Justices to consider appropriate standard for harmless-error review of state convictions in federal habeas proceedings
on Oct 3, 2021 at 1:26 pm
In Brown v. Davenport, the justices will grapple once again with how a provision in the Antiterrorism and Effective Death Penalty Act of 1996 affects state prisoners’ access to federal habeas corpus review. This case, which will be argued on Tuesday, addresses the standard of review that federal habeas courts should use when assessing state courts’ prior determinations that constitutional trial errors were harmless.
Fifty-four years ago, in Chapman v. California, the Supreme Court held that most preserved, constitutional trial errors will result in reversal on direct appeal unless the state can prove that the errors were “harmless beyond a reasonable doubt.” But in 1993, in Brecht v. Abrahamson, the court adopted a different standard of harmless-error review for federal habeas corpus cases, citing interests in finality and the deference due state decisions in our federalist system. For habeas cases, Brecht requires a showing of “actual prejudice,” meaning that the constitutional violation in question must have had a “substantial and injurious effect or influence in determining the jury’s verdict,” before a prisoner may obtain relief.
Three years after Brecht, Congress passed AEDPA, which provides that no state prisoner may obtain habeas relief on the basis of a constitutional claim that has been “adjudicated on the merits in State court proceedings” unless the state court adjudication “resulted in a decision that was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States” or involved an unreasonable determination of the facts. Questions immediately arose about the relationship between AEDPA and the Brecht harmless error standard. In Fry v. Pliler and Davis v. Ayala, the Supreme Court held that AEDPA did not displace Brecht, but rather that Brecht “subsumes” AEDPA. Ervine Lee Davenport’s case now asks the court to address exactly what that means.
Davenport was convicted of first-degree, premeditated murder and sentenced to life in prison without the possibility of parole after a trial where he was visibly shackled in front of the jury. Three years before Davenport’s trial, the Supreme Court held in Deck v. Missouri that shackling a criminal defendant in front of the jury is inherently prejudicial and violates a defendant’s due process rights unless the state can demonstrate a “special need” to justify the shackling. No such justification was offered in Davenport’s case.
On appeal, Davenport argued that his conviction should be reversed because of this constitutional violation, but the Michigan appellate courts determined that the error was harmless. At trial, there had been no question that Davenport had killed the victim. The only questions were whether he had acted in self-defense and whether he had premeditated the killing. In a post-trial evidentiary hearing in state court, jurors testified that they had seen the shackles, had discussed the shackles during the trial, and thought that Davenport was dangerous. But the jurors also testified that they did not discuss the shackling during deliberations and did not believe that it affected their verdict. The Michigan trial and intermediate appellate courts relied on this latter testimony to hold that the error was harmless. The Michigan Supreme Court denied leave to appeal, but only after rejecting the Michigan Court of Appeals’ reliance on juror testimony to support its harmless-error analysis. The state high court chastised the intermediate appellate court’s reasoning as inconsistent with Holbrook v. Flynn – a case in which the Supreme Court rejected reliance solely on jurors’ assessments of whether they were prejudiced when addressing whether courtroom security procedures were so inherently prejudicial as to violate due process. But the Michigan Supreme Court denied leave to appeal noting “the substantial evidence of guilt presented at trial.”
Davenport then filed a federal habeas corpus petition challenging the constitutionality of his continued confinement. The state admitted the due process violation. The only question was whether the error was harmless. The federal district court denied relief, but a divided panel of the U.S. Court of Appeals for the 6th Circuit reversed. The 6th Circuit majority applied Brecht and held that the shackling had a substantial and injurious effect on the verdict. Having determined that Brecht was satisfied, the 6th Circuit believed that Davenport was entitled to habeas relief. It did not believe that it needed to conduct a separate AEDPA analysis, because Brecht “subsumes” AEDPA. Other circuits require habeas petitioners to satisfy Brecht and also separately demonstrate that the state court’s Chapman determination was unreasonable under AEDPA. The 6th Circuit voted 8-7 to deny en banc review, with two of the eight judges (Chief Judge Jeffrey Sutton and Judge Raymond Kethledge) writing that they believed the panel decision was wrong but not en banc worthy. The Supreme Court then took the case.
Michigan, as petitioner, argues that state prisoners must formally run the gamut of both Brecht and AEDPA before they may obtain relief. That would mean that Davenport would have to show both that the shackling had a substantial and injurious effect or influence on the verdict in his case and that the state court’s determination that the shackling error was harmless was contrary to or involved an unreasonable application of clearly established federal law. The state argues that AEDPA shifted federal habeas review from a regime in which federal judges independently assessed state courts’ judgments on constitutional claims to a regime that requires deference to state courts’ reasonable judgments. This deferential regime, Michigan contends, requires federal courts to assess harmlessness determinations from the state courts’ perspective rather than substituting their own independent judgment as Brecht allows. Additionally, the state notes that AEPDA does not allow federal courts to extend Supreme Court precedents to new contexts, does not permit reliance on circuit court precedent, prohibits consideration of evidence from outside the state court record, and gives state courts broad discretion when applying general standards. Because Brecht has none of those limitations, Michigan argues that it does violence to AEDPA to allow Brecht to replace a separate AEDPA analysis.
Davenport counters by arguing that when a state prisoner draws only on the legal and factual materials permissibly considered under AEDPA, Brecht’s determination that there was actual prejudice necessarily means that a contrary state court determination is unreasonable. Brecht requires “more than a reasonable possibility that the error was harmful.” So, Davenport argues, if there is more than a “reasonable possibility” that the error was harmful, then no fair-minded jurist could agree with the state court’s finding that the prosecution proved that the error was harmless beyond a reasonable doubt. That is why Brecht subsumes AEDPA and obviates the need for a separate AEDPA inquiry. In those circumstances, a separate, formal AEPDA inquiry is duplicative and unnecessary.
Davenport admits that a separate AEDPA inquiry would be required if a state prisoner brought in new evidence outside of the trial record, relied on arguments not advanced in state court or on circuit rather than Supreme Court precedent, or ignored the state court’s analysis. But he contends that he did not do any of those things, so the Brecht analysis should be sufficient. (The state argues that the 6th Circuit relied on circuit precedent, brought in new social science evidence, and addressed new arguments never raised in state court such that, even under Davenport’s own test, a separate AEPDA inquiry is required.)
If Michigan prevails and a separate AEDPA inquiry is required, it remains to be seen whether the Supreme Court will remand the case to the 6th Circuit to conduct that separate inquiry or address the issue itself. Interestingly, the two sides cannot even agree on what state court decision would be the subject of an AEDPA analysis. Davenport argues that the Michigan Supreme Court decision is the last reasoned decision, because it rejected the harmless-error reasoning of the Michigan Court of Appeals. The state argues that the Michigan Supreme Court order was not a reasoned decision, because it denied the application for leave to appeal. Instead, the state wants the federal court to look through the high court order to the Michigan Court of Appeals’ decision on the merits to conduct an AEDPA analysis. Either way, Davenport argues that he can satisfy AEDPA’s strict requirements, and the state contends he cannot.
This case, while technical, has implications for when and whether state prisoners will be able to obtain habeas relief. A holding that Brecht automatically satisfies AEDPA – even in a subset of cases – would mean one less hurdle for habeas petitioners to obtain relief. But the Roberts court has increasingly narrowed the scope of federal habeas review under AEDPA’s re-litigation bar, making it exceedingly difficult for state prisoners whose claims have been adjudicated in state court to obtain relief. And the Supreme Court has been quite willing to scold lower courts – and the 6th Circuit in particular – for failing to show state court decisions sufficient deference under AEDPA.