Backend Check-Processing as Wire-Fraud/Identity-Theft Participation; Narrow Brady/Giglio Materiality; Sampling-Based Loss Upheld with Conservative Discount (11th Cir.)

I. Introduction

Thomas Addaquay controlled United Consolidated Accounting and Business Services (“UC”), nominally a check-cashing business. The government’s theory was that UC’s true function was to convert refund-anticipation-loan checks generated by fraudulent tax returns into spendable funds via a third-party processor, ReliaFund. A jury convicted Addaquay in the “tax-refund-fraud case” on twenty-nine counts including wire fraud and aggravated identity theft; later, Addaquay pleaded guilty in a separate “structuring case,” and the district court sentenced him in both matters together.

On appeal, Addaquay pressed four principal issues: (1) sufficiency of the evidence on selected wire-fraud and aggravated-identity-theft counts; (2) post-trial disclosure of four investigative memoranda under Brady v. Maryland, 373 U.S. 83 (1963), and Giglio v. United States, 405 U.S. 150 (1972); (3) denial (in part) of post-trial motions to compel additional discovery; and (4) loss and IRS restitution calculations based on sampling/extrapolation and an IRS “Transaction Code 841” proxy.

II. Summary of the Opinion

The Eleventh Circuit affirmed across the board. It held:

• Sufficiency: The evidence supported wire-fraud counts premised both on ReliaFund settlement transfers and on electronic filing of unauthorized returns, and supported related aggravated identity theft counts, including on an aiding-and-abetting theory.
• Brady/Giglio: No reversible Brady or Giglio violation occurred because the belatedly produced Awiti and Liady materials were not material in light of the trial record and did not establish perjury.
• Discovery: The district court reasonably compelled production of tax returns already in the government’s possession for loss-testing purposes, but denied broader investigative files and K1’s client list as speculative and insufficiently tied to specific loss items.
• Loss/Restitution: The district court’s $4,123,474.55 loss and matching IRS restitution figure—built from a defense-favorable median-based extrapolation using Transaction Code 841 and then reduced by an additional 25% “in the abundance of caution”—was a reasonable estimate supported by the record.

III. Analysis

A. Precedents Cited

1. Standards for sufficiency of evidence

• United States v. Deason, 965 F.3d 1252 (11th Cir. 2020) and United States v. Merrill, 513 F.3d 1293 (11th Cir. 2008): framed the appellate posture—evidence viewed in the light most favorable to the verdict and convictions affirmed unless no reasonable construction supports guilt beyond a reasonable doubt.
• United States v. Jiminez, 564 F.3d 1280 (11th Cir. 2009) (citing United States v. Williams, 144 F.3d 1397 (11th Cir. 1998)): reiterated the “light most favorable to the verdict” lens for recounting evidence.

2. Wire fraud elements and “causing” use of wires

• United States v. Machado, 886 F.3d 1070 (11th Cir. 2018): supplied the three-part wire fraud test (scheme, intent, use/causation of interstate wires) and endorsed circumstantial proof of intent.
• United States v. Ward, 486 F.3d 1212 (11th Cir. 2007) and United States v. Watkins, 42 F.4th 1278 (11th Cir. 2022): crucially supported the court’s rejection of Addaquay’s “I didn’t personally send/file” defense. Under these authorities, the government need not prove the defendant personally performed every element if he knowingly joined the scheme and a co-schemer used the mails/wires to execute it.
• United States v. Williams, 527 F.3d 1235 (11th Cir. 2008): used to emphasize the government’s burden is count-specific (the charged transmissions must be in furtherance of the scheme), not an all-checks-must-be-fraudulent showing.

3. Aiding-and-abetting doctrine

• United States v. Hewitt, 663 F.2d 1381 (11th Cir. 1981) (quoting United States v. Martinez, 555 F.2d 1269 (5th Cir. 1977)): provided the classic tripartite formulation—association with the venture, participation as something wished to bring about, and action to make it succeed; no need to participate in every phase.
• United States v. Schwartz, 666 F.2d 461 (11th Cir. 1982): reinforced that aiding/abetting requires more than mere association; it requires knowing participation aimed at success.

4. Out-of-circuit comparison and non-binding precedent

• United States v. Ragan, 24 F.3d 657 (5th Cir. 1994): Addaquay’s principal reliance; the Eleventh Circuit distinguished it because Ragan lacked evidence tying the defendant to the charged transactions, whereas here each charged wire was traced to unauthorized returns/checks and through UC/ReliaFund to accounts the jury could find Addaquay controlled.
• Bonner v. City of Prichard, 661 F.2d 1206 (11th Cir. 1981) (en banc): invoked for the proposition that decisions of other circuits are not binding on the Eleventh Circuit.

5. Aggravated identity theft and aiding/abetting “advance knowledge”

• United States v. Barrington, 648 F.3d 1178 (11th Cir. 2011): supplied the elements for 18 U.S.C. § 1028A (knowing use, without lawful authority, of another person’s means of identification during and in relation to a predicate felony, here wire fraud).
• Rosemond v. United States, 572 U.S. 65 (2014): Addaquay’s key authority; the court applied it to hold the evidence permitted inference that Addaquay knew identities were being used while he still had the meaningful choice to withdraw—given his role in receiving PII, coordinating with “runners,” and directing forged endorsements.

6. Brady/Giglio materiality and perjury thresholds

• United States v. Jordan, 316 F.3d 1215 (11th Cir. 2003) and United States v. Gallardo, 977 F.3d 1126 (11th Cir. 2020): set review standards (de novo for Brady/Giglio; abuse of discretion for dismissal/new trial).
• United States v. Noriega, 117 F.3d 1206 (11th Cir. 1997) (quoting Kyles v. Whitley, 514 U.S. 419 (1995)): anchored the “reasonable probability” test and the “undermine confidence in the verdict” formulation; Kyles also drove the court’s required cumulative-evidence assessment.
• United States v. Vallejo, 297 F.3d 1154 (11th Cir. 2002): provided the Eleventh Circuit’s four-part Brady framework (possession, defense lack/diligence, suppression, materiality).
• Ford v. Hall, 546 F.3d 1326 (11th Cir. 2008): articulated Giglio as a “species of Brady” involving perjury and prosecutorial knowledge, with materiality framed as “any reasonable likelihood” the falsehood affected the judgment.
• United States v. Jones, 601 F.3d 1247 (11th Cir. 2010) and United States v. Michael, 17 F.3d 1383 (11th Cir. 1994): underpinned the court’s conclusion that additional impeachment that is cumulative or does not show falsity is typically not material and does not establish Giglio perjury.

7. Post-trial discovery and “Brady is not a discovery device”

• United States v. Espinosa-Hernandez, 918 F.2d 911 (11th Cir. 1990): set abuse-of-discretion review for post-trial discovery orders.
• United States v. Arias-Izquierdo, 449 F.3d 1168 (11th Cir. 2006) and United States v. Jordan, 316 F.3d 1215 (11th Cir. 2003): supported limiting compelled discovery where requests rest on speculation rather than concrete ties to material issues.

8. Loss and restitution estimation

• United States v. Medina, 485 F.3d 1291 (11th Cir. 2007); United States v. Bradley, 644 F.3d 1213 (11th Cir. 2011); United States v. Willis, 560 F.3d 1246 (11th Cir. 2009); United States v. Sepulveda, 115 F.3d 882 (11th Cir. 1997): collectively framed “reasonable estimate” vs impermissible speculation and permitted reliance on circumstantial evidence.
• United States v. Campbell, 765 F.3d 1291 (11th Cir. 2014): supported treating victim-to-enterprise transfers as a starting point where conduct is “permeated with fraud.”
• United States v. Robertson, 493 F.3d 1322 (11th Cir. 2007): required “substantial evidence” supporting factual loss findings even under deferential clear-error review.
• United States v. Foley, 508 F.3d 627 (11th Cir. 2007) and United States v. Stein, 846 F.3d 1135 (11th Cir. 2017): reinforced the government’s burden to prove loss by a preponderance.
• United States v. Presendieu, 880 F.3d 1228 (11th Cir. 2018) and United States v. McCrimmon, 362 F.3d 725 (11th Cir. 2004): governed relevant-conduct attribution for the acts of others (scope, furtherance, foreseeability) under U.S.S.G. § 1B1.3(a)(1)(B).
• United States v. Whitman, 887 F.3d 1240 (11th Cir. 2018) (quoting United States v. Moran, 778 F.3d 942 (11th Cir. 2015)): emphasized deference to district courts’ comparative advantage in evaluating loss evidence.
• United States v. Johnson, 841 F.3d 299 (5th Cir. 2016): cited as an example approving sampling/extrapolation in the tax-loss context.
• United States v. Gupta, 572 F.3d 878 (11th Cir. 2009): distinguished to show the district court here articulated an evidentiary basis for its adjustments rather than selecting an unsupported compromise number.
• United States v. Baldwin, 774 F.3d 711 (11th Cir. 2014) and United States v. Moss, 34 F.4th 1176 (11th Cir. 2022): framed restitution as actual-loss-based, but still allow reasonable estimation when precision is impracticable.

B. Legal Reasoning

1. Wire fraud: “scheme membership” plus “wires used to execute”

The court’s sufficiency analysis turned on an important framing choice: the case was not treated as requiring proof that Addaquay personally filed each fraudulent return or personally initiated each wire transfer. Relying on United States v. Ward and United States v. Watkins, the court asked whether the evidence proved (a) a fraudulent scheme, (b) Addaquay’s knowing participation (including via aiding and abetting), and (c) that the charged transmissions were uses of the wires in execution of that scheme.

The panel then anchored each challenged wire-fraud count to a specific transactional chain: (1) unauthorized return(s) (proved through victim taxpayers’ testimony), (2) resulting refund check(s), (3) UC submission to ReliaFund, (4) ReliaFund “settlement” transfers into UC accounts the jury could find Addaquay controlled. This linkage—count by count—was the direct answer to Addaquay’s Ragan argument that the government failed to connect him to the charged transactions.

2. Aggravated identity theft: evidence of contemporaneous knowledge under Rosemond

For 18 U.S.C. § 1028A, Addaquay argued that even if proceeds reached his accounts, that did not show he had the advance knowledge required by Rosemond v. United States. The court rejected this as an artificial “end-point only” view of the evidence. It emphasized proof that placed Addaquay earlier in the pipeline: possession of PII (files and emails), runner interactions (as described by Edwards), and directing employees to sign taxpayers’ names on refund checks. From that evidence, the jury could infer knowledge of identity misuse while Addaquay was still facilitating the scheme.

3. Brady/Giglio: materiality, cumulative impeachment, and “not perjury”

The court’s Brady/Giglio analysis is best understood as three layered determinations:

• No reasonable probability of a different result (Brady): The Awiti memorandum offered some impeachment of Edwards but did not contradict Edwards’s core testimony about Edwards’s own transactions with Addaquay; it also contained inculpatory statements (Awiti said Addaquay knew checks derived from stolen identities and helped disguise proceeds), and there was substantial independent evidence of Addaquay’s role (employees, PII, forged endorsements, UC operations, ReliaFund channel, victims).
• No showing of falsity (Giglio threshold): Citing United States v. Michael, the court treated discrepancies or additional details as insufficient; Giglio requires perjured testimony, not merely testimony that could have been impeached or supplemented.
• Cumulative assessment (Kyles): Even combined, the Awiti and Liady materials did not “place the case in a sufficiently different light” because Liady’s scheme was separate and unconnected to the charged victims/transactions, and Awiti’s memo did not meaningfully collapse the government’s proof of Addaquay’s knowing check-processing role.

4. Post-trial discovery: tailored production tied to loss, denial of speculative fishing

The district court’s split ruling—compel tax returns already in government possession but deny broader investigative files and K1’s client list—was affirmed as a proper exercise of discretion under United States v. Arias-Izquierdo. The appellate court accepted the district court’s distinction between:

• Directly loss-testing material: underlying tax returns for checks UC processed could bear on authorization, preparer identification, and sampling assumptions.
• Speculative material: broad investigative files and client lists not connected to specific checks in the loss calculation.

5. Loss and restitution: “reasonable estimate” through sampling, median, and conservative discount

The court upheld a multi-step loss process notable for its incremental caution:

• The district court rejected the government’s “all-checks/permeated” approach, signaling a distinction between trial guilt and sentencing loss scope.
• It accepted sampling/extrapolation in principle, but credited the defense statistician’s critique of the mean (outlier sensitivity) and a sampling aggregation error, adopting a median-based estimate.
• Recognizing Transaction Code 841’s imperfections, the court treated it as an indicator rather than a conclusive fraud label and then applied an additional 25% reduction “in the abundance of caution.”

On restitution, the court treated the same figure as a reasonable estimate of the IRS’s actual loss, consistent with United States v. Baldwin and United States v. Moss, given the impracticability of perfectly reconstructing all fraudulent refunds in a large dataset.

C. Impact

• Fraud “backend” liability: The opinion reinforces that a defendant who operates the conversion mechanism (here, check-processing through a merchant processor) can be convicted of wire fraud and identity theft even if he is not proved to be the return-filer—so long as the evidence supports knowing participation and that the charged wires executed the scheme.
• Count-by-count tracing as insulation against Ragan-type arguments: The court’s emphasis on tying each charged wire to unauthorized returns and to the UC/ReliaFund channel offers a roadmap for prosecutors and a warning to defendants that generalized “some checks were legitimate” arguments will not defeat sufficiency where charged counts are specifically supported.
• Brady/Giglio narrowing in practice: Late disclosure of impeachment material is unlikely to be “material” where the witness was already heavily impeached (immunity, criminal participation, bias), where the new material is at most incremental, and where it includes inculpatory content that would likely blunt its impeachment value.
• Sampling and statistics at sentencing: The case illustrates how a defendant can materially improve the loss outcome through statistical critique (mean vs median, sample integrity), while also confirming appellate tolerance for pragmatic estimation techniques (including administrative proxies like Transaction Code 841), especially when the district court adopts conservative adjustments and explains them.
• Discovery discipline post-trial: The affirmance of partial denial underscores that “Brady is not a discovery device” and that post-trial motions to compel must be tied to concrete, material issues (here, specific loss components), not speculative alternative-scheme theories.

IV. Complex Concepts Simplified

• Wire fraud (18 U.S.C. § 1343): It is enough that the defendant knowingly joined a fraudulent plan and that someone in the plan used interstate electronic communications to advance it. The defendant need not personally press “send” on the wire transfer or file the return.
• Aiding and abetting (18 U.S.C. § 2): Helping a crime succeed—by providing tools, access, processing, or coordination—can make a person guilty “as a principal” even if another person performs the headline act.
• Aggravated identity theft (18 U.S.C. § 1028A): Adds a mandatory consecutive penalty when someone knowingly uses a real person’s identifying information (like a Social Security number) during and in relation to certain felonies (including wire fraud).
• Rosemond’s “advance knowledge” idea: To be an aider/abettor, a person must learn enough about the crime while there is still time to stop participating. Here, evidence of Addaquay’s role earlier in the pipeline (PII handling; directing forged endorsements) allowed the inference that he knew identities were being misused before the scheme’s benefits were complete.
• Brady vs. Giglio: Brady concerns suppression of favorable evidence (including impeachment) that is material to outcome; Giglio is a subset focused on perjured testimony knowingly used or left uncorrected by the prosecution.
• Materiality (“reasonable probability”): Not “might have helped,” but “enough to undermine confidence in the verdict.” Cumulative impeachment often fails this standard when the jury already heard strong reasons to distrust the witness.
• Loss vs. restitution: Guidelines “loss” can include intended loss for sentencing; restitution is limited to victims’ actual losses. Still, both can be reasonably estimated when exact reconstruction is impractical.
• Sampling/extrapolation; mean vs median: When reviewing every transaction is infeasible, a random sample can be used to estimate totals. The mean is sensitive to extreme outliers; the median often better reflects a “typical” value in skewed fraud datasets.

V. Conclusion

United States v. Thomas Addaquay affirms a robust theory of fraud liability for defendants who control the “cash-out” infrastructure of a broader identity-theft and tax-refund scheme. The Eleventh Circuit endorsed a count-specific tracing method for wire-fraud proof, applied Rosemond to find sufficient evidence of contemporaneous knowledge for aggravated identity theft, and set a high bar for post-trial Brady/Giglio relief where new materials are incremental, partly inculpatory, and outweighed by substantial independent evidence.

On sentencing, the decision highlights both the acceptability of sampling-based estimates and the importance of methodological rigor: the district court’s adoption of a defense-favorable median-based extrapolation and additional conservative discount was treated as a “reasonable estimate” rather than speculation, supporting both the Guidelines loss enhancement and the IRS restitution award.