dissenting:
I respectfully dissent from the panel’s disposition of this case. The district court carefully considered the impact of the state’s use of statements elicited from Alford in violation of Miranda v. Arizona, 384 U.S. 436, 86 S.Ct. 1602, 16 L.Ed.2d 694 (1966), After a full review of the record, the district court concluded that the introduction of Alford’s interrogation statements had a “substantial and injurious effect or influence in determining the jury’s verdict.” Brecht v. Abrahamson, 507 U.S. 619, 637, 113 S.Ct. 1710, 123 L.Ed.2d 353 (1993).1 I agree. The fact that statements taken in violation of Miranda generally may be introduced for impeachment purposes does not alter my conclusion that the unconstitutional use of Alford’s statements at his trial was not harmless.
First, it is not clear that the interrogation statements at issue actually would have been admissible as impeachment evidence. The interrogation statements appear to have been consistent with Alford’s statements at trial. The prosecutor’s criticism of Alford’s interrogation statements was that they were less detailed than the account Alford offered at trial, not that he had changed his story. The prosecutor focused not on what Alford said during his interrogation, but what he did not say once he had invoked his right to counsel. Interrogation statements violating Miranda only are admissible “should the Defendant testify inconsistently.” Pollard v. Galaza, 290 F.3d 1030, 1033 (9th Cir. 2002). I do not think Alford did so here.
Second, the majority does not dispute that the jury was never instructed that the interrogation statements could not be used as substantive evidence of guilt, which is required whenever such statements are proffered for impeachment purposes. As the district court found, there is a very real possibility that the jury did consider his statements as substantive evidence of guilt. Indeed, they were repeatedly implored to do just that by the prosecutor. During her closing argument, the prosecutor argued at length that Alford’s demean- or and statements were incongruous with how one “would expect somebody to behave if they had just killed somebody by accident.” As set forth in the district court’s description of the record, page after page of the prosecutor’s closing repeated this theme. Although the prosecution introduced other substantive evidence of guilt, the interrogation apparently loomed large in the jury’s decisionmaking because it asked to watch the recorded interrogation on the third day of deliberation. I would find that the introduction of the interrogation statements, in combination with the prosecutor’s extensive use of those statements as substantive evidence of guilt, had a sufficiently large effect on the jury’s verdicts—beyond the possible use of the interrogation as impeachment evidence—to warrant habeas relief.
Third, even if Alford did not exhaust an argument under Harrison v. United States, 392 U.S. 219, 88 S.Ct. 2008, 20 L.Ed.2d 1047 (1968), that the erroneous introduction of his interrogation statements induced his trial testimony, I believe the district court was correct to consider the fact that Alford’s testimony might have differed if not for the introduction of the interrogation. Alford argued that reliance on his testimony as actually proffered was inappropriate. He made this argument in response to the state’s contention that the introduction of his interrogation statements was harmless because Alford’s trial testimony constituted sufficient evidence of guilt. Alford was not raising a freestanding claim for habeas relief under Harrison', he was responding to the state’s argument that its admitted substantive use of tainted testimony was harmless. His point was that, even if he might have still testified at trial, the court should not assume his testimony would have been elicited in the same way. Exhaustion is simply not an issue for that argument, Thus, even if, as the district court concluded, it is not clear whether Alford’s testimony was induced in its entirety by the introduction of the interrogation statements, we must assume that the introduction of those statements at the very least “affected [Alford’s] strategy” at trial. Garcia v. Long, 808 F.3d 771, 784 (9th Cir. 2015). The district court properly .considered this effect in considering whether the admission was harmless.
The fact that the interrogation statements do not likely even qualify as impeachment evidence, combined with the prosecutor’s undue emphasis on those statements, the trial court’s failure to properly instruct the jury regarding the proper purpose for which such statements may be considered, and the district court’s factual finding that Alford’s testimony might well have differed in the absence of the state’s constitutional violation, convince me that the district court reached the right conclusion here. We should affirm.
. "Because it is more stringent, the Brecht test ‘subsumes’ the AEDPA/'Chapman standard for review of a state court determination of a constitutional violation.” Deck v. Jenkins, 814 F.3d 954, 985 (9th Cir. 2016) (quoting Fry v. Pliler, 551 U.S. 112, 120, 127 S.Ct. 2321, 168 L.Ed.2d 16 (2007)). As this court has explained, "[a] determination that the (trial court's] error resulted in ‘actual prejudice’ necessarily means that the state court’s harmlessness determination was not merely incorrect, but objectively unreasonable.” Id, (quoting Brecht, 507 U.S. at 637, 113 S.Ct. 1710) (citing Davis v. Ayala, — U.S. -, 135 S.Ct. 2187, 2198-99, 192 L.Ed.2d 323 (2015)). "A federal habeas court therefore need not formally apply both the Brecht test and the AED-PA standard; it is sufficient to apply Brecht alone.” Id. (citing Fry, 551 U.S. at 120, 127 S.Ct. 2321).