Petition for Writ of Certiorari — Mike Brown, Acting Warden, Petitioner v. Ervine Davenport
Supreme Court briefDec 14, 2020
Ask Donna
What actually matters in this document.
Text
No. _________
In the Supreme Court of the United States
MIKE BROWN, ACTING-WARDEN, PETITIONER
v.
ERVINE DAVENPORT
ON PETITION FOR A WRIT OF CERTIORARI
TO THE UNITED STATES COURT OF APPEALS
FOR THE SIXTH CIRCUIT
PETITION FOR A WRIT OF CERTIORARI
Dana Nessel
Michigan Attorney General
Fadwa A. Hammoud
Solicitor General
Counsel of Record
P.O. Box 30212
Lansing, Michigan 48909
HammoudF1@michigan.gov
(517) 335-7628
Jared Schultz
Assistant Attorney General
Criminal Trials & Appeals
Division
Attorneys for Petitioner
i
QUESTION PRESENTED
In Brecht v. Abrahamson, 507 U.S. 619 (1993), the
Court held that the test for determining whether a
constitutional error was harmless on habeas review is
whether the defendant suffered “actual prejudice.”
Congress later enacted 28 U.S.C. § 2254(d)(1), which
prohibits habeas relief on a claim that was adjudicated on the merits by a state court unless the adjudication “resulted in a decision that was contrary to, or
involved an unreasonable application of, clearly established Federal law.” Although the Court has held that
the Brecht test “subsumes” § 2254(d)(1)’s requirements, the Court declared in Davis v. Ayala, 576 U.S.
257, 267 (2015), that those requirements are still a
“precondition” for relief and that a state-court harmlessness determination under Chapman v. California,
386 U.S. 18 (1967), still retains “significance” under
the Brecht test. The question presented is:
May a federal habeas court grant relief based
solely on its conclusion that the Brecht test is
satisfied, as the Sixth Circuit held, or must
the court also find that the state court’s Chapman application was unreasonable under
§ 2254(d)(1), as the Second, Third, Seventh,
Ninth, and Tenth Circuits have held?
ii
PARTIES TO THE PROCEEDING
Petitioner, Mike Brown, is the Acting Warden
where Respondent Ervine Davenport is currently held
in custody. Duncan MacLaren, who previously was
warden at the facility where Davenport was held, was
appellee in the court below.
RELATED PROCEEDINGS
•
People v. Davenport, Michigan Court of Appeals, Docket No. 287767, Opinion issued August 5, 2010 (affirming conviction).
•
People v. Davenport, Michigan Supreme Court,
Docket No. 141832, Order issued March 9, 2011
(reversing Michigan Court of Appeals judgment
in part and remanding to trial court for evidentiary hearing).
•
People v. Davenport, Kalamazoo County Circuit
Court, Docket No. 07-0165, Opinion After Remand issued October 20, 2011 (finding no reversible error).
•
People v. Davenport, Michigan Court of Appeals, Docket No. 306868, Opinion issued December 13, 2012 (affirming circuit court decision).
•
People v. Davenport, Michigan Supreme Court,
Docket No. 146652, Order issued July 3, 2013
(denying application for leave to appeal).
iii
•
Davenport v. MacLaren, United States District
Court for the Western District of Michigan, Report and Recommendation issued November 7,
2016 (recommending that petition for writ of
habeas corpus be denied).
•
Davenport v. MacLaren, United States District
Court for the Western District of Michigan,
Opinion and Order issued September 26, 2017
(adopting report and recommendation and
denying petition for writ of habeas corpus).
•
Davenport v. MacLaren, United States Court of
Appeals for the Sixth Circuit, Judgment issued
June 30, 2020 (reversing district court and
granting petition for writ of habeas corpus).
•
Davenport v. MacLaren, United States Court of
Appeals for the Sixth Circuit, Order issued September 15, 2020 (denying petition for rehearing
en banc).
iv
TABLE OF CONTENTS
Question Presented...................................................... i
Parties to the Proceeding ........................................... ii
Related Proceedings.................................................... ii
Table of Authorities .................................................. vii
Opinions Below ........................................................... 1
Jurisdiction ................................................................. 1
Constitutional, Statutory, and Regulatory
Provisions Involved .............................................. 1
Introduction ................................................................ 2
Statement of the Case ................................................ 4
Reasons for Granting the Petition ........................... 10
I.
The Sixth Circuit’s Brecht-only approach
conflicts with AEDPA and with the Court’s
analysis in Ayala. ............................................... 10
II. The Sixth Circuit’s Brecht-only approach
creates a split with the Second, Third,
Seventh, Ninth, and Tenth circuits. .................. 14
III. As stated by Judge Thapar in his en banc
dissent, this case presents an “ideal vehicle”
to resolve the conflict over federal review of
state-court harmlessness rulings in habeas. ..... 21
A. The Sixth Circuit relied on circuit
precedent and social-science studies to
find harmful error. ....................................... 21
B. The Sixth Circuit extended this Court’s
decision in Holbrook v. Flynn. ..................... 24
v
C. The Sixth Circuit failed to give
substantial leeway to the state court’s
application of the Chapman standard. ....... 26
Conclusion ................................................................. 29
PETITION APPENDIX TABLE OF CONTENTS
United States Court of Appeals
for the Sixth Circuit
Case No. 17-2267
Opinion
Issued June 30, 2020 ........................................ 1a–69a
United States Court of Appeals
for the Sixth Circuit
Case No. 17-2267
Judgment
Issued June 30, 2020 .............................................. 70a
United States District Court
Western District of Michigan
Case No. 1:14-cv-1012
Opinion and Order
Issued September 26, 2017 ............................. 71a–76a
United States District Court
Western District of Michigan
Case No. 1:14-cv-1012
Judgment
Issued September 26, 2017 ..................................... 77a
vi
United States District Court
Western District of Michigan
Case No. 1:14-cv-1012
Report and Recommendation
Issued November 7, 2016 ............................... 78a–92a
Michigan Supreme Court
Case No. 146652
Order
Issued July 3, 2013 ......................................... 93a–94a
Michigan Court of Appeals
Case No. 306868
Opinion
Issued December 13, 2012 ............................ 95a–100a
United States Court of Appeals
for the Sixth Circuit
Case No. 17-2267,
Order Denying Petition for Rehearing En Banc
Issued September 15, 2020 ......................... 101a–137a
vii
TABLE OF AUTHORITIES
Cases
Brecht v. Abrahamson,
507 U.S. 619 (1993) ............................................... i
Chapman v. California,
386 U.S. 18 (1967) ........................................... i, 10
Czech v. Melvin,
904 F.3d 570 (7th Cir. 2018) .............................. 18
Darden v. Wainwright,
477 U.S. 168 (1986) ............................................ 18
Davis v. Ayala,
576 U.S. 257 (2015) .................................... passim
Deck v. Missouri,
544 U.S. 622 (2005) ............................................ 26
Ford v. Peery,
976 F.3d 1032 (9th Cir. 2020) ...................... 18, 19
Fry v. Pliler,
551 U.S. 112 (2007) ...................................... 11, 12
Hall v. Haws,
861 F.3d 977 (9th Cir. 2017) .............................. 19
Harrington v. Richter,
562 U.S. 86 (2011) .................................... 3, 13, 15
Holbrook v. Flynn,
475 U.S. 560 (1986) .................................. 6, 24, 25
Johnson v. Acevedo,
572 F.3d 398 (7th Cir. 2009) ........................ 16, 17
Johnson v. Lamas,
850 F.3d 119 (3d Cir. 2017) ................................ 15
viii
Johnson v. Superintendent Fayette SCI,
949 F.3d 791 (3d Cir. 2020) ................................ 16
Kernan v. Cuero,
138 S. Ct. 4 (2017) .............................................. 22
Malone v. Carpenter,
911 F.3d 1022 (10th Cir. 2018) .......................... 18
Marshall v. Rodgers,
569 U.S. 58 (2013) .............................................. 26
Mitchell v. Esparza,
540 U.S. 12 (2003) ........................................ 11, 16
O’Neal v. McAninch,
513 U.S. 432 (1995) ............................................ 13
Orlando v. Nassau County District Attorney’s
Office,
915 F.3d 113 (2d Cir. 2019) ................................ 14
People v. Johnson,
597 N.W.2d 73 (Mich. 1999) ............................... 28
Reiner v. Woods,
955 F.3d 549 (6th Cir. 2020) ................................ 8
Rhoden v. Rowland,
172 F.3d 633 (9th Cir. 1999) .............................. 22
Richardson v. Griffin,
866 F.3d 836 (7th Cir. 2017) .............................. 18
Ruelas v. Wolfenbarger,
580 F.3d 403 (6th Cir. 2009) ................................ 7
Ruimveld v. Birkett,
404 F.3d 1006 (6th Cir. 2005) ............................ 22
Spencer v. Capra,
788 F. App’x 21 (2d Cir. 2019) ........................... 15
ix
Welch v. Hepp,
793 F.3d 734 (7th Cir. 2015) .............................. 17
Wharton v. Vaughn,
722 F. App’x 268 (3d Cir. 2018) ......................... 20
White v. Woodall,
572 U.S. 415 (2014) ............................................ 25
Woodford v. Visciotti,
537 U.S. 19 (2002) .............................................. 14
Yarborough v. Alvarado,
541 U.S. 652 (2004) ............................................ 26
Statutes
28 U.S.C. § 1254(1) ..................................................... 1
28 U.S.C. § 2254(d) ..................................................... 2
28 U.S.C. § 2254(d)(1) .................................................. i
Other Authorities
Antonin Scalia,
The Rule of Law as a Law of Rules, 56 U.
Chi. L. Rev. 1175 (1989) ....................................... 9
Constitutional Provisions
U.S. Const. amend. V. ................................................. 1
1
OPINIONS BELOW
The Sixth Circuit’s order denying the State’s petition for rehearing en banc, App. 101a–137a, is reported at 975 F.3d 537. The Sixth Circuit’s panel opinion granting habeas relief, App. 1a–69a is reported at
964 F.3d 448. The district court’s opinion and order
denying habeas relief, App. 71a–76a, is not reported
but available at 2017 WL 4296808. The Michigan Supreme Court’s order denying Davenport’s application
for leave to appeal, App. 93a–94a, is reported at 832
N.W.2d 389. The Michigan Court of Appeals’ opinion
affirming Davenport’s conviction, App. 95a–100a, is
not reported but available at 2012 WL 6217134.
JURISDICTION
The judgment of the court of appeals was entered
on June 30, 2020. The order of the court of appeals
denying the State’s petition for rehearing en banc was
entered on September 15, 2020. Petitioner invokes
this Court’s jurisdiction under 28 U.S.C. § 1254(1).
CONSTITUTIONAL, STATUTORY, AND
REGULATORY PROVISIONS INVOLVED
The Fifth Amendment to the United States constitution provides, “No person shall be . . . deprived of
life, liberty, or property, without due process of law
. . . .”
2
INTRODUCTION
Ervine Davenport was unconstitutionally shackled during his trial, at which a jury found him guilty
of premeditated murder. All 12 jurors later testified
that the shackling did not affect their verdict. Their
decision, they said, was based on the evidence, which
was highlighted by uncontroverted testimony that
Davenport strangled a woman over a foot shorter and
nearly 200 pounds lighter than he.
On direct appeal, the Michigan courts evaluated
the juror’s testimony and the evidence of guilt and
found that the shackling error was harmless beyond a
reasonable doubt. But the Sixth Circuit found those
determinations irrelevant. So long as it makes an independent finding on habeas review that Davenport
was prejudiced, the court held below, the State must
retry him for his heinous crime nearly 14 years ago.
The independent test that the Sixth Circuit used
—from Brecht v. Abrahamson—is certainly a demanding one. But it takes no account of the deference due
state court decisions that Congress commanded when
it enacted 28 U.S.C. § 2254(d) a few years after Brecht.
And it conflicts with this Court’s analysis in Davis v.
Ayala, where this Court stated that AEDPA forms a
“precondition” to relief and retains “significance”
when reviewing a state court’s harmless-error finding.
Not only does the Sixth Circuit’s Brecht-only approach contradict AEDPA and Ayala, it also creates a
circuit split. The Second, Third, Seventh, Ninth, and
Tenth circuits have all required that a federal habeas
court must determine whether AEDPA’s deferential
limitations have been overcome before granting relief.
3
Even the Sixth Circuit is fractured here—it denied rehearing en banc in an 8-to-7 vote.
The requirement that a federal habeas court must
independently examine the state court’s harmless-error finding before addressing Brecht is a meaningful
restraint. It not only respects the state court by examining the court’s analysis on the very issue, it requires
the federal court to conclude that the state court’s
harmless-error analysis was an “extreme malfunction,” as explained in Harrington v. Richter, 562 U.S.
86, 102 (2011), before granting relief. This step restrains the federal court from the all-too-tempting
wish to review de novo and inadvertently reduce the
threshold necessary for granting relief. Review under
AEDPA is legally necessary and practically valuable.
Taking a Brecht-only approach to this case, the
Sixth Circuit restricted the state court’s application of
Chapman’s general harmless-error standard, extended this Court’s prohibitions on considering pretrial juror assumptions to posttrial juror observations,
and relied on circuit precedent and extra-judicial resources to conclude that habeas relief was warranted.
Those are all practices that AEDPA forbids. Had the
court appropriately applied AEDPA’s highly deferential standard, it would have reached an unmistakable
conclusion: the Michigan courts did not unreasonably
apply Chapman by denying Davenport relief.
The Sixth Circuit’s decision below conflicts with
AEDPA. It conflicts with Ayala. It creates a circuit
split. And it offers an ideal vehicle for review because
this circuit’s failure to apply AEDPA to the facts and
circumstances of this case affected its decision. The
Court should grant this petition.
4
STATEMENT OF THE CASE
Ervine Davenport killed Annette White. He admitted it. His jury rejected his self-defense claim at
trial and found him guilty of first-degree premeditated murder.
The trial
The prosecution presented a strong case. Davenport had bragged that, if he had a problem with someone, he would choke them. App. at 4a–5a. And days
before the murder, he did just that—he strangled a
woman until she lost consciousness and urinated on
herself. App. at 121a. Davenport later admitted to
killing White, saying he “offed her.” App. 5a. Davenport also went to White’s home after the murder and
stole property. App. at 120a–121a. When the police
questioned him, he gave three stories: first, that he
knew nothing about White’s death; second, that he
helped dispose of her body but that another person
had killed White; and third, that he had killed White
in self-defense. See App. at 5a n.1.
A forensic pathologist examined White’s body and
opined that she died of manual strangulation. See
App. at 4a. The pathologist explained that a strangled
victim could lose consciousness after 30 seconds. Id.
But death does not occur until the victim is without
air for at least four to five minutes. Id.
At trial, Davenport explained that he and White
had been using cocaine, after which he drove her
home. App. at 3a. She became agitated and tried to
grab the steering wheel, but Davenport pushed her
5
back. Id. He claimed that White then pulled out a box
cutter and cut his arm. Id.
So Davenport extended one arm and pinned her
back against the passenger side of the car, with his
hand under her chin. Id. He said that White dropped
the box cutter but tried to kick him, so he kept his arm
extended across her neck. See App. at 88a. Davenport
eventually noticed that White had stopped struggling.
App. at 3a. Discovering that she was no longer breathing, he panicked and left her body in a field. Id.
Davenport’s testimony was significantly undermined at trial. Although the police found a box cutter
in the trunk of his car, no blood was found on the tool.
App. at 89a. Also, the prosecution elicited from Davenport that he repeatedly lied to the police. See App.
at 90a. He even testified: “[I]t’s not gonna help me any
to tell the truth.” App. at 67a. And during rebuttal,
the forensic pathologist testified that Davenport’s version of events was not possible. App. at 4a. White’s injuries were on both sides of her neck, consistent with
strangulation and inconsistent with broad force being
applied across the front of her neck. Id.
Direct appeal
Davenport appealed and argued that he was unconstitutionally shackled during his trial and that the
trial court had not made any findings regarding the
need for shackles. The Michigan Court of Appeals
found error but determined that it was harmless; however, the Michigan Supreme Court remanded the case
to the trial court for an evidentiary hearing to determine whether Davenport was prejudiced by the shackling. See App. at 95a–96a.
6
On remand, the trial court heard testimony from
all 12 jurors. “Only five jurors testified that they observed defendant’s shackles during trial. While some
of the jurors remembered a comment being made
about the shackling from one of the jurors, all 12 jurors testified that defendant’s shackles were not discussed during deliberations and did not influence the
verdict.” App. at 96a. The trial court found that the
prosecution had proved beyond a reasonable doubt
that the shackling did not affect the jury’s verdict. See
id.
Davenport appealed again. The Michigan Court of
Appeals discussed the jurors’ testimony in detail. App.
at 97a–98a. It found that “[a]ll of the evidence indicated that the shackling did not affect the verdict in
any way.” App. at 98a. Though three jurors testified
that they believed Davenport might be dangerous, the
appellate court found that all three explained that
this belief was based on the first-degree murder
charge itself, not the shackling. Id. The court also
found that “it was proper for the jurors to testify regarding how viewing the shackles affected their deliberations.” App. at 99a. Ultimately, the court held that
“[t]he trial court did not err in finding that the prosecution proved beyond a reasonable doubt that the
shackling error did not affect the verdict.” App. at
100a. The Michigan Supreme Court subsequently denied Davenport’s application for leave to appeal. It explained that although the Michigan Court of Appeals
had erred by failing to apply Holbrook v. Flynn, 475
U.S. 560, 570 (1986), that “error was harmless under
the facts of this case.” App. at 93a–94a.
7
Federal habeas proceedings
Davenport sought federal habeas relief on his
shackling claim. Magistrate Judge Ellen S. Carmody
recommended that the district court deny habeas relief. App. at 91a–92a. In doing so, Judge Carmody discussed the evidence of Davenport’s guilt along with
the juror’s testimony at the evidentiary hearing and
found that the state courts’ harmlessness determination did not meet AEDPA’s stringent requirements.
App. at 87a–91a. Judge Janet T. Neff adopted the report and recommendation and denied habeas relief.
App. at 76a.
In a 2-to-1 decision, a panel of the Sixth Circuit
reversed. The majority rejected the argument that, before granting relief, a habeas court must find actual
prejudice under the Brecht test and conclude that the
state court’s decision was an unreasonable application
of clearly established federal law under AEDPA. App.
at 10a. Instead, noting that “the Brecht standard ‘subsumes’ AEDPA’s unreasonableness inquiry,” the majority held that “ ‘Brecht is always the test’ ” and a habeas court need not also ask whether the state court’s
decision was unreasonable. App. at 10a (quoting
Ruelas v. Wolfenbarger, 580 F.3d 403, 412 (6th Cir.
2009)). Finding only that Davenport was actually
prejudiced, without reference to the state court’s contrary opinions, the majority ruled that the error was
not harmless and granted relief. See App. at 38a.
Judge Readler dissented. He found that a habeas
court must engage in both inquiries when assessing
the harmlessness of a constitutional error. App. at
39a–40a. Failing to do so, the opinion reasoned, contradicts Ayala and left the Sixth Circuit “stand[ing]
8
alone as the only Circuit to award habeas relief without expressly applying the requirements of both
Brecht and AEDPA.” App. at 41a (citing Reiner v.
Woods, 955 F.3d 549, 556–57 (6th Cir. 2020) and noting that it “recogniz[ed] a ‘colorable argument’ that
Ruelas and its progeny are incorrect in light of
Ayala”). The dissent found that the state court harmless-error determinations did not run afoul of AEDPA
and habeas relief was not warranted. See App. at 63a–
64a.
The State filed a petition for rehearing en banc,
arguing that the panel majority applied the wrong
standard for reviewing a state court’s harmless-error
determination. The Sixth Circuit denied rehearing by
an 8-to-7 vote. In a dissenting opinion, Judge Thapar
(joined by five other judges) rejected the approach
taken in the majority panel opinion. 1 According to
Judge Thapar, that approach—that “federal judges
can simply ignore AEDPA’s guardrails whenever they
find that a petitioner has suffered actual prejudice under Brecht”—“casts aside AEDPA and misinterprets
Supreme Court precedent.” App. at 119a. To demonstrate why application of AEDPA deference is required even if the Brecht test is met, Judge Thapar
discussed the differences between the two standards:
AEDPA prohibits extending Supreme Court precedent, it prohibits relying on circuit precedent, it gives
state courts broad discretion when applying a general
standard from the Supreme Court, and it prohibits
considering evidence from outside the state court
1 In a separate dissenting opinion, Judge Griffin also rejected the
approach taken by the majority panel opinion. App. at 114a–
118a.
9
record—the Brecht test does not require adhering to
these clear rules. App. at 129a–133a. “Jettisoning
these clear, rule-based requirements will make appellate review in habeas cases more difficult and unpredictable.” App. at 133a (citing Antonin Scalia, The
Rule of Law as a Law of Rules, 56 U. Chi. L. Rev. 1175
(1989)). And doing so in this case, Judge Thapar said,
shows why application of AEDPA deference would
have required a different result. See App. at 132a.
Judge Thapar also highlighted the “deepening split
among the various federal Courts of Appeals” App. at
133–135a.
Additionally, Judge Sutton (joined by Judge Kethledge) wrote a separate concurring opinion expressing
that he was skeptical that the panel majority applied
the correct standard. App. at 109a–114a. He nevertheless voted to deny rehearing, in part because of his belief that it would be “inefficien[t]” to do so given that
this Court has the “final” say on the matter. App. at
114a; see also id. (“. . . Not only are we fallible, we are
not final either.” (emphasis added)).
10
REASONS FOR GRANTING THE PETITION
I.
The Sixth Circuit’s Brecht-only approach
conflicts with AEDPA and with the Court’s
analysis in Ayala.
By applying the Brecht standard without any regard to the analysis employed by the state courts, the
Sixth Circuit’s decision conflicts with both AEDPA
and Davis v. Ayala.
In Ayala, the Court noted that the test for determining whether a constitutional error was harmless
on habeas review is the Brecht standard—whether
there is “grave doubt” that the error had “substantial
and injurious effect or influence in determining the
jury’s verdict.” 576 U.S. at 267–68. But Ayala reaffirmed that “AEDPA nevertheless ‘sets forth a precondition to the grant of habeas relief,’ ” meaning that relief cannot be granted unless the state court unreasonably applied the harmless-error standard prescribed
in Chapman v. California, 386 U.S. 18, 24 (1967).
Ayala, 576 U.S. at 268. Though reiterating that both
tests need not be formally applied, the Ayala Court focused on the standards outlined in both tests and determined that the prisoner could not meet Brecht nor
could he show that the state court’s determination
was an unreasonable application of Supreme Court
precedent. Id. at 270–86.
When seeking federal habeas corpus relief, it is
undoubtedly true that the prisoner “must meet the
Brecht standard.” Id. at 268. Yet Ayala added an important caveat—“but that does not mean, as the Ninth
Circuit thought, that a state court’s harmlessness determination has no significance under Brecht.” Id.
11
Despite this Court expressly disapproving the Ninth
Circuit’s disregard of a state-court harmlessness ruling, the Sixth Circuit made the same mistake here.
The Sixth Circuit ignored Ayala’s analysis and instead focused on a single clause within its 29-page
opinion: “[I]f the state court adjudicated his claim on
the merits, the Brecht test subsumes the limitations
imposed by AEDPA.” Id. at 270. According to the Sixth
Circuit, this language gives a habeas court license to
ignore AEDPA altogether. But Ayala’s lengthy outline
of AEDPA’s limitations and its application of the
AEDPA/Chapman standard to the case at hand refute
the Sixth Circuit’s interpretation of that lone passage.
As support, the Sixth Circuit pointed to its own
(pre-Ayala) precedents. App. at 10a–11a. Those precedents, in turn, focused on a statement in Fry v. Pliler,
551 U.S. 112, 120 (2007)—the Brecht standard “obviously subsumes” AEDPA/Chapman. App. at 10a–11a.
But the Court in Fry assumed that there was no state
court harmlessness determination in the case at hand,
Fry, 551 U.S. at 116 n.1, meaning AEDPA deference
was not relevant to its decision. The petitioner had argued that an earlier decision, Mitchell v. Esparza, 540
U.S. 12 (2003)—which applied AEDPA/Chapman and
not Brecht—implied that AEDPA had eliminated the
need to apply Brecht on habeas review. Fry, 551 U.S.
at 119. Fry explained why it was not necessary for Esparza to apply the Brecht standard after finding that
AEDPA/Chapman had not been met. Id. at 119–120.
It reasoned that “it is implausible that, without saying
so, AEDPA replaced the Brecht standard of actual
prejudice with the more liberal AEDPA/Chapman
standard.” Id. (internal quotations, citations omitted).
12
But Fry made clear that it was deciding a very
limited question: “whether Brecht or Chapman provides the appropriate standard of review when constitutional error in a state-court trial is first recognized
by a federal court.” Id. at 120. It had no reason to decide the standard—or standards—applicable when a
state court recognizes an error but finds it harmless.
Ayala recognized Fry’s limits: “The Fry Court did
not hold—and would have had no possible basis for
holding—that Brecht somehow abrogates the limitation on federal habeas relief that § 2254(d) plainly sets
out.” Ayala, 576 U.S. at 268. Thus, while reiterating
that “the Brecht test subsumes the limitations imposed by AEDPA,”—meaning Brecht cannot be met if
AEDPA has not been met—it clarified that AEDPA’s
limitations still must be overcome before relief is
granted.
True, AEDPA’s limitations need not be formally
applied, but the Sixth Circuit below expressly
acknowledged that it did not consider the state court
decisions at all, opining instead that the Brecht test
does that work for them. App. at 11a (Stranch, J., majority opinion) (“a habeas court may [also] go straight
to Brecht”) (internal quotes and citation omitted). The
panel dissent explained why this approach is wrong:
. . . AEDPA requires a federal habeas court
to assess whether Supreme Court precedent
put a state court on notice of precise constitutional limitations. See Yarborough, 541 U.S.
at 665. Brecht, on the other hand, writes
largely on a clean slate.
13
Unchecked by then-existing Supreme
Court precedent, Brecht simply asks a federal
habeas court to assess the prejudice arising
from an alleged error. And that distinction can
make all the difference. A habeas claim alleging a deeply prejudicial trial error may easily
clear Brecht’s “actual prejudice” bar. But the
claim may nonetheless fail AEDPA’s comityinspired requirements if the reviewing court
must create new law or extend existing Supreme Court precedent to find underlying legal error, or that the error was not harmless.
Id. at 666.
App. at 49a–50a (Readler, J., dissenting) (paragraph
break added).
As Judge Readler has aptly explained, AEDPA’s
standard requires a federal court to defer to a state
court’s determination; it may displace a sovereign jurisdiction’s “good-faith attempt[ ] to honor constitutional rights” only if it contravenes or unreasonably
applies holdings of this Court. See Harrington v. Richter, 562 U.S. 86, 103 (2011). The Brecht test, though
certainly a high one, does not contain those requirements; a federal habeas court may extend holdings of
this Court, look at circuit precedent, and consider extra-judicial resources. See App. at 129a–133a
(Thapar, J., dissenting) (discussing practices that are
appropriate under Brecht but prohibited under
AEDPA). So long as a court independently concludes
that it is in “grave doubt” as to an error’s impact on
the verdict, it must grant habeas relief. O’Neal v.
McAninch, 513 U.S. 432, 435 (1995).
14
But applying Brecht, without more, fails to give
effect to AEDPA’s purpose of ensuring “that statecourt decisions be given the benefit of the doubt.” See
Woodford v. Visciotti, 537 U.S. 19, 24 (2002). This is
the chief failing of relying on Brecht alone. The Sixth
Circuit’s Brecht-only approach is incompatible with
Ayala and with AEDPA. The Court should grant this
petition.
II. The Sixth Circuit’s Brecht-only approach
creates a split with the Second, Third,
Seventh, Ninth, and Tenth circuits.
Five circuits have rejected the Sixth Circuit’s
Brecht-only approach. These circuits require a review
of a state-court’s harmless-error determination under
AEDPA before habeas relief may be granted.
Start with the Second Circuit. In Orlando v. Nassau County District Attorney’s Office, the court noted
that, “[w]hen a state court makes a harmless error determination on direct appeal, we owe the harmlessness determination itself deference under [AEDPA].”
915 F.3d 113, 127 (2d Cir. 2019) (internal quotations
omitted). Id. Only because the state court did not
make a harmlessness determination in the case at
hand did the Second Circuit apply the Brecht test. A
2019 unpublished decision fully captures the point:
Although the district court correctly inquired as to whether the error at [the petitioner]’s trial yielded a “substantial and injurious effect or influence” on the jury verdict,
recent Supreme Court precedent required
more.
15
Specifically, the inquiry requires considering whether the state courts’ harmlessness determinations were unreasonable. . . . [T]he district court “appears to have treated the unreasonableness question as a test of its confidence
in the result it would reach under de novo review” and thus “overlook[ed] arguments that
would otherwise justify the state court’s result.”
Spencer v. Capra, 788 F. App’x 21, 23 (2d Cir. 2019)
(emphasis added, and paragraph break added) (citing
Ayala, and quoting Harrington v. Richter, 562 U.S. at
102).
The Third Circuit has also rejected a Brecht-only
approach. In a post-Ayala case, the court discussed
the impact that this Court’s decision had on this question and required that the court “defer” to the state
court’s determination:
In Davis v. Ayala, the Court confirmed that
the Brecht standard still governs our harmless
error analysis on collateral review. However,
where a state court has concluded that the error was harmless on direct review, the Supreme Court clarified that we must defer to
that determination under AEDPA unless the
state court unreasonably applied Chapman v.
California.
Johnson v. Lamas, 850 F.3d 119, 133 (3d Cir. 2017)
(footnotes and citations omitted; and emphasis
added).
16
In a more recent decision, the Third Circuit underscored the importance of applying AEDPA deference to a state-court’s harmless-error analysis. In
Johnson v. Superintendent Fayette SCI, the court
noted that the state court had not conducted a harmless-error analysis, and it found actual prejudice under the Brecht test. 949 F.3d 791, 798–99, 804 (3d Cir.
2020). The court then referenced Johnson v. Lamas
and explained why that precedential opinion did not
compel a different result. Id. at 804. Despite the similarity of the errors, the state court in Lamas had undergone a harmless-error analysis, meaning that
AEDPA deference applied. Id. A finding of harmlessness in Lamas was compelled by the different standard of review. Id. In the case at bar, the court contrasted, the harmless-error question was reviewed de
novo under Brecht. Id. By implication, the court reiterated its conclusion in Lamas that AEDPA is a separate standard that could necessitate a different outcome than that reached under a Brecht-only approach.
The Seventh Circuit first grappled with this issue
before Ayala was decided. In Johnson v. Acevedo, the
court noted that the “harmless-error question has
some difficulties” before discussing the two competing
approaches. 572 F.3d 398, 403 (7th Cir. 2009). At the
time, two decisions seemingly governed the answer—
Esparza and Fry. The Seventh Circuit first focused on
the pronouncement in Esparza that “ ‘habeas relief is
appropriate only if the [State] Court of Appeals applied harmless-error review in an “objectively unreasonable” manner.’ ” Acevedo, 572 F.3d at 404 (quoting
Esparza, 540 U.S. at 18). The court then looked to Fry,
asking whether the decision held that the Brecht test
is the only applicable test for all collateral attacks. Id.
17
The Seventh Circuit rejected that interpretation because the issue in Fry was merely the standard to apply when a state court did not conduct a harmless-error analysis. Id.; see also supra at 11–12. The court
ultimately outlined its approach, which reviewed the
reasonableness of the state court’s analysis on harmlessness:
If the state court has conducted a harmlesserror analysis, the federal court must decide
whether that analysis was a reasonable application of the Chapman standard. If the answer is yes, then the federal case is over and
no collateral relief issues. That’s the holding
of Esparza. If the answer is no—either because the state court never conducted a harmless-error analysis, or because it applied
Chapman unreasonably—then § 2254(d)
drops out of the picture and the federal court
must make an independent decision, just as if
the state court had never addressed the subject at all. And we know from Fry that, when
this is so, a federal court must apply the
Brecht standard to determine whether the error was harmless. . . .
Acevedo, 572 F.3d at 404 (emphasis added).
Post-Ayala, the Seventh Circuit still uses this approach. In Welch v. Hepp, the court pointed out
Ayala’s requirement that actual prejudice under
Brecht be shown, but it stated that, because the state
court in the case at bar conducted a harmless-error
analysis, “[t]hat determination is subject to deference.” 793 F.3d 734, 738 (7th Cir. 2015). The Welch
court went on to discuss the reasonableness of the
18
state court’s analysis, ultimately denying relief. Id. at
739. And in Richardson v. Griffin, the Seventh Circuit
again addressed a habeas claim by asking whether a
state court’s harmless-error analysis overcame
AEDPA’s deferential standard of review. 866 F.3d
836, 841 (7th Cir. 2017). The Richardson court found
that the state court’s analysis was contrary to and an
unreasonable application of clearly established federal law because the state court had not applied the
Chapman standard in its harmless-error analysis. Id.
at 844. Only then did the Seventh Circuit go on to discuss whether the petitioner was actually prejudiced
under Brecht. 2 Id. at 844–45.
A recent decision from the Ninth Circuit similarly
reflects this approach. After finding a due-process violation under Darden v. Wainwright, 477 U.S. 168
(1986), the Ninth Circuit in Ford v. Peery determined
that the Darden violation necessitated a finding of actual prejudice. 976 F.3d 1032, 1044 (9th Cir. 2020).
Nevertheless, the court found it necessary to also ask
whether the state court’s harmless-error determination was unreasonable under Chapman. Id. at 1044–
45. After finding both standards met, the court
granted habeas relief. Id. at 1045.
The Tenth Circuit takes a similar approach. In
Malone v. Carpenter, the court said that “satisfaction
of the AEDPA/Chapman standard is a necessary condition for relief.” 911 F.3d 1022, 1030 (10th Cir. 2018).
2 But see Czech v. Melvin, 904 F.3d 570, 577 (7th Cir. 2018). Alt-
hough the court in Czech focused only on the Brecht test, it determined that there was no actual prejudice, so it had no reason
to address whether AEDPA was necessarily met after succeeding
under Brecht.
19
The Tenth Circuit then found that the state court’s
harmless-error determination was not an unreasonable application of Supreme Court precedent. Id. at
1031–37. It only reached Brecht in a one-sentence
analysis that it said it would have employed if it alternatively were to accept the argument that the state
court applied the wrong harmless-error standard. Id.
at 1037. Thus, the Tenth Circuit applies Brecht only if
the state court’s harmless-error analysis was unreasonable.
The Sixth Circuit acknowledged many of these decisions from its sister circuits but nevertheless found
no conflict. According to the court, that some courts
choose to perform a harmless-error analysis under
both standards does not mean that they have rejected
a Brecht-only approach. See App. at 25a–27a (finding
that, even though a federal habeas court may choose
to apply both tests, “it is not necessary” because a
Brecht-only approach encompasses both). But no identified decision—except the one below—has indicated
that the application of AEDPA before granting habeas
relief was merely an option. 3
3 The panel majority argued that the Ninth Circuit has found
that analysis under AEDPA is optional, citing Hall v. Haws, 861
F.3d 977, 992 (9th Cir. 2017). App. at 22a–23a. But despite saying that it “need not apply both” tests, the Hall court nevertheless did go on to “separately analyze the state court’s harmless
error determination under AEDPA/Chapman.” Hall, 861 F.3d at
992. And, regardless, as discussed above, the same circuit three
years later expressly applied the AEDPA/Chapman standard after already finding actual prejudice under Brecht. Ford, 976 F.3d
at 1044–45. And even if the Sixth Circuit were correct, it would
just mean that it is a four-to-two circuit split, rather than a fiveto-one split.
20
Indeed, performing an optional analysis that requires detailing the reasoning behind a state court
opinion and contrasting it with this Court’s holdings
would be an unnecessary waste of time and resources
if it was understood that the case was won or lost under Brecht. In other words, a court’s decision to stop
its discussion after finding the Brecht test unsatisfied
says nothing about whether it employs the Sixth Circuit’s express Brecht-only approach.
The Sixth Circuit also pointed out that none of its
sister-circuit decisions had undergone an analysis under AEDPA after finding actual-prejudice existed under Brecht. But that is because none of those cases
found actual prejudice under Brecht. Because the failure to meet either standard precludes relief—as Ayala
says—it would make little sense to perform a second
test after the first fails. 4 Conversely, the Sixth Circuit
fails to appreciate that none of these circuits have
granted habeas relief before making an explicit determination regarding the reasonableness of the underlying state-court decision. Put differently, each of
these circuits acknowledges that habeas relief may be
denied solely under the Brecht test, but habeas relief
may not be granted unless AEDPA’s deferential
standard is also met.
4 See App. at 23a (Stranch, J., majority opinion) (pointing to an
unpublished decision, Wharton v. Vaughn, 722 F. App’x 268, 277
(3d Cir. 2018), for its proposition that the Third Circuit applies a
similar Brecht-only approach, but failing to recognize that the
decision merely limited its harmless-error inquiry to the Brecht
test because no actual prejudice was found, see id. at 278; it thus
had no reason to subsequently apply AEDPA and is entirely consistent with the State’s argument).
21
Simply put, Ayala instructed federal courts that
AEDPA forms a “precondition” to habeas relief and
fully applies to state-court harmlessness determinations. 576 U.S. at 268 (internal quotation omitted).
The Second, Third, Seventh, Ninth, and Tenth circuits
have all recognized that this instruction demands a
deferential review of a state court’s harmless-error determination. The Sixth Circuit below, in contrast, reasoned that no deferential inquiry is necessary. Because there is a circuit split, the Court should grant
the petition.
III. As stated by Judge Thapar in his en banc
dissent, this case presents an “ideal vehicle”
to resolve the conflict over federal review of
state-court harmlessness rulings in habeas.
The question presented was squarely raised to the
Sixth Circuit. It was fully analyzed in a panel majority
opinion, a dissenting opinion, and competing opinions
of an almost equally divided circuit when deciding
whether to rehear the case en banc. And this case exemplifies why the differences between the Brecht test
and AEDPA’s highly deferential standard are important.
A. The Sixth Circuit relied on circuit
precedent and social-science studies to
find harmful error.
In finding actual prejudice under Brecht, the Sixth
Circuit relied on authority that cannot be considered
under AEDPA’s standard.
22
When discussing whether the evidence against
Davenport could prove the shackling error harmless,
the court relied in part on Ruimveld v. Birkett, 404
F.3d 1006, 1016 (6th Cir. 2005), one of its own precedents that found actual prejudice stemming from a
shackling error and that affirmed a habeas grant.
App. at 23a–24a, 32a. But such a comparison of lower
federal court cases is not allowed when performing an
analysis under AEDPA. The Court has warned lower
courts time and again that “circuit precedent does not
constitute clearly established Federal law, as determined by the Supreme Court.” Kernan v. Cuero, 138
S. Ct. 4, 9 (2017) (internal quotations omitted).
The Sixth Circuit repeated its mistake when rejecting the argument that the juror testimony supported a harmless-error finding. Pointing out that the
Ninth Circuit had found that a shackling error was
not harmless in a similar case (though the juror testimony in that case occurred during habeas proceedings, not in the state courts), the court relied on it to
show that the Brecht test was met here. App. at 37a–
38a (quoting Rhoden v. Rowland, 172 F.3d 633, 637
(9th Cir. 1999)). Never did the Sixth Circuit determine
whether this Court’s precedents prohibit a state court
from finding harmless error when all 12 jurors explicitly disclaim any effect caused by the shackling.
And to further discredit the jurors’ explicit testimony, the Sixth Circuit relied on “a voluminous body
of social-science research.” App. at 34a. The court attempted to cloak its reliance on this extra-judicial research by stating that it “merely provides further support for the Supreme Court’s determination” in Flynn
that shackling is “inherently prejudicial.” App. at 35a.
23
But that statement is hard to reconcile with the
Sixth Circuit panel majority’s footnote inserted just
three lines earlier where it explicitly relied on a 2019
study to buttress its conclusion that shackling Davenport created a presumption of dangerousness and
guilt because he is “a 6′5″ tall black man weighing approximately 300 pounds.” App. at 34a n.13. This race
and size component to the shackling harmless-error
determination has no support in the record and is
found nowhere in Supreme Court precedent.
The Sixth Circuit’s actual-prejudice analysis in
this case wholly circumvented the reins that AEDPA
places on habeas review. Judge Thapar explained why
this case demands further review:
. . . [T]he differences between Brecht and
AEDPA matter. If AEDPA applies, the panel
decision is plainly erroneous since it extends
Supreme Court precedent, relies on circuit
precedent, creates a new standard for harmless-error review in the shackling context, and
introduces evidence not presented in the state
court proceedings. Thus, this case presents an
ideal vehicle for clarifying the relationship between Brecht and AEDPA.
App. 133a (Thapar, J., dissenting) (emphasis added;
internal citation omitted). Put simply, because applying both standards in this case would have changed
the panel majority’s decision here, this case is an ideal
vehicle to clarify a recurring problem in habeas cases.
24
B. The Sixth Circuit extended this Court’s
decision in Holbrook v. Flynn.
The Sixth Circuit also found that the state court
improperly relied on the juror testimony because “the
Supreme Court has made clear that jurors’ subjective
testimony about the effect shackling had on them
bears little weight.” App. at 34a. But this Court has
adopted no such holding.
In Holbrook v. Flynn, this Court reviewed
whether the presence of uniformed state troopers in
the courtroom for security was so “inherently prejudicial” so as to deny a fair trial. 475 U.S. at 562, 569. It
noted that many of the prospective jurors were asked
about the troopers’ presence during voir dire and the
trial court found that the answers indicated that the
troopers’ presence did not create an inference of guilt.
Id. at 565. Finding the “assessment of jurors’ states of
mind cannot be dispositive here,” the Court said
“[e]ven though a practice may be inherently prejudicial, jurors will not necessarily be fully conscious of
the effect it will have on their attitude toward the accused.” Id. at 570. But this did not foreclose the approach taken here by the state courts for two reasons.
First, the juror testimony in Flynn occurred before
trial, see id. at 565, where they were asked to explain
the effect the troopers’ presence might have. As Judge
Thapar put it: “That is a far cry from post-trial testimony from jurors about how shackling actually affected their verdict.” App. at 129a. Indeed, part of
Flynn’s reasoning rested on the fact that the jurors
could “only speculate on how they will feel after being
exposed to a practice daily over the course of a long
trial.” 475 U.S. at 570.
25
Here, the jurors did not need to speculate—they
had already been exposed to the shackling error and
could assess the impact it had on their verdict. The
Sixth Circuit therefore had to extend Flynn, an action
that is prohibited by AEDPA. See White v. Woodall,
572 U.S. 415, 426 (2014) (“Section 2254(d)(1) provides
a remedy for instances in which a state court unreasonably applies this Court’s precedent; it does not require state courts to extend that precedent or license
federal courts to treat the failure to do so as error.”).
Second, Flynn’s discussion of the jurors’ testimony
was limited to whether there was an underlying constitutional violation. 475 U.S. at 570. This Court ultimately found that there was no violation because the
troopers’ presence was not “inherently prejudicial.”
Id. at 570–71. It thus had no reason to employ a harmless-error analysis. Although juror testimony cannot
be considered when determining whether a courtroom
security practice is “inherently prejudicial,” it does not
necessarily follow that juror testimony cannot be considered when determining whether that practice
amounted to harmless error.
Here, it is uncontroverted that Davenport’s shackling was “inherently prejudicial” and was error: the
question was whether that error was harmless. Because neither Flynn, nor any other Supreme Court
precedent, prohibits consideration of juror testimony
in answering that question, the Michigan courts’ consideration of the jurors’ post-trial testimony was well
within the range of reasonable jurisprudence.
By extending Flynn beyond its holding, the Sixth
Circuit’s Brecht-only approach failed to appreciate the
limits imposed by AEDPA.
26
C. The Sixth Circuit failed to give
substantial leeway to the state court’s
application of the Chapman standard.
Under AEDPA, lower federal courts cannot “refine
or sharpen a general principle of Supreme Court jurisprudence into a specific legal rule that [the Supreme] Court has not announced.” Marshall v. Rodgers, 569 U.S. 58, 63 (2013) (per curiam) (citation omitted). So, when “evaluating whether a rule application
was unreasonable,” a habeas court must “consider[ ]
the rule’s specificity. The more general the rule, the
more leeway courts have in reaching outcomes in caseby-case determinations.” Yarborough v. Alvarado, 541
U.S. 652, 664 (2004).
The rule at issue here is a general one—Chapman’s harmless-beyond-a-reasonable-doubt standard.
Although Deck v. Missouri provided a framework for
identifying a shackling error, it did not discuss how to
analyze such an error under the Chapman standard.
544 U.S. 622, 635 (2005). Thus, the general principles
announced in Chapman are all that guide a federal
court’s review of a state court’s determination that a
shackling error was harmless.
Taking its Brecht-only approach, the Sixth Circuit
failed to explain why the Michigan court determinations fell outside the substantial leeway it was allowed
in applying Chapman to the facts. The Michigan
Court of Appeals found that all of the jurors who observed Davenport in shackles “believed that there was
nothing unusual” about it and that “every juror testified that defendant’s shackles were not discussed during jury deliberations and that the verdict was based
solely on the evidence presented at trial.” App. at 98a.
27
The state court concluded that the error was harmless
beyond a reasonable doubt. Id. In its analysis, the
court also noted that, although several jurors testified
that they believed that Davenport was dangerous,
each testified that their belief was based on the firstdegree murder charge itself, not the shackling. Id.
Finally, the state court noted that its harmlessness finding was buttressed by the fact that there was
overwhelming evidence of Davenport’s guilt. App. at
99a n.2. Then, in denying leave to appeal, the Michigan Supreme Court said that the juror testimony
should have been evaluated differently, but it too held
that “the substantial evidence of guilt presented at
trial” rendered any error harmless. App. at 93a–94a.
Instead of considering the reasonableness of this
approach and analyzing whether it contravened
Chapman, the Sixth Circuit reviewed the harmlessness inquiry de novo. It set forth a lengthy dissertation on the difference between first- and second-degree murder, applying its interpretation of Michigan
law to its own independent balancing of the facts. App.
28a–33a. While a state court could have found, as the
Sixth Circuit did, that it was a close call whether Davenport committed first-degree premeditated murder,
it was not unreasonable to conclude that the evidence
overwhelmingly supported first-degree murder.
Indeed, the record allows for a reasonable ruling
that there was overwhelming evidence that this murder was premeditated. Davenport described to a witness his penchant for choking people. App. at 4a–5a.
Shortly before the murder, he strangled another
woman until she lost consciousness and urinated.
App. at 121a.
28
After murdering White, Davenport left her body
in a field and stole property from her home. App. at
120a–121a. He even confessed that he had “off[ed]”
White, and he repeatedly lied to the police. App. at 5a
and n.1. He later claimed self-defense, but his physical explanation was not consistent with her injuries.
App. at 3a, 4a. The Sixth Circuit found that this evidence did not overwhelmingly prove Davenport guilty
of first-degree premeditated murder. App. at 32a. This
finding was in turn supported by two other findings:
(1) that the murder occurred during a fight, and (2)
that Michigan law provides that strangulation alone
cannot prove premeditation. App. at 30a–32a. Both
findings ignored critical features in Davenport’s case.
First, whether a fight occurred is largely irrelevant—Davenport strangled White to death, an act
that required completely depriving her of air for at
least four to five minutes and could only have been
completed after White had already lost consciousness.
App. at 4a. By finding that the jury “easily could have
found that this was second-degree [non-premeditated]
murder,” App. at 31a, the Sixth Circuit ignored evidence showing that Davenport continued to strangle
her for a significant duration after she no longer posed
any threat. Nothing in the Sixth Circuit’s Brecht-only
approach explains why a rational jurist could not have
weighed this evidence differently than it did.
Second, although strangulation alone cannot
prove premeditation in Michigan, People v. Johnson,
597 N.W.2d 73, 79 (Mich. 1999), there was much more
here. Along with the actual act of strangling White—
which enable him to take a “second look,” id. (quotation, citation omitted)—here the jury also heard more:
29
•
Davenport had a plan to use deadly force when
in a confrontation,
•
Davenport had done so before, and
•
Davenport took actions after the murder that
revealed his state of mind.
The Sixth Circuit’s independent review fails to see the
true picture of the crime and, in any event, fails to appreciate that a rational jurist could have viewed this
case differently.
In ruling that the error was not harmless, the
Sixth Circuit discussed its own view of the facts as it
applied to state law and failed to provide the appropriate leeway that AEDPA demands. The state court
decision here was reasonable in denying relief.
CONCLUSION
The petition for writ of certiorari should be
granted.
Respectfully submitted,
Dana Nessel
Michigan Attorney General
Fadwa A. Hammoud
Solicitor General
Counsel of Record
Dated: DECEMBER 2020
Jared Schultz
Assistant Attorney General
Criminal Trials & Appeals Division
Attorneys for Petitioner
This is a copy of a public record, reproduced as it was published. It is not legal advice, and it may not be the version a court would rely on. Check the official source before you cite it.