AVAA Did Not Displace Paroline: NCMEC Hash-Based Proof and the “Relative Role” Restitution Method Remain the Governing Framework in § 2259 Child-Pornography Restitution

I. Introduction

In United States v. Moises Abraham Sotelo (11th Cir. Mar. 10, 2025), the Eleventh Circuit affirmed (1) a 121-month prison sentence for receipt of child pornography and (2) a $30,000 restitution award to seven depicted victims. The consolidated appeals raised two core questions:

  1. Sentencing: whether the district court abused its discretion by refusing the parties’ joint 84-month recommendation and imposing 121 months instead.
  2. Restitution after the AVAA amendments to 18 U.S.C. § 2259: whether the Amy, Vicky, and Andy Child Pornography Victim Assistance Act of 2018 (“AVAA”) changed the procedure for calculating restitution—particularly whether courts must quantify total “viewings” and whether the traditional evidentiary approach (including reliance on NCMEC reports) remains permissible.

The parties were the United States (Plaintiff–Appellee) and Moises Abraham Sotelo (Defendant–Appellant). Sotelo admitted using a LiveMe profile (“piper954”) to facilitate trading child sexual abuse material and possessed thousands of files, including depictions of very young children and sadistic/masochistic content.

II. Summary of the Opinion

The Eleventh Circuit held:

  • Sentence affirmed: the 121-month sentence—already below the advisory Guideline range—was substantively reasonable. The district court considered relevant 18 U.S.C. § 3553(a) factors, including the nature of the images, the volume of material, and Sotelo’s “moderator” role.
  • Restitution affirmed: the AVAA did not alter the basic Paroline v. United States approach for trafficking/possession restitution under § 2259. Courts need not find the total number of image viewings to determine a defendant’s “relative role.” The district court could rely on NCMEC and law-enforcement materials under the established reliability standard.
  • Evidentiary holding: the NCMEC report and associated spreadsheet bore sufficient indicia of reliability; the defendant had an opportunity to challenge them.
  • Awareness not required: victims need not already know their images are online to receive restitution, particularly where losses are reasonably projected.

III. Analysis

A. Precedents Cited

1. Child-pornography restitution causation and methodology

  • Paroline v. United States, 572 U.S. 434 (2014): The central pillar. The Supreme Court recognized that the victim’s losses are caused by the “broader causal process” of trafficking and possession, making traditional but-for allocation impossible. It directed district courts to use “discretion and sound judgment” guided by non-rigid “guideposts” (often called the “Paroline factors”) to set an amount reflecting the defendant’s “relative causal role.”
    Use in Sotelo: The Eleventh Circuit treated Paroline as still governing after the AVAA. It rejected Sotelo’s claim that AVAA requires a new, more quantifiable metric (such as total viewings).
  • Rothenberg v. United States, 923 F.3d 1309 (11th Cir. 2019): The court’s in-circuit benchmark for review and proof. It emphasizes deference to restitution amounts and permits awards based on “reasonable estimates” supported by reliable evidence. It also approves reliance on victim-counsel declarations to substantiate losses.
    Use in Sotelo: The Eleventh Circuit applied Rothenberg to (i) confirm the deferential abuse-of-discretion review of amounts and (ii) uphold several awards based on expert reports and counsel declarations.
  • United States v. Osman, 853 F.3d 1184 (11th Cir. 2017): Confirms future expenses (e.g., therapy) may be included if reasonably estimated and record-supported.
    Use in Sotelo: Supports restitution to victims with projected future costs, including where victims are currently unaware of online circulation.
  • United States v. Danser, 270 F.3d 451 (7th Cir. 2001): Cited (via Osman) for Congress’s intent to provide expansive relief for the “full amount” of losses.
    Use in Sotelo: Reinforces the breadth of compensable loss categories under § 2259.
  • United States v. Groover, 2021 WL 3205719 (11th Cir. July 29, 2021): Example of similar restitution outcomes for a particular victim referenced in the record.
    Use in Sotelo: Used to show the district court’s award fit within a familiar range in Eleventh Circuit practice.
  • United States v. Sainz, 827 F.3d 602 (7th Cir. 2016): Referenced for allowing restitution for future lost earnings.
    Use in Sotelo: Supports the inclusion of lost income and future earnings estimates within “full amount of losses.”

2. Evidentiary standards for restitution and hearsay

  • United States v. Hairston, 888 F.2d 1349 (11th Cir. 1989) (quoting United States v. Rodriguez, 765 F.2d 1546 (11th Cir. 1985)): Restitution determinations may consider hearsay if the defendant can refute it and it bears “minimal indicia of reliability.”
    Use in Sotelo: The court held the NCMEC report and spreadsheet met this standard, particularly given corroboration across the record and the role of NCMEC.

3. NCMEC’s role and hash identification context

  • United States v. McDaniel, 631 F.3d 1204 (11th Cir. 2011): Describes NCMEC’s identification/notification function.
    Use in Sotelo: The Eleventh Circuit relied on this background to contextualize why NCMEC-based identification is integral to modern enforcement and restitution.
  • United States v. Ackerman, 831 F.3d 1292 (10th Cir. 2016): Discusses NCMEC’s statutory obligations and quasi-clearinghouse function.
    Use in Sotelo: Supports the institutional reliability and statutory role of NCMEC.
  • United States v. Miller, 982 F.3d 412 (6th Cir. 2020): Explains “hash values” as digital fingerprints used to match known child sexual abuse material without repeatedly viewing the content.
    Use in Sotelo: Provides technical grounding for why a hash-based NCMEC report can reliably link files to known series/victims.

4. Sentencing reasonableness framework

  • United States v. Livesay, 587 F.3d 1274 (11th Cir. 2009): Abuse-of-discretion review for substantive reasonableness.
    Use in Sotelo: Sets the appellate lens for reviewing the 121-month sentence.
  • United States v. Irey, 612 F.3d 1160 (11th Cir. 2010) (en banc) (quoting United States v. Campa, 459 F.3d 1121 (11th Cir. 2006) (en banc)): Defines abuse-of-discretion errors in weighing § 3553(a) factors.
    Use in Sotelo: Frames why disagreement with the district court’s weighting is not enough to show unreasonableness.
  • United States v. Ramirez-Gon- zalez, 755 F.3d 1267 (11th Cir. 2014) and United States v. Shabazz, 887 F.3d 1204 (11th Cir. 2018): Weighting of § 3553(a) factors is within the district court’s discretion; challenger bears the burden.
    Use in Sotelo: Supports affirmance where the district court made a record-based judgment call.
  • United States v. Rosales-Bruno, 789 F.3d 1249 (11th Cir. 2015) (quoting Ledford v. Peeples, 605 F.3d 871 (11th Cir. 2010)): Appellate review asks whether the sentence falls “in the ballpark of permissible outcomes.”
    Use in Sotelo: The “ballpark” framing is the decisive response to Sotelo’s variance arguments.
  • United States v. Duperval, 777 F.3d 1324 (11th Cir. 2015): Sentencing disparity comparisons require similarly situated defendants.
    Use in Sotelo: Undercuts Sotelo’s reliance on a generalized “usual” seven-year statement without comparable case details.
  • United States v. Castaneda, 997 F.3d 1318 (11th Cir. 2021): Within-Guidelines sentences are ordinarily expected to be reasonable; below statutory maximum is an indicator of reasonableness.
    Use in Sotelo: Reinforces affirmance where the sentence was below both the Guideline range and the statutory maximum.

5. Plain error and constitutional arguments raised late

  • United States v. Bobal, 981 F.3d 971 (11th Cir. 2020) and United States v. Lejarde-Rada, 319 F.3d 1288 (11th Cir. 2003): No “plain error” absent explicit statutory language or binding precedent directly resolving the issue.
    Use in Sotelo: The court refused to find plain error on Sotelo’s new argument that AVAA’s $3,000 floor triggers heightened proof requirements.
  • Apprendi v. New Jersey, 530 U.S. 466 (2000) and Alleyne v. United States, 570 U.S. 99 (2013): Jury-finding requirements for facts increasing statutory maximums/minimums in criminal punishment.
    Use in Sotelo: The court did not reach the merits because the argument was forfeited and not plainly resolved by precedent in this restitution context.
  • Dohrmann v. United States, 442 F.3d 1279 (11th Cir. 2006): Referenced as part of the Eleventh Circuit’s existing landscape suggesting Apprendi limits may not map neatly onto restitution.
    Use in Sotelo: Helps explain why there was no plain error.
  • United States v. Sperrazza, 804 F.3d 1113 (11th Cir. 2015): Invoked by analogy to reject (on plain-error review) an excessive-fines challenge to the $3,000 minimum as not grossly disproportionate.

B. Legal Reasoning

1. Sentencing: why the district court could reject the joint recommendation

The panel emphasized the breadth of district court discretion under § 3553(a). The sentencing judge:

  • Accepted the Guideline calculations and expressed that the offense level was “absolutely appropriate.”
  • Varied downward primarily because criminal history category II overstated Sotelo’s history (driven by a DUI and probation status).
  • Declined a larger variance due to aggravating facts: extraordinary file volume, depictions of infants/toddlers, sadistic material, and Sotelo’s group moderator role facilitating sharing.

The Eleventh Circuit treated the government’s “usual seven years” comment as too unparticularized to establish unwarranted disparity, especially without evidence that those “usual” cases shared Sotelo’s aggravators. The 121-month sentence—below the Guideline range and far below the 240-month statutory maximum—fell within the permissible range of outcomes.

2. Restitution after AVAA: Paroline’s methodology remains operative

The opinion’s principal doctrinal contribution is its clear statement that the AVAA amendments to § 2259 did not replace the Paroline framework. The court reasoned:

  • The amended statute requires restitution reflecting the defendant’s “relative role in the causal process.” That concept tracks Paroline’s “relative causal role” inquiry.
  • AVAA adds a floor ($3,000) and clarifies “full amount of losses,” including reasonably projected future costs, but it does not demand a new algorithm.
  • The court rejected Sotelo’s proposed mandatory finding of “total viewings,” noting both practical impossibility and legal mismatch: neither AVAA nor Paroline requires a precise mathematical apportionment.
  • Congress’s findings in AVAA expressly referenced Paroline, signaling ratification rather than displacement.

3. Proof of victim identification and possession: NCMEC report reliability

Sotelo argued the NCMEC report/spreadsheet lacked reliability (unknown spreadsheet author; unclear compilation process). The court affirmed reliance because:

  • Restitution may be based on hearsay with “minimal indicia of reliability” where the defendant can refute it (United States v. Hairston; United States v. Rodriguez).
  • The NCMEC report is produced by the statutorily tasked entity and is hash-based, reducing the need for repeated viewing of contraband.
  • The report/spreadsheet were corroborated throughout the record: consistent series/victim names across claims and extensive supporting documentation (experts, counsel, impact statements).
  • Sotelo’s guilty plea and the factual record (including his role facilitating access/sharing) provided contextual corroboration.

4. Victim awareness of online circulation is not a precondition to restitution

The court held that § 2259 allows “reasonably projected” losses. Thus, victims need not already know their images are online, because future therapy/medical costs, legal costs, and other losses may be projected and compensable, and the statute covers more than therapy.

5. Amounts awarded: deference and “reasonable estimate” support

Applying Rothenberg v. United States, the court deferred to the district court’s judgment where the awards fell in a modest range ($3,000–$10,000) and each victim had evidence of significant total losses. The panel also emphasized the AVAA minimum: for several victims, even disputes about high-end economic projections would not undermine the entitlement to (at least) $3,000 where total losses plainly exceeded prior recoveries.

C. Impact

  • Stability of § 2259 practice in the Eleventh Circuit: The decision cements that AVAA did not require courts to reinvent restitution methodology; litigants should expect continued use of Paroline guideposts and deferential appellate review under Rothenberg.
  • Rejects “total viewings” as a mandatory evidentiary prerequisite: This is practically significant because the internet-scale dissemination problem makes comprehensive viewer counts speculative at best and impossible at worst.
  • Endorses NCMEC hash-based identification as sufficient (with corroboration): This encourages restitution proceedings that do not require victims to repeatedly authenticate or relive the content and reduces incentives for defendants to demand victim-intensive proof.
  • Procedural lesson on constitutional challenges: By applying plain-error review to newly raised Apprendi/Alleyne and Eighth Amendment theories, the opinion signals that defendants must timely preserve challenges to AVAA’s restitution floor if they hope to obtain merits review.
  • Victim-centered approach to “projected” losses: The ruling confirms that minors or otherwise unaware victims are not excluded from restitution, consistent with § 2259’s forward-looking loss language.

IV. Complex Concepts Simplified

What changed under the AVAA?
AVAA amended § 2259 to (1) require courts to determine the victim’s “full amount of losses,” including reasonably projected future costs, (2) order an amount reflecting the defendant’s “relative role,” and (3) set a minimum of $3,000 per victim (subject to a cap at demonstrated losses). Sotelo holds that these changes do not replace the Paroline approach; they largely codify and supplement it.
What are the “Paroline factors”?
They are non-mandatory “guideposts” to help a judge decide a restitution amount when many offenders collectively contribute to the victim’s harm. Examples include how many images the defendant possessed, whether he distributed them, and how many offenders are likely involved. They are not a rigid formula.
What is a “hash value,” and why does it matter?
A hash value is like a digital fingerprint of a file. Matching hash values allow identification of known illegal images/videos without repeatedly opening them. Here, it supported linking Sotelo’s files to known victim series via NCMEC’s database.
What does “minimal indicia of reliability” mean?
In restitution proceedings, courts can consider evidence that might be inadmissible at trial (including hearsay), as long as it has basic signs of trustworthiness and the defendant can challenge it.
Why wasn’t the court required to find the total number of times an image was viewed?
Because § 2259 (as amended) and Paroline require an amount reflecting a defendant’s relative role, not a precise mathematical allocation. Total viewings are typically unknowable and not a statutory prerequisite.
Why could victims recover restitution even if they don’t know their images are online?
The statute allows “reasonably projected” losses, and restitution covers many categories (medical, therapy, attorneys’ fees, lost income, and other losses), not solely distress tied to current awareness.

V. Conclusion

United States v. Moises Abraham Sotelo is a consolidating decision for post-AVAA § 2259 restitution in the Eleventh Circuit. It confirms that AVAA’s amendments—despite adding a $3,000 minimum and elaborating loss categories—do not displace Paroline v. United States or require courts to undertake impracticable “total viewings” findings. It also approves reliance on NCMEC hash-based identification materials under the established reliability standard and reinforces that restitution may be based on reasonably projected future harms, even for victims not yet aware of online dissemination. On sentencing, the decision reiterates the breadth of district court discretion to reject joint recommendations when the record supports a higher sentence.