AEDPA’s Relitigation Bar Controls Brady Impeachment Claims—Federal Habeas Relief Requires “No Fairminded Disagreement” on Materiality
Commentary on Holberg v. Guerrero (5th Cir. Aug. 14, 2026) (en banc)

I. Introduction

Holberg v. Guerrero is a federal habeas capital case arising from Brittany Marlowe Holberg’s 1996 murder of A.B. Towery in Amarillo, Texas. After a capital-murder conviction and death sentence were affirmed on direct review, Holberg sought state and then federal postconviction relief. The federal district court denied relief under the Antiterrorism and Effective Death Penalty Act of 1996 (“AEDPA”), 28 U.S.C. § 2254(d).

A Fifth Circuit panel later granted relief on one slice of Holberg’s multi-claim petition: a Brady impeachment theory targeting guilt-phase testimony by jailhouse witness Vicki Kirkpatrick, based on Kirkpatrick’s paid informant activities with an Amarillo police officer in other, unrelated cases. The en banc court granted rehearing and—by an equally divided vote—affirmed the district court’s denial of habeas relief.

The central doctrinal dispute is not the classic de novo Brady question (“Was this impeachment evidence Brady material?”), but the AEDPA question: whether the state court’s rejection of Holberg’s Brady theory was so wrong that it was beyond any possibility of fairminded disagreement. A secondary dispute concerns Holberg’s Strickland/Wiggins mitigation-investigation claim.

II. Summary of the Opinion

The en banc court’s judgment is a one-sentence per curiam: “The judgment is AFFIRMED by an equally divided en banc court.” As a result, the district court’s denial of federal habeas relief stands.

Multiple separate opinions provide the reasoning on each side. A principal opinion (authored by Judge Oldham, joined by several judges) would hold that AEDPA bars relief because reasonable jurists could agree with the state court’s conclusions—particularly that the undisclosed impeachment evidence was not material under Brady. Judge Southwick concurred in the judgment on narrower grounds: he would find a Brady violation but would still deny relief because AEDPA requires deference to a reasonable state-court “not material” determination. Several dissents (including Chief Judge Elrod and Judges Higginbotham and Higginson) would grant relief, emphasizing the importance of compensated-informant impeachment in a capital case and arguing the state court’s disposition was unreasonable even under AEDPA.

Because the en banc court is evenly divided, the affirmance leaves the district court’s judgment intact without a single majority rationale. Practically, the case functions as a high-profile illustration of AEDPA’s “relitigation bar” constraining Brady-based habeas relief, even amid serious disagreement over the underlying constitutional merits.

III. Analysis

A. Precedents Cited

1. AEDPA’s relitigation bar and “fairminded disagreement”

  • Harrington v. Richter, 562 U.S. 86 (2011): The principal opinion treats Richter as the cornerstone of § 2254(d) deference, repeatedly invoking its framing of AEDPA as aimed at correcting only “extreme malfunctions,” and its instruction that a petitioner must show there was “no reasonable basis” for the state court to deny relief when the state decision is unexplained.
  • Langley v. Prince, 926 F.3d 145 (5th Cir. 2019) (en banc) (quoting Shoop v. Hill, 586 U.S. 45 (2019) (per curiam); Woods v. Etherton, 578 U.S. 113 (2016) (per curiam)): These cases supply the principal opinion’s articulation of “unreasonable application” as error “beyond any possibility for fairminded disagreement.”
  • Klein v. Martin, 607 U.S. 213 (2026) (per curiam): Used as a modern exemplar of how demanding AEDPA is in Brady materiality disputes; even potentially strong Brady claims can fail under § 2254(d) if reasonable jurists could deem the undisclosed evidence immaterial.
  • Early v. Packer, 537 U.S. 3 (2002) (per curiam), and McCarthy v. Hernandez, 146 S. Ct. 1873 (2026) (per curiam): Cited for the proposition that state courts need not cite federal law—or write a reasoned opinion—to receive AEDPA deference.
  • White v. Woodall, 572 U.S. 415 (2014), Renico v. Lett, 559 U.S. 766 (2010), Parker v. Matthews, 567 U.S. 37 (2012) (per curiam), Marshall v. Rodgers, 569 U.S. 58 (2013) (per curiam), Lopez v. Smith, 574 U.S. 1 (2014) (per curiam): Invoked to reinforce that “clearly established” law comes only from Supreme Court holdings, not circuit precedent, and to caution against expanding doctrine in habeas.
  • Waddington v. Sarausad, 555 U.S. 179 (2009); White v. Wheeler, 577 U.S. 73 (2015) (per curiam); Sexton v. Beaudreaux, 585 U.S. 961 (2018) (per curiam); Shinn v. Kayer, 592 U.S. 111 (2020) (per curiam); Mays v. Hines, 592 U.S. 385 (2021) (per curiam): Cited by the principal opinion to argue the dissents replicate a common error—deciding constitutional merits de novo and then labeling the state court “unreasonable.”

2. Brady doctrine, impeachment evidence, and materiality

  • Brady v. Maryland, 373 U.S. 83 (1963): The foundational rule—suppression of favorable, material evidence violates due process—frames the entire dispute.
  • Kyles v. Whitley, 514 U.S. 419 (1995): Supplies multiple controlling principles: (i) the prosecutor’s duty to learn of favorable evidence “known to the others acting on the government’s behalf in the case, including the police,” and (ii) materiality as the “cumulative effect” undermining confidence in the verdict.
  • Strickler v. Greene, 527 U.S. 263 (1999): Provides the three elements (suppression, favorability, materiality) and the “reasonable probability” materiality formulation.
  • Turner v. United States, 582 U.S. 313 (2017): Reinforces the “reasonable probability” standard for materiality.
  • Banks v. Dretke, 540 U.S. 668 (2004): Cited for the proposition that the State cannot “suppress” information the defendant actually possesses and, in dissent, for the heightened concern with informant testimony in capital cases.
  • United States v. Agurs, 427 U.S. 97 (1976): Used by the principal opinion to reject relief premised on “mere possibility” of helpfulness rather than constitutionally material prejudice.
  • Wood v. Bartholomew, 516 U.S. 1 (1995) (per curiam): Cited to illustrate how defense strategy can bear on materiality/prejudice analysis.
  • Murphy v. Davis, 901 F.3d 578 (5th Cir. 2018): Used to argue cumulative impeachment is less likely to be material.
  • United States v. Bagley, 473 U.S. 677 (1985): Prominent in dissents as authority that non-monetary or contingent inducements affecting credibility are classic Brady impeachment material.

3. AEDPA deference to state-court factfinding

  • Brumfield v. Cain, 576 U.S. 305 (2015): Cited for “substantial deference” and the rule that reasonable disagreement over facts is insufficient under § 2254(d)(2).
  • Burt v. Titlow, 571 U.S. 12 (2013), and Wood v. Allen, 558 U.S. 290 (2010): Used to emphasize that federal courts cannot substitute their view of the record for reasonable state-court factual determinations.

4. Ineffective assistance and mitigation investigation

  • Strickland v. Washington, 466 U.S. 668 (1984), and Wiggins v. Smith, 539 U.S. 510 (2003): Frame the mitigation-investigation claim; the principal opinion stresses that the reasonableness of the investigation, not merely presentation choices, is the core inquiry.
  • Porter v. McCollum, 558 U.S. 30 (2009) (per curiam), Rompilla v. Beard, 545 U.S. 374 (2005), Terry Williams v. Taylor, 529 U.S. 362 (2000): Used to contrast cases where counsel failed to take basic investigative steps.
  • Cullen v. Pinholster, 563 U.S. 170 (2011): Invoked for the “double-edged” nature of mitigation evidence.
  • Dunn v. Reeves, 594 U.S. 731 (2021) (per curiam): Supplies the “doubly deferential” AEDPA + Strickland framing.
  • Thornell v. Jones, 602 U.S. 154 (2024): Cited (with a caveat about timing under Greene v. Fisher, 565 U.S. 34 (2011)) to illustrate how courts weigh aggravation versus mitigation in prejudice analysis.

5. Capital-case systemic framing in dissents

  • Furman v. Georgia, 408 U.S. 238 (1972), Gregg v. Georgia, 428 U.S. 153 (1976), Jurek v. Texas, 428 U.S. 262 (1976): The dissents use these to situate Brady disclosure as part of the legitimacy infrastructure of modern capital punishment.
  • On Lee v. United States, 343 U.S. 747 (1952), and Hoffa v. United States, 385 U.S. 293 (1966): Cited for longstanding Supreme Court skepticism toward compensated informant credibility, making disclosure and cross-examination crucial.
  • Glossip v. Oklahoma, 604 U.S. 226 (2025): Cited in dissent to underscore the gravity of state non-disclosure in capital prosecutions.
  • Tassin v. Cain, 517 F.3d 770 (5th Cir. 2008): Used in dissent to argue this circuit has previously found Brady/AEDPA violations in analogous inducement-driven witness settings.

B. Legal Reasoning

1. The case’s threshold move: Brady merits vs. AEDPA review

The principal opinion insists the panel’s “fundamental error” was treating the controversy as a first-instance Brady dispute rather than an AEDPA dispute: under § 2254(d), the question is whether the state court’s rejection of the Brady claim was beyond the bounds of fairminded disagreement. This reframing draws heavily on Harrington v. Richter and the “extreme malfunctions” concept, and it uses Klein v. Martin to illustrate how even substantial Brady arguments can fail once they are filtered through AEDPA deference.

2. Suppression: how far does the State’s Brady duty to “learn” extend?

On suppression, Holberg’s theory required the prosecutor to discover a witness’s informant/payment arrangement arising from police activity “totally unrelated” to the defendant’s case. The principal opinion emphasizes what it characterizes as a missing Supreme Court directive requiring prosecutors to interview officers unconnected to the prosecution team and suggests the information was not in the “case file” under an open-file policy (citing Kyles v. Whitley for the proposition that Brady does not constitutionalize open-file discovery).

Notably, the principal opinion declines to rest its holding on suppression because it concludes materiality is the dispositive AEDPA bottleneck. The dissents, and Judge Southwick’s concurrence (which views the merits differently), lean on Kyles v. Whitley’s “duty to learn” language and treat compensated-informant impeachment as paradigmatic Brady material, regardless of whether the compensation was arranged in the defendant’s case specifically.

3. Materiality: guilt and punishment as separate lenses

The principal opinion assumes arguendo suppression and resolves the case on materiality: it argues that Kirkpatrick’s testimony was peripheral, cumulative of other confession evidence, discounted by the prosecution as “jailhouse talk,” and dwarfed by independent evidence undermining self-defense and supporting robbery/burglary. Under that view, the state court could reasonably conclude no “reasonable probability” of a different outcome existed (Strickler v. Greene; Kyles v. Whitley).

For punishment, the principal opinion emphasizes the brutality of the crime and additional punishment-phase evidence of prior violence and manipulation, concluding a reasonable jurist could find that impeaching Kirkpatrick would not have undermined the future-dangerousness and mitigation determinations.

The dissents respond that Kirkpatrick was a “key prosecution witness” whose compensated-informant status would have fundamentally altered credibility assessments—especially regarding capital eligibility (robbery intent) and future dangerousness—invoking Banks v. Dretke, United States v. Bagley, and the general due process logic that jurors must have the tools to evaluate incentivized informants.

Judge Southwick’s concurrence occupies a middle position: he views the non-disclosure as a “serious” Brady violation on the merits but believes AEDPA deference still precludes relief because fairminded jurists could treat the impeachment as not material given the rest of the record.

4. Factfinding deference and the “conspiracy” narrative

A significant portion of the principal opinion is devoted to § 2254(d)(2) and the presumption of correctness in § 2254(e)(1), citing Brumfield v. Cain and Burt v. Titlow. It criticizes the dissents for relying on inferences about quid pro quo “bought-and-paid-for” testimony that the state habeas court rejected as factually unsupported, particularly crediting a prosecutor’s testimony that no deal existed in the Holberg case.

The dissents counter that Brady materiality does not require proof of an explicit deal; the witness’s compensated informant status and incentives are themselves impeachment material, and the absence of disclosure deprived the defense and jury of core truth-testing tools.

5. The ineffective-assistance claim: mitigation investigation under AEDPA + Strickland

On Strickland/Wiggins, the principal opinion characterizes the state habeas proceedings as extensive (a multi-day evidentiary hearing; numerous interviews; records review) and credits the state court’s conclusion that counsel pursued a strategic mitigation approach. It then frames the decisive point as lack of Strickland prejudice under AEDPA’s “doubly” deferential standard (Dunn v. Reeves; Harrington v. Richter), reasoning that the aggravating facts of the crime and the mitigation presented at trial made it reasonable for the state court to find no reasonable probability of a different sentence.

C. Impact

1. Practical takeaway for federal habeas litigants

Whatever one’s view of the underlying Brady merits, Holberg underscores a recurring Fifth Circuit theme (and a Supreme Court insistence): in § 2254 cases, the decisive battleground is often not the constitutional standard itself but whether the state court’s application was beyond fairminded disagreement. The opinions demonstrate how materiality—especially as a case-specific, record-intensive judgment—can become the central AEDPA barrier even where judges agree misconduct occurred.

2. Brady impeachment evidence sourced from “other cases”

The disagreement exposes an unresolved fault line: how far Kyles v. Whitley’s duty to learn extends when impeachment evidence exists within the same police department but arises from unrelated investigations. The principal opinion treats the lack of Supreme Court specificity as room for reasonable state-court outcomes; the dissents treat the compensated-informant context as a fundamental permutation where Brady’s application is “beyond doubt.”

3. Capital sentencing narratives and jury confidence

The dissents forcefully frame Brady disclosure as part of capital punishment’s institutional legitimacy (drawing on Furman v. Georgia, Gregg v. Georgia, and Jurek v. Texas). The competing opinions illustrate a broader, recurring tension: whether AEDPA’s deference regime can leave in place capital judgments that some judges deem constitutionally unreliable, so long as a state court’s contrary view is arguable.

4. Precedential effect of an equally divided en banc affirmance

Because the en banc court affirmed by an equally divided vote, the district court’s judgment stands, but no single merits rationale commands a majority. The separate writings nevertheless provide a detailed map of competing approaches to Brady materiality under AEDPA and are likely to be cited in future briefing as persuasive authority on both sides of the Brady/AEDPA interface.

IV. Complex Concepts Simplified

  • AEDPA “relitigation bar” (28 U.S.C. § 2254(d)): After a state court has decided a federal constitutional claim “on the merits,” a federal court cannot grant habeas relief just because it thinks the state court was wrong. Relief is allowed only if the state court’s decision was not merely incorrect but objectively unreasonable under Supreme Court holdings (or based on an unreasonable factual determination).
  • Brady claim (suppression + favorability + materiality): The Constitution is violated when (1) the State fails to disclose (suppresses) evidence, (2) the evidence helps the defense (including impeachment of a key witness), and (3) it is “material,” meaning there is a reasonable probability that disclosure would have changed the outcome or undermined confidence in it.
  • Materiality is not sufficiency: Even if other evidence could support a conviction or sentence, undisclosed impeachment can still be “material” if it would have changed how jurors evaluated critical testimony—especially when the case hinges on credibility.
  • § 2254(d)(2) and § 2254(e)(1): These provisions restrict federal courts from re-deciding factual disputes. State factual findings are strongly presumed correct, making it difficult to win federal habeas relief by recharacterizing the record.
  • Strickland/Wiggins mitigation claims: In capital sentencing, counsel must reasonably investigate mitigating evidence. But under AEDPA, even a debatable deficiency may not justify relief unless the state court’s no-prejudice conclusion is beyond fairminded disagreement.
  • Equally divided en banc affirmance: The lower court’s judgment is affirmed because there is no majority to reverse. The affirmance resolves the parties’ case but does not yield a single controlling rationale from the en banc court.

V. Conclusion

Holberg v. Guerrero is a stark demonstration of AEDPA’s constraining force on Brady-based federal habeas relief. The competing opinions agree on core doctrine—Brady’s elements and AEDPA’s deference—but diverge sharply on how “materiality” and “reasonableness” operate when the withheld impeachment concerns a jailhouse witness with a compensated-informant history in other cases. The judgment—affirmance by an equally divided en banc court—leaves the district court’s denial of relief intact, while the separate writings collectively illuminate the modern fault line in postconviction law: the distance between proving a constitutional violation and proving that a state court’s rejection of it was so unreasonable that no fairminded jurist could agree.