Retroactive Terrorism Limitations Waiver Before Expiration; Ex Post Facto Limits on § 3A1.4 “Involved” Prong When Predicate Listing Post-Dates Conduct

1. Introduction

United States v. Saab addresses an unusual but consequential temporal problem in terrorism prosecutions: what happens when the charged conduct straddles (or even largely predates) the enactment of the very statute of conviction, and when other terrorism-adjacent provisions (a limitations waiver and a Guidelines enhancement) become applicable only after Congress amends a cross-referenced definitional list.

Parties: The United States prosecuted Alexei Saab (also known as Ali Hassan Saab, Alex Saab, Rachid). Saab, born and raised in Lebanon, was recruited into Hizballah and later lived in the United States while traveling back to Lebanon for training and direction.

Charges and posture: After a 2022 jury trial, Saab was convicted on Count Three for receiving military-type training from Hizballah in violation of 18 U.S.C. § 2339D (and § 3238). The district court applied U.S.S.G. § 3A1.4, sentencing him principally to ten years on Count Three (and a total of twelve years on all counts of conviction). On appeal, Saab challenged only Count Three—both conviction and sentence—under plain error review.

Key issues: (i) Due process limits on conviction where the jury was not instructed on § 2339D’s enactment date; (ii) whether 18 U.S.C. § 3286(b) (the “Limitations Waiver”) could be applied retroactively to Saab’s § 2339D offense; (iii) whether the jury was properly instructed to find the § 3286(b) “foreseeable risk” element (and whether evidence was sufficient); and (iv) whether applying U.S.S.G. § 3A1.4 violated the Ex Post Facto Clause given the timing of when § 2339D was added to § 2332b(g)(5)(B).

2. Summary of the Opinion

• Conviction affirmed: The panel held that the district court plainly erred by not instructing the jury that it could convict only for post-enactment conduct (after Dec. 17, 2004), but the error did not affect substantial rights because the trial record contained substantial evidence Saab received qualifying training in 2005.
• Limitations waiver upheld: The panel held that applying § 3286(b) to Saab’s § 2339D prosecution was permissible because the original limitations period had not run; retroactive extension before expiration does not violate the Ex Post Facto Clause and is consistent with retroactivity principles under Landgraf v. USI Film Products.
• Foreseeable-risk instruction error but harmless: The panel agreed the jury should have been instructed that the “foreseeable risk of death or serious bodily injury” must be created by Saab’s post-enactment conduct, but found no reasonable probability of acquittal given the evidence.
• Sentence vacated and remanded: The panel held the district court plainly erred in applying U.S.S.G. § 3A1.4 on the theory the § 2339D offense “involved” a federal crime of terrorism, because § 2339D was not in § 2332b(g)(5)(B) until March 9, 2006—after Saab’s conduct ended. The Government’s alternative “intended to promote” theory was not sufficiently developed in the record to deem the error harmless, so resentencing was required.
• Dissent on sentencing: Judge Sullivan would have affirmed the sentence, concluding Saab’s conduct “involved” and was “intended to promote” other then-enumerated terrorism crimes, making remand unnecessary.

3. Analysis

3.1. Precedents Cited and Their Influence

A. Plain error and pre-enactment conduct: the Marcus framework

The court’s due-process analysis is explicitly anchored in the Supreme Court’s and Second Circuit’s Marcus trilogy. The panel cites United States v. Marcus, 560 U.S. 258 (2010) ("Marcus II") for the basic proposition that a conviction based exclusively on noncriminal, pre-enactment conduct would violate due process if the jury was not instructed on the enactment date. It then applies the operational Second Circuit standard from United States v. Marcus, 628 F.3d 36 (2d Cir. 2010) ("Marcus III"): whether there is a “reasonable probability” the jury would have acquitted absent the error.

The panel also distinguishes Saab’s attempted reliance on the concurrence from United States v. Marcus, 538 F.3d 97 (2d Cir. 2008) ("Marcus I"), noting the concurrence’s lower “reasonable possibility” phrasing was not adopted in Marcus III. This clarification matters as a doctrinal signal: Saab reinforces that the Second Circuit’s plain-error prejudice threshold in this setting is the “reasonable probability of acquittal” test, not a more defendant-friendly formulation.

The panel’s citation to United States v. Hild, 147 F.4th 103 (2d Cir. 2025) underscores that the “no reasonable probability of acquittal” approach is not unique to Marcus facts and reflects a broader circuit practice in assessing prejudice under plain error.

B. Ex Post Facto principles and retroactivity

On the Ex Post Facto Clause, the panel restates classic categories via United States v. Harris, 79 F.3d 223 (2d Cir. 1996) (citing Collins v. Youngblood, 497 U.S. 37 (1990)). It uses Garner v. Jones, 529 U.S. 244 (2000) and Miller v. Florida, 482 U.S. 423 (1987) to emphasize the “significant risk” / “disadvantage” inquiry. But the decisive precedent for limitations retroactivity is the “unexpired limitations period” rule, described as “long-standing” in the Second Circuit and reaffirmed in United States v. Weinlein, 109 F.4th 91 (2d Cir. 2024), with Supreme Court support from Stogner v. California, 539 U.S. 607 (2003).

For statutory retroactivity more generally, the court deploys the two-step framework of Landgraf v. USI Film Products, 511 U.S. 244 (1994), filtered through Second Circuit applications such as Weingarten v. United States, 865 F.3d 48 (2d Cir. 2017) and the Supreme Court’s Fernandez-Vargas v. Gonzales, 548 U.S. 30 (2006). The presumption against retroactivity is reiterated with Olivieri v. Stifel, Nicolaus & Co., 112 F.4th 74 (2d Cir. 2024).

C. Statutes of limitations as procedural under Landgraf step two

The panel’s conclusion that applying the amended coverage of § 3286(b) does not create impermissible retroactive effects draws on civil and criminal analogies. It cites Vernon v. Cassadaga Valley Cent. Sch. Dist., 49 F.3d 886 (2d Cir. 1995) for the proposition that changed limitations rules are typically procedural and may apply to later-filed actions without upsetting reliance interests.

To buttress this in the criminal context, the panel cites out-of-circuit decisions that treat limitations changes similarly under Landgraf, including United States v. Norwood, 49 F.4th 189 (3d Cir. 2022), United States v. Carlson, 235 F.3d 466 (9th Cir. 2000), United States v. Piette, 45 F.4th 1142 (10th Cir. 2022), and a district court decision affirmed in-circuit, United States v. Maxwell, 534 F. Supp. 3d 299 (S.D.N.Y. 2021), aff’d, 118 F.4th 256 (2d Cir. 2024).

D. Waiver versus forfeiture of jury instruction claims

On whether Saab “waived” his jury instruction argument by proposing similar language, the panel rejects a bright-line approach and instead adheres to the “intentional relinquishment” concept. It invokes United States v. Crowley, 318 F.3d 401 (2d Cir. 2003) (no waiver where conduct reflects acquiescence rather than invitation), distinguishes United States v. Giovanelli, 464 F.3d 346 (2d Cir. 2006) (waiver where omission was requested and approved), and acknowledges cases like United States v. Caltabiano, 871 F.3d 210 (2d Cir. 2017) while stressing context.

The court’s reasoning is reinforced by persuasive authorities outside the circuit—United States v. Perez, 116 F.3d 840 (9th Cir. 1997) (en banc) and Gov't of Virgin Islands v. Rosa, 399 F.3d 283 (3d Cir. 2005)—to support the principle that one cannot intentionally waive a right one did not know existed, particularly where the controlling enactment-date problem was apparently unknown to all participants at trial.

E. Plain error based on statutory text

The panel’s determination that the erroneous foreseeable-risk instruction was “plain” even without binding precedent relies on United States v. Polouizzi, 564 F.3d 142 (2d Cir. 2009): plainness can be established by the “plain language of the statute,” here § 3286(b) requiring the risk be created by “the commission of such offense,” which—given the due-process timing limit—must be post-enactment conduct.

F. Foreseeability content

To define foreseeability, the panel analogizes to criminal and civil formulations: United States v. Turk, 626 F.3d 743 (2d Cir. 2010) (harm the defendant knew or reasonably should have known was a potential result), and Sec. & Exch. Comm'n v. Rashid, 96 F.4th 233 (2d Cir. 2024) (reasonable prudence anticipation), while noting probative value of prior similar experience via Gardner v. Q. H. S., Inc., 448 F.2d 238 (4th Cir. 1971).

G. Sentencing: ex post facto and harmlessness standards

For ex post facto in the Guidelines setting, the panel relies on United States v. Ortiz, 621 F.3d 82 (2d Cir. 2010). On harmlessness and the “would have imposed the same sentence” inquiry, it cites United States v. Arigbodi, 924 F.2d 462 (2d Cir. 1991) and United States v. Jass, 569 F.3d 47 (2d Cir. 2009), and emphasizes the relatively relaxed approach to noticing sentencing errors under United States v. Gamez, 577 F.3d 394 (2d Cir. 2009), United States v. Williams, 399 F.3d 450 (2d Cir. 2005), and United States v. Wernick, 691 F.3d 108 (2d Cir. 2012).

H. U.S.S.G. § 3A1.4’s meaning: “involved” versus “intended to promote”

Although the majority does not decide the alternative “intended to promote” theory on the merits, it sets the legal frame by citing United States v. Awan, 607 F.3d 306 (2d Cir. 2010) for the definition of “intended to promote,” and it treats remand as necessary because the district court did not address the alternative predicate.

The dissent’s analysis further illustrates the interpretive stakes: it would apply § 3A1.4 because Saab’s conduct “involved” and “intended to promote” other then-enumerated federal crimes of terrorism, citing United States v. Mingo, 340 F.3d 112 (2d Cir. 2003) and United States v. Stewart, 590 F.3d 93 (2d Cir. 2009) (plain language controls), and invoking the appellate principle from United States v. Barker, 723 F.3d 315 (2d Cir. 2013) that the court may affirm on any supported ground. The dissent also notes the Government’s legal concession is not binding, citing Orloff v. Willoughby, 345 U.S. 83 (1953) and United States v. Smith, 621 F.2d 483 (2d Cir. 1980).

3.2. The Court’s Legal Reasoning

A. Due process: enactment-date limitation as an element of lawful conviction, filtered through plain error

The court treats the enactment-date issue as a due-process constraint on what conduct may serve as the basis for conviction. Because 18 U.S.C. § 2339D was enacted on December 17, 2004, the jury should have been instructed to convict only if Saab received military-type training after that date. The court accepts (and the Government concedes) that omission was plain error.

But the decision’s real doctrinal work lies in the prejudice analysis: under Marcus III, the question is not whether pre-enactment evidence predominated, but whether there is a reasonable probability the jury would have acquitted had it been properly instructed. The panel then methodically catalogues post-enactment proof—explosives and surveillance training in 2005, corroborated by admissions, residue testing, passport/visa stamps, and physical evidence—finding the probability of acquittal remote.

The court also stresses the “material difference” inquiry: if pre- and post-enactment conduct differ materially (as they did in the vacated Marcus sex-trafficking count), an incorrect charge may matter. Here, by contrast, the court finds continuity: training in weapons/explosives/surveillance both before and after enactment, and post-enactment conduct that is “quintessential” military-type training within § 2339D(c)(1).

B. Retroactivity of the limitations waiver: substantive versus definitional change and the “unexpired limitations” rule

A central analytic move is the court’s refusal to treat the March 2006 amendment to § 2332b(g)(5)(B) as the relevant “retroactive statute” in the abstract. The panel characterizes § 2332b(g)(5) as definitional—its “substantive effects” arise only through other provisions that cross-reference it, including § 3286(b) and U.S.S.G. § 3A1.4. Accordingly, the retroactivity question is directed at the substantive mechanism: does applying § 3286(b) to a § 2339D prosecution for 2005 conduct violate Ex Post Facto or Landgraf?

The Ex Post Facto analysis is straightforward: extending a limitations period is permissible if done before the original period expires. Because Saab’s 2005 conduct was not yet time-barred when Congress expanded the cross-referenced list in March 2006, the amendment did not resurrect a dead prosecution (which Stogner v. California forbids) but merely extended an unexpired period (which Stogner permits).

Under Landgraf v. USI Film Products, even assuming textual ambiguity, the panel treats the limitations change as procedural: it regulates “secondary rather than primary conduct” and diminishes reliance interests. The opinion’s civil analogy (Vernon v. Cassadaga Valley Cent. Sch. Dist.) is used to support the same conclusion in criminal cases: extending an unexpired limitations period does not attach new legal consequences to completed primary conduct in the way retroactivity doctrine primarily guards against.

C. The “foreseeable risk” predicate must be tied to post-enactment conduct

The panel adopts a clean textual logic: § 3286(b) requires that “the commission of such offense” created a foreseeable risk. If the offense can only be “committed” (for due-process purposes) after the statute’s enactment, then the risk finding must also be anchored in post-enactment conduct. The jury charge failed to convey that temporal constraint, which the court deems plain error under United States v. Polouizzi because the statutory text makes the requirement obvious.

The harmlessness analysis is fact-intensive: explosives training (including building/testing IEDs) plus surveillance aimed at facilitating destruction of targets, in the context of Hizballah’s known activities and Saab’s prior operational experience, makes death or serious bodily injury a foreseeable potential consequence. The panel thus finds no reasonable probability that a properly instructed jury would have rejected foreseeability.

D. Sentencing: ex post facto constraints on § 3A1.4 and institutional competence on alternative theories

The sentencing component is where Saab creates the most consequential “new law” effect. The district court applied § 3A1.4 because it found the § 2339D offense “involved” a federal crime of terrorism. The Government conceded error on appeal: since § 2339D was not listed in § 2332b(g)(5)(B) until March 9, 2006, treating a 2005 § 2339D offense as “involving” a federal crime of terrorism on that basis created an ex post facto risk of a harsher sentence under United States v. Ortiz.

The Government attempted to salvage the sentence by shifting to an “intended to promote” rationale (citing other enumerated terrorism crimes). The majority’s key reasoning is institutional and procedural: this alternative theory was not presented below; the record does not “clearly indicate” the district court would apply the enhancement on that basis; and the enhancement has a “dramatic impact” on the Guidelines range (invoking United States v. Wernick). Therefore, resentencing is the proper corrective mechanism rather than appellate fact-finding or speculation.

Judge Sullivan’s partial dissent frames the counterpoint: § 3A1.4’s “involved” and “intended to promote” language is broad (as interpreted in United States v. Awan), and the district court’s adoption of PSR facts could support the enhancement via other then-enumerated terrorism offenses, making remand unnecessary. The majority, however, treats the absence of developed adversarial presentation on these alternative routes as a reason to remand rather than decide the enhancement’s applicability in the first instance.

3.3. Impact

A. Trial practice: enactment-date instructions are now a highlighted fault line

Saab makes enactment-date awareness a non-negotiable feature of prosecutions under statutes enacted mid-course of alleged criminal conduct. While the panel affirms on prejudice, it confirms that failure to instruct on the enactment date is plain error. Practically, district courts and litigants should expect:

• more frequent enactment-date limiting instructions in cases charging long time windows, and
• more refined verdict forms or special interrogatories when a charge spans pre/post enactment periods.

B. Limitations waivers: extending coverage via cross-referenced definitional lists is treated as procedural (if not expired)

The opinion strengthens the Government’s ability to rely on later-expanded terrorism limitations mechanisms—at least where the original period had not expired. Defendants will face an uphill battle arguing that a cross-reference amendment that expands a limitations waiver is impermissibly retroactive, unless they can show the prosecution was already time-barred (a Stogner-style problem) or that Congress clearly imposed a substantive change beyond limitations procedure.

C. Jury instructions under § 3286(b): the “foreseeable risk” finding must match the legally chargeable time slice

Even though Saab finds harmless error, it sets a clear expectation: when a limitations waiver’s application turns on a risk element, the jury should be asked to find that risk for the legally cognizable period of the offense (here, post-enactment). This is likely to affect how terrorism prosecutors draft instructions when charging spans periods before a statute’s applicability or before cross-reference amendments.

D. Sentencing: ex post facto scrutiny for Guidelines cross-references is sharpened; remand likely where alternative theories were not litigated

Saab’s sentencing holding has two forward-looking effects:

• Ex post facto guardrails: District courts must verify not merely the defendant’s conduct date, but also the date the offense became enumerated under § 2332b(g)(5)(B) when using § 3A1.4’s “involved” theory tied to the offense of conviction itself.
• Appellate reluctance to supply new enhancement predicates: When the Government pivots on appeal to alternative predicates (“intended to promote” or other enumerated crimes), Saab signals that the Second Circuit may prefer remand unless the record clearly demonstrates the district court would have imposed the same enhancement/sentence on that basis.

The dissent simultaneously signals litigation risk: on remand, the Government may press (and the district court may accept) a broader Awan-style “involved” or “intended to promote” theory based on other enumerated crimes, potentially reaching the same sentence. Saab thus may shift the battleground from whether § 3A1.4 can ever apply to whether the record supports alternate enumerated predicates and the required “calculated to influence” intent.

4. Complex Concepts Simplified

• Plain error review: When the defendant did not object at trial, an appellate court corrects an error only if it is (1) an error, (2) clear/obvious, (3) likely changed the outcome, and (4) seriously harms fairness or integrity. Saab turns mostly on step (3): whether the error probably changed the verdict or sentence.
• Due process and enactment dates: You cannot be convicted for conduct that was not a crime yet. If a statute is enacted on a certain date, the jury must effectively be told (explicitly or functionally) to base guilt on conduct after that date.
• Ex Post Facto Clause (limitations): Government cannot revive a prosecution after the limitations period already expired (that would be like re-criminalizing by time). But it can extend the limitations period for crimes not yet time-barred.
• Landgraf retroactivity test: Courts ask (1) did Congress clearly say it applies retroactively? If not, (2) would applying it to past events unfairly attach new legal consequences? Procedural changes—like filing deadlines—often apply to pending/future filings without being deemed impermissibly retroactive.
• “Foreseeable risk of death or serious bodily injury” in § 3286(b): Not whether someone actually died, but whether the offense created a reasonably anticipatable potential for death/serious injury.
• U.S.S.G. § 3A1.4 (“Terrorism Enhancement”): Adds 12 offense levels and assigns criminal history category VI if the felony “involved” or was “intended to promote” a “federal crime of terrorism,” a term defined by a statutory list plus an intent requirement (“calculated to influence government”). Saab holds that using an after-the-fact listing to justify the enhancement can raise ex post facto problems; whether other predicates can support it is left for remand.

5. Conclusion

United States v. Saab is a temporal-precision decision with significant terrorism-prosecution consequences. It (1) reaffirms that enactment-date limiting instructions are constitutionally required but may be harmless under Marcus III when post-enactment evidence is substantial; (2) holds that the § 3286(b) terrorism limitations waiver can apply retroactively when the original limitations period had not expired, consistent with Stogner v. California and Landgraf v. USI Film Products; (3) clarifies that the § 3286(b) “foreseeable risk” finding must be tied to the legally chargeable (post-enactment) portion of the offense; and (4) vacates a terrorism-enhanced sentence where the enhancement depended on a terrorism-listing amendment that post-dated the offense conduct, remanding to permit the district court to address alternative enhancement theories in the first instance.

The broader significance lies in Saab’s insistence on “date-aware” adjudication: not only must courts track when conduct occurred, but also when cross-referenced definitional lists became operative for limitations and sentencing—an increasingly important discipline in long-horizon national security cases.