AEDPA “Atkins-Only” Review of Intellectual-Disability Claims: Hall/Moore Non-Retroactivity and Deference to State Findings (Franqui)

1. Introduction

In Leonardo Franqui v. Secretary, Florida Department of Corrections (11th Cir. Aug. 10, 2026) (per curiam) (not for publication), the Eleventh Circuit affirmed the denial of federal habeas relief under 28 U.S.C. § 2254 on Leonardo Franqui’s claim that the Eighth Amendment bars his execution because he is intellectually disabled.

Franqui was under two Florida death sentences arising from two separate murders: the “Hialeah case” (Franqui v. State, 699 So. 2d 1312 (Fla. 1997)) and the “North Miami case” (Franqui v. State, 804 So. 2d 1185 (Fla. 2001)). Although his federal habeas actions were separate, Florida consolidated state postconviction proceedings after Hall v. Florida, 572 U.S. 701 (2014) and held a single evidentiary hearing on whether Franqui was intellectually disabled.

The central issues on appeal were narrowly defined by a certificate of appealability (COA): (1) whether the Florida Supreme Court’s decision in Franqui v. State, 301 So. 3d 152 (Fla. 2020) (“Franqui III”) was “contrary to, or an unreasonable application of,” Atkins v. Virginia, 536 U.S. 304 (2002); and (2) whether it rested on an unreasonable determination of the facts. Because the COA excluded Hall and Moore v. Texas, 581 U.S. 1 (2017) as non-retroactive on federal collateral review, the Eleventh Circuit’s analysis was explicitly “Atkins-only” under AEDPA.

2. Summary of the Opinion

The Eleventh Circuit held:

  • No AEDPA legal error under § 2254(d)(1): Franqui III was not contrary to, nor an unreasonable application of, Atkins v. Virginia, because Atkins set a constitutional floor and left to the States the task of implementing procedures and standards for determining intellectual disability.
  • No AEDPA factual unreasonableness under § 2254(d)(2): The state courts’ findings that Franqui failed to prove (i) significantly subaverage intellectual functioning, (ii) significant adaptive deficits, and (iii) onset before age eighteen were not objectively unreasonable in light of the record (including disputed expert testimony, evidence of malingering, and evidence of day-to-day functioning).

Accordingly, the court affirmed the district court’s denial of Franqui’s habeas petition.

3. Analysis

3.1. Precedents Cited

A. Substantive Eighth Amendment framework

  • Atkins v. Virginia, 536 U.S. 304 (2002)
    Role in the opinion: The governing “clearly established Federal law” for the COA issues. The Eleventh Circuit emphasized that Atkins prohibits executing an intellectually disabled person but “le[ft] to the State[s] the task of developing appropriate ways to enforce the constitutional restriction,” making Atkins a constitutional floor rather than a detailed diagnostic mandate. This framing supported deference to Florida’s three-prong clinical model (intellectual functioning, adaptive deficits, onset before 18) and to Florida’s factfinding.
  • Ford v. Wainwright, 477 U.S. 399 (1986)
    Role in the opinion: Quoted via Atkins to reinforce that the Constitution can set a substantive limit while leaving procedural implementation to the States—an important move for rejecting arguments that Florida’s appellate review standard or level of explanation violated federal law.

B. Hall/Moore retroactivity and the COA’s narrowing function

  • Hall v. Florida, 572 U.S. 701 (2014)
    Role in the opinion: Contextual, not governing for § 2254(d)(1) because the COA excluded Hall as non-retroactive on federal collateral review. The court still used Hall to explain Franqui’s arguments (SEM and holistic evaluation) and then concluded that, even if considered, the state courts did not engage in the kind of rigid IQ cutoff approach condemned by Hall.
  • Moore v. Texas, 581 U.S. 1 (2017)
    Role in the opinion: Also contextual but excluded from the governing federal law by the COA’s retroactivity limitation. The Eleventh Circuit noted Florida considered and rejected a Moore argument on the merits, but federal habeas review could not treat Moore as controlling “clearly established” law for this case.
  • In re Henry, 757 F.3d 1151 (11th Cir. 2014) and Smith v. Comm'r, Ala. Dep't of Corr., 924 F.3d 1330 (11th Cir. 2019)
    Role in the opinion: These cases supplied the Eleventh Circuit’s binding circuit law that Hall and Moore do not apply retroactively on collateral review, justifying the COA’s confinement to Atkins.
  • Castillo v. United States, 816 F.3d 1300 (11th Cir. 2016)
    Role in the opinion: Invoked for the principle that appellate review is limited to issues specified in the COA—critical to the court’s refusal to expand the merits inquiry to Hall/Moore.
  • Hamm v. Smith, No. 24-872, slip op. at 1 (U.S. May 21, 2026) (per curiam)
    Role in the opinion: The court treated the Supreme Court’s dismissal as improvidently granted as providing no merits holding and therefore no new “clearly established Federal law” for § 2254(d)(1).

C. AEDPA deference and standards of review

  • Harrington v. Richter, 562 U.S. 86 (2011)
    Role in the opinion: Supplied the “fairminded disagreement” benchmark—habeas relief is unavailable unless the state decision is so unjustified that no fairminded jurist could agree.
  • Williams v. Taylor, 529 U.S. 362 (2000)
    Role in the opinion: Defined “contrary to” and “unreasonable application” under § 2254(d)(1).
  • Early v. Packer, 537 U.S. 3 (2002)
    Role in the opinion: Used to reject the argument that the Florida Supreme Court’s failure to cite Atkins itself showed legal error; a state court need not cite Supreme Court cases if its reasoning/result do not contradict them.
  • Wood v. Allen, 558 U.S. 290 (2010), Miller-El v. Cockrell, 537 U.S. 322 (2003), and Rice v. Collins, 546 U.S. 333 (2006)
    Role in the opinion: Anchored the court’s refusal to reweigh evidence and reinforced the deference owed to state factfinding under § 2254(d)(2) and the presumption of correctness under § 2254(e)(1).
  • Morrow v. Warden, 886 F.3d 1138 (11th Cir. 2018), Pye v. Warden, Ga. Diagnostic Prison, 50 F.4th 1025 (11th Cir. 2022) (en banc), Sealey v. Warden, Ga. Diagnostic Prison, 954 F.3d 1338 (11th Cir. 2020), and Evans v. Sec'y, Dep't of Corr., 703 F.3d 1316 (11th Cir. 2013)
    Role in the opinion: These decisions supplied the Eleventh Circuit’s articulation of AEDPA’s highly deferential framework and the definitions of § 2254(d)(1)/(d)(2) in circuit practice.

D. Intellectual-disability evidentiary disputes: SEM, Flynn Effect, malingering

  • Raulerson v. Warden, 928 F.3d 987 (11th Cir. 2019)
    Role in the opinion: Cited to explain the “Flynn Effect” and to underscore that Hall addressed SEM, not a judicial mandate to apply Flynn adjustments. This supported the conclusion that declining a Flynn adjustment was not unreasonable.
  • Quince v. State, 241 So. 3d 58 (Fla. 2018)
    Role in the opinion: Used by the state postconviction court (and noted in the federal opinion’s recitation) to justify declining to adjust IQ scores for the Flynn Effect where substantial evidence supported that choice. It functioned as Florida authority legitimizing the state court’s evidentiary resolution.
  • United States v. Stein, 964 F.3d 1313 (11th Cir. 2020)
    Role in the opinion: Used to normalize trial-level choices between “dueling experts.” The Eleventh Circuit relied on this concept to hold that choosing among competing testimony about SEM and Flynn adjustments was not an unreasonable fact determination.

E. Additional deference cases on conflicting evidence

  • Eggers v. Alabama, 876 F.3d 1086 (11th Cir. 2017) and Holsey v. Warden, Ga. Diagnostic Prison, 694 F.3d 1230 (11th Cir. 2012)
    Role in the opinion: Reinforced that conflicting record evidence does not itself establish unreasonableness under § 2254(d)(2); if fairminded jurists could agree with the state court, federal habeas relief is barred.

3.2. Legal Reasoning

A. The court’s threshold move: narrowing the governing law to Atkins

A defining feature of the decision is its insistence on the COA’s limits and AEDPA’s “clearly established” requirement. The Eleventh Circuit treated Atkins as the sole Supreme Court decision controlling the § 2254(d)(1) inquiry, because circuit precedent deems Hall and Moore non-retroactive on federal collateral review. This eliminated a common path for capital habeas petitioners—recasting an older Atkins claim as a Hall/Moore claim—and forced the court to ask only whether Florida’s approach fell below the constitutional floor described in Atkins.

B. Atkins as a “floor,” not a detailed method—and why that matters under AEDPA

The opinion reads Atkins as delegating substantial implementation authority to the States, so long as state standards remain clinically grounded and recognize the core elements (intellectual functioning, adaptive deficits, childhood onset). From that premise, the Eleventh Circuit rejected:

  • The “wrong law” argument based on Florida’s discussion of Hall/Moore; state courts may consider later-decided cases without converting the federal habeas “clearly established” inquiry.
  • The “no Atkins citation” argument; under Early v. Packer, citations are unnecessary if the outcome is not inconsistent with Supreme Court precedent.
  • The “wrong appellate standard” argument (Florida’s “competent, substantial evidence” review); Atkins does not prescribe state appellate standards or require de novo reweighing.

C. Application of § 2254(d)(2): deference to state factfinding on contested expert questions

The decision’s fact analysis is fundamentally an AEDPA exercise: the question is not whether Franqui might plausibly be intellectually disabled, but whether the state courts were objectively unreasonable in rejecting his proof. The Eleventh Circuit emphasized that the record contained:

  • multiple IQ scores on approved tests above 70,
  • expert disagreement on SEM calculations and the Flynn Effect’s validity/applicability,
  • validity testing interpreted as malingering (especially in Dr. Suarez’s evaluation), and
  • substantial evidence of functional behavior pre-incarceration and in prison relevant to adaptive functioning.

Given that evidentiary landscape, the Eleventh Circuit treated the state court’s choice among expert views as the paradigm of deference: so long as the choice is supported by record evidence, a federal court cannot substitute its judgment.

3.3. Impact

Practical rule reinforced (as persuasive authority; the decision is unpublished): In Eleventh Circuit federal habeas proceedings where Hall and Moore are unavailable due to non-retroactivity and COA limits, an intellectual-disability claim is evaluated under Atkins plus AEDPA deference; state refusals to apply the Flynn Effect and state credibility choices among experts will rarely satisfy § 2254(d)(2).

Potential consequences include:

  • Strategic emphasis on state-court records: Because federal review is tightly constrained, petitioners must build the most complete evidentiary record in state postconviction proceedings, anticipating that federal courts will not reweigh expert disputes.
  • Flynn Effect remains optional, not mandatory: The decision underscores that a state court may decline Flynn adjustments where the record supports skepticism or lack of consensus, and that such refusal is unlikely to be deemed unreasonable under AEDPA.
  • Malingering/validity testing can be dispositive in practice: Even when SEM arguments might bring a score closer to the disability range, evidence interpreted as poor effort can support findings that obtained IQ results understate true ability.
  • Procedural compartmentalization matters: Separate state judgments can lead to consolidated state disability litigation, but federal habeas relief remains tethered to the specific judgment under review (here, the Hialeah case).

4. Complex Concepts Simplified

  • AEDPA (§ 2254(d)): A federal habeas court can overturn a state decision only if it is (i) contrary to or an unreasonable application of Supreme Court holdings, or (ii) based on an unreasonable view of the facts. It is not enough that the federal court would have decided differently.
  • Certificate of appealability (COA): A gatekeeping order identifying which issues may be appealed. Here, it limited the governing law to Atkins, excluding Hall and Moore.
  • Intellectual disability (clinical triad used by Florida): (1) significantly subaverage intellectual functioning, (2) significant adaptive deficits, and (3) onset before age 18.
  • Standard error of measurement (SEM): IQ scores are estimates; SEM expresses a margin of error. A score should be treated as a range rather than a single fixed number.
  • Flynn Effect: The observed tendency for population IQ scores to rise over time, which can make older test norms produce inflated scores. Courts and experts disagree about whether and how to apply this to an individual.
  • Adaptive behavior domains: (i) conceptual (language, academics, reasoning), (ii) social (interpersonal skills, social judgment), and (iii) practical (self-care, work, daily living).
  • Dismissed as improvidently granted (DIG): When the Supreme Court dismisses a granted case without deciding the merits; it creates no new substantive rule for AEDPA purposes.

5. Conclusion

This Eleventh Circuit decision is a study in habeas constraint: by treating Atkins v. Virginia as the only controlling Supreme Court authority available under the COA and AEDPA, the court framed the case as a deference inquiry rather than a fresh intellectual-disability determination. Within that framework, the court held that Florida’s rejection of Franqui’s claim—grounded in expert disputes over IQ interpretation (including SEM and Flynn Effect), evidence of malingering, and evidence of adaptive functioning—was not beyond the bounds of fairminded disagreement.

The broader significance lies less in any new diagnostic rule than in the procedural message: absent retroactively applicable Supreme Court precedent expanding Atkins, federal habeas courts in the Eleventh Circuit will seldom disturb state intellectual-disability findings where the state court held an evidentiary hearing, credited one set of experts over another, and made domain-based adaptive findings supported by record evidence.