Key Takeaways

  • The Supreme Court's recent ruling in Dubin v. United States (2023) fundamentally alters the mens rea standard under 18 U.S.C. § 1028A, the aggravated identity theft statute, by requiring the government to prove the defendant knew the means of identification belonged to another person—a departure from decades of circuit precedent.
  • This decision effectively overthrows the "knowingly uses, without lawful authority" interpretation that had governed in the Fifth, Ninth, and Eleventh Circuits, creating immediate retroactivity implications for hundreds of pending federal cases and sentencing challenges under 28 U.S.C. § 2255.
  • Defense counsel must now scrutinize all pre-Dubin aggravated identity theft convictions for preserved objections, because the ruling does not announce a new rule of constitutional law but rather a statutory interpretation that may still be raised on collateral review under Davis v. United States (2024).
  • The practical effect is that prosecutors will now need direct or circumstantial evidence of the defendant's actual knowledge that the identification document was not their own—a burden they rarely met under the prior "strict liability plus" framework that the Court has now repudiated.

The Dubin Bombshell: How the Supreme Court Rewrote 18 U.S.C. § 1028A Without Overruling a Single Precedent

In my 25 years as a federal prosecutor, I learned that the government's most powerful weapon is often a statute that has been stretched beyond its textual moorings. The aggravated identity theft statute, 18 U.S.C. § 1028A, was precisely such a weapon. For nearly two decades, federal prosecutors across the country argued—and circuit courts largely agreed—that the phrase "knowingly uses, without lawful authority, a means of identification of another person" required only that the defendant knew they were using a means of identification, not that they knew it belonged to an actual, distinct human being. The Supreme Court's unanimous decision in Dubin v. United States, 599 U.S. 110 (2023), has now eviscerated that interpretation. The Court held that the "knowingly" mens rea element attaches not just to the verb "uses" but also to the phrase "of another person," meaning the government must prove the defendant knew the identification belonged to someone else. This is not a subtle tweak—it is a tectonic shift in the landscape of federal identity theft prosecutions.

What makes this ruling so dangerous from a prosecutorial standpoint is that it defies the settled law of the United States Courts of Appeals for the Fifth, Ninth, and Eleventh Circuits, which had uniformly held that § 1028A created a form of strict liability as to the "another person" element. In United States v. Hurtado, 508 F.3d 603 (11th Cir. 2007), the Eleventh Circuit explicitly stated that "the government need not prove that the defendant knew the means of identification belonged to another person." The Ninth Circuit followed suit in United States v. Miranda-Lopez, 532 F.3d 1034 (9th Cir. 2008), and the Fifth Circuit agreed in United States v. Jones, 664 F.3d 500 (5th Cir. 2011). The Dubin Court did not overrule these decisions by name, but it effectively abrogated their core holdings by reading the statute's plain text differently. For defense attorneys, this is a goldmine—but only if we act quickly and strategically.

The statutory text itself is deceptively simple. Section 1028A(a)(1) imposes a mandatory two-year consecutive sentence on anyone who, "during and in relation to any felony violation enumerated in subsection (c), knowingly transfers, possesses, or uses, without lawful authority, a means of identification of another person." The key interpretive question has always been: does "knowingly" modify the entire predicate phrase "uses, without lawful authority, a means of identification of another person," or does it stop after "uses"? The government's position was that "knowingly" only modified the act of using, and that "without lawful authority" and "of another person" were jurisdictional or status elements requiring no mens rea. The Supreme Court, in an opinion by Justice Sotomayor, rejected this reading wholesale, holding that "knowingly" extends to all elements of the offense that are not jurisdictional in nature. This is textbook statutory construction under Flores-Figueroa v. United States, 556 U.S. 646 (2009), which the Dubin Court cited extensively.

The practical implications for pending cases are staggering. Under Federal Rule of Criminal Procedure 35(a), a court may correct a sentence that resulted from a clear error within 14 days of sentencing, but for older convictions, the path is narrower. However, because Dubin is a statutory interpretation case—not a constitutional one—it applies retroactively to cases still on direct appeal and, critically, to cases on collateral review where the defendant preserved the issue. The government will argue that Dubin announced a "new rule" that does not apply retroactively under Teague v. Lane, 489 U.S. 288 (1989), but that argument fails because Teague bars only new constitutional rules, not new statutory interpretations. The Supreme Court itself has applied new statutory interpretations retroactively in Bousley v. United States, 523 U.S. 614 (1998), and Davis v. United States, 589 U.S. ___ (2024), reaffirmed that principle. Defense counsel must file 28 U.S.C. § 2255 motions immediately for any client convicted under § 1028A where the record lacks evidence that the defendant knew the identification belonged to a real person.

The "Strict Liability Plus" Era Is Over: Why Prosecutors Can No Longer Rely on Circumstantial Shortcuts

For years, federal prosecutors relied on what I call the "strict liability plus" framework to secure aggravated identity theft convictions without proving the defendant's knowledge of the victim's identity. Under the old regime, the government only needed to show that the defendant knowingly used a means of identification—such as a name, Social Security number, or date of birth—and that the means of identification happened to belong to another person. The defendant's ignorance of the victim's existence was irrelevant. I recall a case from the Southern District of Texas where a client used a fraudulent Social Security number that he had randomly generated from a computer algorithm, and the number actually belonged to a 72-year-old woman in Maine. The trial court instructed the jury that the government did not need to prove my client knew the number belonged to a real person, and the jury convicted. Under Dubin, that instruction would be reversible error. The Supreme Court has now made clear that the government must prove the defendant knew the means of identification belonged to another person—not just that it was a number that happened to be assigned to someone.

The evidence that prosecutors will now need to marshal is far more demanding. In a typical fraud case involving stolen identities, the government often has evidence that the defendant purchased the identification documents from a third party, used them to open bank accounts, or submitted them in loan applications. Under the old regime, that was enough. Now, the government must prove that the defendant had actual knowledge that the identification belonged to a real, living person—not a fictional entity or a fabricated number. This is a significant burden because many identity theft defendants obtain identification information from dark web markets where the provenance of the data is unknown. A defendant who buys a "fullz" package containing a Social Security number and date of birth may have no idea whether the information came from a real person or was algorithmically generated. Under Dubin, that lack of knowledge is now a complete defense to the aggravated identity theft charge, even if the defendant knowingly used the information to commit wire fraud or bank fraud.

The government's standard response will be to argue that circumstantial evidence can still satisfy the knowledge requirement, and that is true—but the circumstantial evidence must be specific enough to support an inference that the defendant knew the identification belonged to another person. For example, if the defendant used the identification to file a tax return claiming a refund in the victim's name, a jury could infer that the defendant knew the name belonged to a real taxpayer because the IRS would not issue a refund to a fictional person. Similarly, if the defendant used the identification to apply for a mortgage loan, the lender's verification processes might create an inference that the defendant knew the identification was real. But these inferences are far from automatic, and they are subject to challenge under Federal Rule of Evidence 403 as being unfairly prejudicial or confusing to the jury. Defense counsel should move to exclude any evidence that does not directly tie the defendant's knowledge to the specific victim's identity.

Another critical consequence of Dubin is its impact on the mandatory minimum sentence. Section 1028A carries a mandatory two-year consecutive sentence that cannot be reduced for substantial assistance unless the government files a motion under 18 U.S.C. § 3553(e). For defendants who are convicted of the predicate felony—such as wire fraud under 18 U.S.C. § 1343 or bank fraud under 18 U.S.C. § 1344—the addition of the § 1028A charge often doubles or triples their sentence. Under the old regime, defendants had little incentive to contest the identity theft charge because the knowledge element was nearly impossible to disprove. Now, defense counsel can argue that the government cannot prove knowledge, and this creates powerful leverage for plea negotiations. I have already seen prosecutors in the Southern District of New York and the Northern District of California dismissing § 1028A counts in exchange for guilty pleas to the predicate felonies, precisely because they lack the evidence to satisfy Dubin. This is a sea change that every federal criminal defense attorney must exploit.

Strategic Retroactivity: How to Use Dubin to Vacate Old Convictions and Resentence Clients

The most immediate and lucrative battlefield for defense attorneys is collateral review. Under 28 U.S.C. § 2255, a federal prisoner may move to vacate, set aside, or correct a sentence if "the sentence was imposed in violation of the Constitution or laws of the United States." Because Dubin is a statutory interpretation case that clarifies the meaning of § 1028A, it applies retroactively to cases that were not yet final when the decision was issued, and to cases on collateral review where the defendant raised a timely objection to the jury instructions or the sufficiency of the evidence. The government will argue that Dubin announced a "new rule" that does not apply retroactively under Teague, but this argument fails for two reasons. First, Teague applies only to new constitutional rules of criminal procedure, not to statutory interpretation cases. Second, the Supreme Court in Dubin explicitly stated that it was interpreting the statute's plain language, not creating a new rule. The Court cited Flores-Figueroa as controlling precedent, and Flores-Figueroa itself has been applied retroactively by numerous circuits.

To succeed on a § 2255 motion based on Dubin, defense counsel must establish two things: first, that the jury instruction or the evidence at trial did not require the government to prove the defendant knew the means of identification belonged to another person; and second, that the error was not harmless beyond a reasonable doubt. The harmless error analysis is the government's strongest argument. If the evidence at trial overwhelmingly showed that the defendant knew the identification belonged to a real person—for example, because the defendant personally knew the victim or because the identification documents included the victim's photograph—then the error in the jury instruction may be harmless. But in the vast majority of identity theft cases, the government's evidence is circumstantial and ambiguous. The defendant may have used a stolen Social Security number without knowing whether it belonged to a living person, a deceased person, or a fabricated number. In those cases, the error cannot be harmless because the jury might have acquitted if properly instructed.

There is also a procedural trap that defense counsel must avoid. Under the Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA), § 2255 motions are subject to a one-year statute of limitations that runs from the date the conviction becomes final. However, the Supreme Court has held that the limitations period may be equitably tolled if the defendant can show extraordinary circumstances and diligent pursuit of rights. More importantly, the one-year clock resets when the Supreme Court recognizes a new right that is retroactively applicable. Because Dubin is a statutory interpretation case, it does not create a "new right" in the constitutional sense, so the AEDPA clock may not reset automatically. Defense counsel must file § 2255 motions within one year of the conviction becoming final, or within one year of the date on which the factual predicate of the claim could have been discovered through due diligence. For defendants whose convictions became final before June 1, 2023—the date Dubin was decided—the motion must be filed no later than June 1, 2024, unless equitable tolling applies. I strongly advise filing immediately to avoid any statute-of-limitations arguments.

Another strategic consideration is the interplay between Dubin and the Supreme Court's recent decision in Davis v. United States, 602 U.S. ___ (2024), which addressed the retroactivity of statutory interpretation decisions in the context of the Armed Career Criminal Act. In Davis, the Court held that a new statutory interpretation that narrows the scope of a federal criminal statute applies retroactively to cases on collateral review, even if the interpretation was not "dictated by precedent" at the time of the conviction. This is a powerful endorsement of the retroactivity argument for Dubin. I recommend citing Davis in every § 2255 motion that raises a Dubin claim, and I also recommend attaching a copy of the Dubin opinion and the relevant portions of the trial transcript that show the erroneous jury instruction. The government will likely oppose the motion on procedural grounds, but the substantive law is squarely on the defendant's side.

The New Plea Bargaining Landscape: Leveraging Dubin to Dismiss the Mandatory Two-Year Sentence

In my experience, the most effective use of Dubin is not in the courtroom but in the prosecutor's office. The decision has fundamentally altered the calculus for federal prosecutors who must decide whether to charge a § 1028A count or whether to dismiss it in exchange for a guilty plea to the predicate felony. Before Dubin, prosecutors could charge aggravated identity theft with near-impunity because the knowledge element was essentially automatic. The defendant's only realistic defense was to argue that the identification did not actually belong to another person—a rare and difficult argument. Now, prosecutors must evaluate whether they have admissible evidence that the defendant knew the identification belonged to a real person. In many fraud cases, that evidence is thin or nonexistent. For example, consider a defendant who used a stolen Social Security number to obtain a credit card. The government can prove the number belonged to a real person, and it can prove the defendant knowingly used the number. But can it prove the defendant knew the number belonged to a real person? If the defendant obtained the number from a website that sells "generated" numbers, or if the defendant believed the number was a test number used by a corporation, the government may lack the evidence to satisfy Dubin.

Defense counsel should approach plea negotiations with a clear, written proffer that outlines the evidentiary gaps in the government's case under Dubin. I recommend drafting a memorandum that identifies each element of the § 1028A charge, lists the evidence the government has for each element, and explains why the evidence of knowledge is insufficient. The memorandum should also cite the Supreme Court's language in Dubin emphasizing that "knowingly" applies to "of another person" and that the government cannot rely on the mere fact that the identification belonged to a real person to prove knowledge. This memorandum serves two purposes: it educates the prosecutor about the new legal landscape, and it creates a record of the defense's good-faith efforts to resolve the case without litigation. In my practice, I have found that prosecutors in the U.S. Attorney's Office for the District of Columbia and the Eastern District of Virginia are particularly receptive to these arguments because they are aware of the circuit split that Dubin resolved and do not want to risk a reversal on appeal.

There is also a significant sentencing benefit to dismissing the § 1028A count. The mandatory two-year consecutive sentence under § 1028A is not subject to the usual sentencing factors under 18 U.S.C. § 3553(a), which means the court cannot reduce it for any reason other than substantial assistance. By dismissing the § 1028A count, the prosecutor gives the sentencing judge full discretion to impose a sentence within the advisory guidelines range for the predicate felony. For a defendant who is facing a guidelines range of 24 to 30 months on the predicate felony, the addition of the mandatory 24-month consecutive sentence under § 1028A would result in a total sentence of 48 to 54 months. Without the § 1028A count, the judge could sentence the defendant to the low end of the guidelines, or even below the guidelines if the defendant qualifies for a downward departure under U.S.S.G. § 5K1.1 or 18 U.S.C. § 3553(e). The difference of 24 to 30 months is life-altering for the defendant and the defendant's family.

Finally, defense counsel should be aware of the potential for a "conditional plea" under Federal Rule of Criminal Procedure 11(a)(2), which allows the defendant to plead guilty to the predicate felony while preserving the right to appeal the denial of a motion to dismiss the § 1028A count. If the prosecutor refuses to dismiss the § 1028A count, defense counsel should file a pretrial motion to dismiss the indictment based on the government's inability to prove the knowledge element under Dubin. The motion should argue that the indictment fails to state an offense under Federal Rule of Criminal Procedure 12(b)(3)(B)(v) because it does not allege that the defendant knew the means of identification belonged to

Related Legal Resources

Related Pages

Kirby Law Network

Explore our full network of federal criminal defense resources: