Post–Loper Bright Enterprises v. Raimondo Reaffirmation of AEDPA § 2254(d) Constitutionality and Deferential Review of State Insanity-Instruction and Strickland v. Washington Rulings

Case: David Lee Sanders v. Laura Plappert, Warden (with United States of America as Intervenor)  |  Court: Sixth Circuit  |  Decided: March 3, 2026

1. Introduction

This capital habeas case arises from David Lee Sanders’s 1987 robbery-murders of two men at a Kentucky convenience store. Sanders admitted the shootings but asserted a single defense at guilt: not guilty by reason of insanity. The jury rejected that defense and imposed death.

The case’s procedural posture is unusually long and layered: four Kentucky Supreme Court decisions (Sanders I–IV), three U.S. Supreme Court certiorari denials, and federal habeas proceedings culminating in a Sixth Circuit merits decision. On appeal, Sanders pressed two overarching themes:

• Structural challenge: the constitutionality of AEDPA’s deference regime, 28 U.S.C. § 2254(d), especially after the Supreme Court’s elimination of Chevron deference in Loper Bright Enterprises v. Raimondo.
• Case-specific constitutional claims: (i) ambiguous insanity-defense jury instructions allegedly confusing the burden of proof; (ii) multiple ineffective-assistance claims under Strickland v. Washington; and (iii) cumulative error.

The Sixth Circuit’s published opinion is noteworthy not because it alters the substantive standards of insanity or ineffective assistance, but because it squarely reaffirms AEDPA’s constitutionality and clarifies why Loper Bright Enterprises v. Raimondo does not destabilize § 2254(d).

2. Summary of the Opinion

On the merits, the court concluded:

• AEDPA constitutionality: Congress may define the scope of collateral relief and cabin habeas remedies; § 2254(d) sets a constitutionally valid “rule of decision” limiting when federal courts may grant relief.
• Insanity instruction: Kentucky’s “believe from the evidence” phrasing for insanity defenses was not unconstitutionally ambiguous under AEDPA’s deferential standard.
• Ineffective assistance: Under “doubly deferential” review (Knowles v. Mirzayance), the Kentucky Supreme Court reasonably rejected key Strickland claims; other claims failed de novo for lack of prejudice.
• Cumulative error: procedurally defaulted because Kentucky rejected it on a state procedural bar in Sanders III; no cause shown under Murray v. Carrier.

Judge Stranch dissented at length, arguing AEDPA is unconstitutional as applied because it effectively forces federal courts to give effect to unconstitutional state-court decisions, and also contending Sanders was entitled to relief on the merits (especially the insanity instructions and expert-preparation issues).

3. Analysis

3.1. Precedents Cited

The opinion’s reasoning is built from three clusters of authority: (A) AEDPA’s statutory framework and Supreme Court gloss; (B) separation-of-powers and “rule of decision” cases; and (C) underlying constitutional standards for jury instructions, insanity, and ineffective assistance.

A. AEDPA framework and Supreme Court gloss

• Williams v. Taylor (529 U.S. 362 (2000)) supplies the canonical interpretation of § 2254(d)(1)’s “contrary to” and “unreasonable application” clauses and anchors the court’s insistence that habeas relief is limited to clearly established Supreme Court holdings.
• Harrington v. Richter (562 U.S. 86 (2011)) is used for the “fairminded disagreement” and “extreme malfunctions” language—key to the court’s conclusion that AEDPA’s demanding standard is compatible with federalism and finality.
• White v. Woodall (572 U.S. 415 (2014)), Lockyer v. Andrade (538 U.S. 63 (2003)), Price v. Vincent (538 U.S. 634 (2003)), Brumfield v. Cain (576 U.S. 305 (2015)), and Davis v. Ayala (576 U.S. 257 (2015)) reinforce the “objectively unreasonable” threshold and the deference to state factfinding.
• Brown v. Davenport (596 U.S. 118 (2022)) is pivotal for the proposition that AEDPA supplies a “constitutionally valid rule of decision,” a formulation the Sixth Circuit treats as strongly confirmatory of AEDPA’s constitutionality.
• Shinn v. Ramirez (596 U.S. 366 (2022)) and 28 U.S.C. § 2254(e)(2) support the court’s refusal to expand the record with new mitigation affidavits where the relevant claim was adjudicated on the merits in state court.

B. Separation-of-powers and “rule of decision” cases

• Marbury v. Madison (1 Cranch 137 (1803)) is the rhetorical and doctrinal touchstone for Sanders’s argument that AEDPA impermissibly interferes with Article III’s duty to “say what the law is.” The court distinguishes “independent interpretation” from “availability of habeas relief.”
• Felker v. Turpin (518 U.S. 651 (1996)) is invoked for the proposition that Congress may impose “new conditions” on habeas relief and that judgments about the writ’s proper scope are “normally for Congress to make.”
• Boumediene v. Bush (553 U.S. 723 (2008)) appears as an important caveat: congressional control over habeas is significant but not plenary; there are constitutional limits.
• United States v. Klein (80 U.S. 128 (1871)) is raised by Sanders as a “rules of decision” constraint; the court uses Robertson v. Seattle Audubon Society (503 U.S. 429 (1992)) and Bank Markazi v. Peterson (578 U.S. 212 (2016)) to narrow Klein and uphold Congress’s power to impose new legal standards that constrain outcomes.
• Lauf v. E.G. Shinner & Co. (303 U.S. 323 (1938)) is used to illustrate Congress’s authority to limit remedies (there, injunctions) by defining inferior federal courts’ jurisdiction.
• Federalism/structure references such as The Federalist No. 82 and Tafflin v. Levitt (493 U.S. 455 (1990)) support the premise that state courts are “presumptively competent” to adjudicate federal rights, undercutting the notion that Article III requires universal de novo federal re-litigation via habeas.

C. Underlying merits standards: jury instructions, insanity, and ineffective assistance

• Boyde v. California (494 U.S. 370 (1990)) supplies the due process test for ambiguous instructions: whether there is a “reasonable likelihood” the jury applied the instruction in a constitutionally impermissible way.
• Leland v. Oregon (343 U.S. 790 (1952)) and Engle v. Isaac (456 U.S. 107 (1982)) support the baseline rule that states have latitude to allocate the insanity burden of proof.
• Hicks v. Oklahoma (447 U.S. 343 (1980)) is cited for the narrower proposition that once a state sets the burden, due process protects a defendant’s interest in having that burden properly applied.
• Strickland v. Washington (466 U.S. 668 (1984)) governs the ineffective-assistance claims, with “doubly deferential” layering via Knowles v. Mirzayance (556 U.S. 111 (2009)) and emphasized by the Sixth Circuit’s own en banc decision Fields v. Jordan (86 F.4th 218 (6th Cir. 2023) (en banc)).
• Ake v. Oklahoma (470 U.S. 68 (1985)) is used to frame the “mental-health expert” component; the court stresses counsel discretion and resource-balancing, echoing Harrington v. Richter and Dunn v. Reeves.
• Competency standards derive from Drope v. Missouri (420 U.S. 162 (1975)) and Godinez v. Moran (509 U.S. 389 (1993)), applied through a Strickland prejudice lens.

3.2. Legal Reasoning

A. The new clarification: why Loper Bright Enterprises v. Raimondo does not destabilize AEDPA

Sanders attempted a novel bridge: if Loper Bright Enterprises v. Raimondo condemns Chevron-style deference as inconsistent with Article III (at least per concurrences), then “AEDPA deference” should fall too. The court rejects the analogy on the ground that the two forms of deference operate differently.

• Chelvron’s problem (as described in Loper Bright Enterprises v. Raimondo): it required courts to defer to an agency’s reasonable interpretation and thereby displaced the court’s independent judgment on statutory meaning.
• AEDPA’s design (as the Sixth Circuit characterizes it): federal courts still conduct their own constitutional analysis, but § 2254(d) limits the remedial consequence of disagreement unless the state court’s decision is not just wrong but unreasonably so under clearly established Supreme Court holdings.

This distinction is the opinion’s central move: AEDPA does not “commandeer” interpretation; it “channels” collateral relief.

B. AEDPA’s constitutionality as jurisdiction/remedy regulation

The court places AEDPA within a constitutional-historical frame it labels the “Madisonian Compromise” and derives three supportive propositions:

1. State courts are competent constitutional adjudicators (citing The Federalist No. 82 and Tafflin v. Levitt), and habeas should not be conceived as a constitutionally required mechanism for routine federal error correction of state criminal judgments.
2. Congress has broad authority to regulate inferior federal jurisdiction, including the scope of habeas relief (citing Felker v. Turpin and related discussion).
3. Congress may limit remedies (illustrated via Lauf v. E.G. Shinner & Co. and habeas-adjacent doctrines like nonretroactivity under Teague v. Lane, procedural default under Wainwright v. Sykes, and restrictions on successive petitions under McCleskey v. Zant).

Against the “rules of decision” objection rooted in United States v. Klein, the court reads Klein narrowly and relies on Robertson v. Seattle Audubon Society and Bank Markazi v. Peterson to argue that Congress may prescribe a new legal standard—even one that effectively dictates outcomes—so long as it does not usurp judicial power by dictating findings without changing the applicable law. AEDPA, the court concludes, “announced such a rule” (quoting Brown v. Davenport) for collateral relief and is therefore constitutionally permissible.

C. Merits application: insanity instruction and due process

On the insanity instruction, the Sixth Circuit’s key move is to keep the question at AEDPA’s altitude: not whether the instruction could have been better, but whether the Kentucky Supreme Court’s approval was an unreasonable application of clearly established Supreme Court law.

Although the Kentucky Supreme Court’s treatment in Sanders I was brief, the Sixth Circuit presumes a merits adjudication under Johnson v. Williams. It then applies the ambiguity framework associated with Boyde v. California, emphasizing general presumptions about juror competence (Aetna Life Ins. v. Ward, Francis v. Franklin) and Kentucky’s established phrasing (“believe from the evidence”) recognized in Gall v. Commonwealth. The bottom line is institutional: a “fairminded jurist” could conclude jurors understood the difference between the “beyond a reasonable doubt” elements instruction and the “believe from the evidence” insanity instruction; therefore AEDPA forecloses relief.

D. Merits application: ineffective assistance under “double deference”

The opinion methodically segments Sanders’s ineffective-assistance allegations and repeatedly returns to two constraints:

• Substantive constraint: Strickland v. Washington is already deferential to counsel’s strategic choices and requires a concrete showing of prejudice.
• Collateral constraint: when state courts adjudicated the claim on the merits, AEDPA adds a second layer, making the review “doubly deferential” under Knowles v. Mirzayance (and framed in Sixth Circuit terms via Fields v. Jordan and Dunn v. Reeves).

Notable applications include:

• Flenning report / KCPC team: the court treats counsel’s failure to uncover a non-disclosed sub-report as not unreasonably deficient, given the official Walker report’s presentation of a unified team conclusion, and also credits the state court’s view that the report could have enabled damaging impeachment about Sanders’s inconsistent narratives.
• Preparation of Cooke: even acknowledging the prosecution’s strong cross and closing points, the court frames the question as whether no reasonable lawyer could have proceeded with Cooke under the circumstances; it answers no under AEDPA.
• Uncalled jail witnesses: the court rejects the claim largely on an evidentiary-prejudice basis: without affidavits or concrete proffers showing what the witnesses would have said, prejudice is speculative under cases like Tinsley v. Million and Baze v. Parker.
• Mitigation investigation: the court’s analysis is shaped by record constraints. Because the Kentucky Supreme Court adjudicated the mitigation Strickland claim on the merits, the Sixth Circuit refuses to consider new affidavits, invoking AEDPA record limits as explained through Shinn v. Ramirez and Sixth Circuit authority (Upshaw v. Stephenson). It also rejects the “new claim” theory under Franklin v. Jenkins.
• Penalty-phase testimony preparation: the court emphasizes the absence of non-speculative proof that better preparation would have changed Sanders’s testimony in a way that would likely have altered the sentence recommendation, treating the claim as an unsupported counterfactual.
• Competency reevaluation request: reviewed de novo because the Kentucky Supreme Court did not reach the merits, but rejected for lack of prejudice—no reasonable probability a second evaluation would have been granted given the recent KCPC finding and the trial court’s competency colloquy.

E. Cumulative error and procedural default

The court treats cumulative prejudice as a classic procedural-default problem. Sanders raised it later (Sanders III), and Kentucky rejected it as barred for not being raised earlier. Under Coleman v. Thompson and related Sixth Circuit procedural default doctrine, the federal court cannot reach the merits absent cause and prejudice; Sanders failed to show “cause” under Murray v. Carrier.

3.3. Impact

1) Structural impact: AEDPA after Loper Bright Enterprises v. Raimondo

The most consequential doctrinal takeaway is the court’s explicit, published reaffirmation that § 2254(d) survives separation-of-powers attacks even in the post–Loper Bright Enterprises v. Raimondo landscape. The opinion positions “AEDPA deference” as remedial restraint rather than interpretive surrender, a framing likely to be cited whenever habeas petitioners attempt to analogize AEDPA to administrative deference doctrines.

2) Practical habeas impact: record development and mitigation claims

The opinion also underscores the tightening vise on mitigation-based ineffective-assistance claims: once a state court merits-adjudicates a claim, federal courts are largely confined to the state record, and attempts to “upgrade” the claim with new affidavits face both exhaustion/procedural-default problems and § 2254(e)(2) limits as reinforced by Shinn v. Ramirez.

3) Kentucky insanity instructions

While the opinion does not announce a new substantive due process rule, it implicitly signals that Kentucky’s traditional “believe from the evidence” insanity instruction—approved in Gall v. Commonwealth and reaffirmed here—will be difficult to challenge on federal habeas absent an unusually stark record of confusion or contradiction.

4) Intra-circuit and inter-circuit dynamics

Because the opinion is “RECOMMENDED FOR PUBLICATION,” it supplies precedential Sixth Circuit authority on AEDPA’s constitutionality (beyond the earlier unpublished Miles v. Floyd). That may reduce incentives for future litigants in the circuit to raise facial constitutional attacks on § 2254(d), shifting focus to claim-specific merits arguments or to Supreme Court review.

4. Complex Concepts Simplified

• AEDPA “deference” (§ 2254(d)): If a state court already decided a federal constitutional claim “on the merits,” a federal habeas court generally cannot grant relief unless the state decision was not just wrong, but unreasonably wrong under clearly established Supreme Court holdings (or rested on unreasonable factfinding).
• “Clearly established Federal law”: Only Supreme Court holdings (not dicta) count; lower-court precedent cannot supply the rule for § 2254(d)(1).
• “Contrary to” vs. “unreasonable application”: A state court is “contrary to” Supreme Court law if it uses the wrong governing rule or reaches a different result on materially indistinguishable facts; it is an “unreasonable application” if it identifies the right rule but applies it in an objectively unreasonable way.
• De novo review: No deference; the federal court decides the issue fresh. Here, some claims were reviewed de novo only because the state court did not reach the merits.
• Procedural default: If a state court rejects a claim on a state procedural rule (e.g., not raised in time), federal habeas generally cannot review it unless the petitioner shows “cause” and “prejudice.”
• Strickland v. Washington: Ineffective assistance requires (1) deficient performance and (2) prejudice—a reasonable probability that, but for counsel’s errors, the result would have been different.
• “Doubly deferential” review: When a state court has already applied Strickland v. Washington, AEDPA adds a second layer of deference, making relief especially hard to obtain.

5. Conclusion

The Sixth Circuit’s opinion has two principal significances. First, it cements within the circuit—explicitly and in a published decision—that AEDPA § 2254(d) is constitutional and not undermined by the demise of Chevron in Loper Bright Enterprises v. Raimondo, because AEDPA limits collateral remedies rather than compelling federal courts to adopt state courts’ interpretations as binding law. Second, it exemplifies how AEDPA’s deference, record limits, and procedural-default rules combine to make relief difficult even in capital cases featuring contested mental-health defenses and extensive postconviction litigation.

At the same time, Judge Stranch’s dissent highlights the unresolved normative and constitutional tension at AEDPA’s core—whether limiting habeas relief in the face of acknowledged constitutional error functionally compromises Article III’s role. The majority’s resolution, however, is clear: habeas is not an engine of ordinary error correction, and Congress may demand federal restraint unless the state court’s adjudication falls outside the bounds of fairminded disagreement.