f)(1) to justify issuing limited remand instructions, and it cited United States v. Anthony, 25 F.4th 792 to confirm restitution is a component of a criminal sentence, making the sentencing-remand framework applicable.

9. United States v. Thomas, 749 F.3d 1302 (10th Cir. 2014)

The panel relied on Thomas for the proposition that the government should not receive another opportunity to make a record it failed to make earlier—reinforcing the limited remand to vacatur of restitution only.

B. Legal Reasoning

1. The governing causation rule: “direct and proximate,” with “direct” meaning but-for

The panel applied the TVPRA restitution framework as articulated in Anthony II: restitution can be ordered only for losses “actually resulting from the offense of conviction,” and the government must prove (by a preponderance) that the defendant’s conduct “directly and proximately caused” the losses. “Direct” causation is defined as but-for causation.

2. Multi-perpetrator requirement: differentiate harms by defendant

The decisive move is the panel’s insistence—again drawn from Anthony I and Anthony II—that where multiple independent perpetrators abused a victim, the government must present evidence allowing the court to conclude that, absent the defendant’s involvement, the victim would not have suffered the claimed losses (or would have suffered less). Aggregation defeats that inquiry because it leaves the court unable to identify which losses were “actually resulting” from the defendant’s offense of conviction.

3. Why the expert evidence failed

The panel’s critique was evidentiary, not moral: the wrongdoing was “reprehensible,” but restitution required proof. Dr. Missar’s reports quantified loss based on the cumulative trauma caused by the mother and multiple men; they did not identify losses caused by Cifuentes-Lopez’s conduct as distinct from losses caused by others.

Wong’s report began from Missar’s aggregated totals and applied financial adjustments, then attempted allocation. Critically, Wong conceded she could not link losses to Cifuentes-Lopez’s direct actions and could not estimate the counterfactual where the victims never encountered him. The panel therefore treated Wong’s work as analytically downstream of an already noncompliant causation model.

4. Apportionment under § 3664(h) cannot replace but-for proof

The district court attributed 35% of D.H.’s losses and 20% of S.H.’s losses to Cifuentes-Lopez under its own assessment of “level of contribution” and “economic circumstances” (18 U.S.C. § 3664(h)). The panel held this was “cart before the horse”: a court cannot apportion a loss to a defendant unless the government first proves that the defendant directly caused the loss (but-for causation) for which restitution is sought. In other words, equitable allocation is a second-stage decision; it does not supply the missing first-stage causation showing.

5. Joint-and-several/common enterprise theory rejected by factual findings and waiver

The government attempted to recast the case as a common enterprise making Cifuentes-Lopez responsible for the entire loss. The panel held the argument was foreclosed because the district court found the defendants were not “working together” but acting in self-interest, and it found no organizing/planning conduct by Cifuentes-Lopez. Because the government did not challenge these findings as clearly erroneous, the panel treated the issue as waived and the findings as dispositive.

6. Remedy: limited remand; no second bite at the apple

The panel denied the government’s request for de novo resentencing and instead remanded solely to vacate the restitution portions of the judgment. It emphasized: (i) the government bore the burden; (ii) the defendant repeatedly identified the precise causation defect (including reliance on Anthony I and later Anthony II); and (iii) the government did not cure the deficiency even after abatement and a restitution hearing. Under Montano, Forsythe, and Thomas, that justified limiting remand to the existing record and denying a do-over.

C. Impact

Heightened evidentiary burden in multi-perpetrator TVPRA restitution: The decision reinforces that, in the Tenth Circuit, aggregated “total harm” models are insufficient where victims were abused by multiple perpetrators. The government must produce evidence that isolates the defendant-specific increment of harm (or a defensible counterfactual).
Constrains the use of experts who quantify global trauma without attribution: Psychological and economic experts must connect specific categories of loss (therapy needs, lost earnings, education costs) to the defendant’s conduct, not merely to “the trafficking” as a generalized experience involving multiple actors.
Limits district courts’ equitable flexibility: Even though § 3664(h) contemplates apportionment among defendants, this decision underscores that apportionment is not a causation workaround; it is available only after the defendant-specific causation threshold is satisfied.
Strategic consequences for litigation: The opinion incentivizes defendants to make early, record-focused causation objections, and it pressures the government to build a defendant-specific causation record at the initial restitution stage, because appellate courts may restrict remand where deficiencies were flagged but not cured.
Doctrinal tension likely to persist: By acknowledging Anthony III’s reference to Paroline while adhering to Anthony I/Anthony II, the decision highlights an unresolved policy and doctrinal dispute: whether but-for differentiation is workable for indivisible harms in multi-perpetrator sexual exploitation cases.

IV. Complex Concepts Simplified

TVPRA restitution (18 U.S.C. § 1593): A mandatory restitution scheme for trafficking victims requiring payment of the “full amount of the victim’s losses,” but only for losses that legally qualify as resulting from the defendant’s offense of conviction under the governing causation standards.
But-for causation: A counterfactual test: would the victim have suffered this loss if the defendant had not committed the offense? If yes, the loss is not “directly caused” by the defendant under the Tenth Circuit’s framework.
Proximate causation: A limiting principle addressing whether the loss is sufficiently connected to the conduct (often framed in terms of foreseeability and the absence of intervening causes). In this case, the panel did not need to elaborate because it found the record failed at the but-for stage.
Aggregation problem: Evidence that totals the victim’s losses from multiple abusers without showing what portion is attributable to a particular defendant. Under Anthony II, such “lumping together” is inadequate to prove but-for causation in multi-perpetrator cases.
Apportionment (18 U.S.C. § 3664(h)): A mechanism allowing a court to divide restitution responsibility among multiple defendants to reflect “level of contribution” and “economic circumstances.” This opinion clarifies that apportionment is not available to fill a causation gap; it presupposes causation is proved.
Clear-error review / abuse of discretion: Appellate standards: factual findings stand unless clearly mistaken; restitution amounts are reviewed for abuse of discretion. A court abuses discretion if it relies on a legally erroneous view (here, apportioning without sufficient but-for proof) or clearly erroneous evidence assessment.
Waiver on appeal: If a party does not challenge critical factual findings or properly raise an issue, an appellate court may treat the argument as forfeited/waived. Here, the government did not challenge the district court’s findings rejecting a common enterprise theory.

V. Conclusion

United States v. Cifuentes-Lopez strengthens and operationalizes the Tenth Circuit’s Anthony I/Anthony II rule: in TVPRA child sex trafficking cases involving multiple perpetrators, restitution requires defendant-specific proof that the victim’s claimed losses would not have occurred absent that defendant’s conduct. Expert reports that quantify global trauma and global economic loss, without isolating the defendant’s causal contribution, are insufficient—even if a district court attempts to apportion the total through equitable judgment.

The opinion’s practical significance lies in its insistence that causation is a threshold requirement, not an equitable afterthought, and in its remedial message: where the government is repeatedly alerted to a proof defect and does not cure it, the appellate court may vacate restitution outright and deny a second opportunity to reconstruct the record.