George Kayer v. Charles L. Ryan

944 F.3d 1147

Opinion

FOR PUBLICATION

UNITED STATES COURT OF APPEALS FOR THE NINTH CIRCUIT

GEORGE RUSSELL KAYER, No. 09-99027 Petitioner-Appellant, D.C. No. v. 2:07-cv-02120-DGC

CHARLES L. RYAN, Director of the Arizona Department of ORDER Corrections, Respondent-Appellee.

Filed December 18, 2019

Before: William A. Fletcher, John B. Owens, and Michelle T. Friedland, Circuit Judges.

Order; Concurrence by Judges W. Fletcher and Friedland; Dissent by Judge Bea

SUMMARY*

Habeas Corpus / Death Penalty

The panel filed an order denying a petition for panel rehearing, and denying on behalf of the court a petition for rehearing en banc, in a case in which the panel (1) reversed in part and affirmed in part the district court’s judgment denying Arizona state prisoner George Russell Kayer’s habeas corpus petition and (2) remanded with directions to grant the writ with respect to Kayer’s death sentence.

Judges W. Fletcher and Friedland concurred in the denial of rehearing en banc. Responding to their dissenting colleagues’ arguments, they wrote that they are acutely aware of the deference required under AEDPA, and that even giving all appropriate deference to the decision of the post- conviction-relief court judge, habeas relief is warranted.

Judge Bea, joined by Judges Bybee, Callahan, M. Smith, Ikuta, Owens, Bennett, R. Nelson, Bade, Collins, Lee, and Bress, dissented from the denial of rehearing en banc. He wrote that by any fair reading of the panel majority’s opinion, it reviewed the post-conviction-review court’s decision de novo as to whether an Arizona court, applying Arizona precedent, would have granted relief—a radical approach unwarranted under the Antiterrorism and Effective Death Penalty Act. He also wrote that beyond the legal errors, Kayer’s proposed mitigating evidence is hardly overwhelming, and reasonable jurists could find that it did not

* This summary constitutes no part of the opinion of the court. It has been prepared by court staff for the convenience of the reader. KAYER V. RYAN 3

undermine confidence in the death sentence, providing no basis for relief under AEDPA’s deferential standard.

COUNSEL

Jennifer Y. Garcia (argued) and Emma L. Smith, Assistant Federal Public Defenders; Jon M. Sands, Federal Public Defender; Office of the Federal Public Defender, Phoenix, Arizona; for Petitioner-Appellant.

John Pressley Todd (argued), Special Assistant Attorney General; Jacinda A. Lanum, Assistant Attorney General; Lacey Stover Gard, Chief Counsel; Dominic Draye, Solicitor General; Mark Brnovich, Attorney General; Office of the Attorney General, Phoenix, Arizona; for Respondent- Appellee.

ORDER

Judges W. Fletcher and Friedland voted to deny the petition for panel rehearing and rehearing en banc. Judge Owens voted to grant the petition for panel rehearing and rehearing en banc.

The full court has been advised of the petition for rehearing en banc. A judge of the court requested a vote on en banc rehearing. The matter failed to receive a majority of the votes of non-recused active judges in favor of en banc consideration. Fed. R. App. P. 35.

Judge Hurwitz was recused and did not participate in the deliberations or vote in this case.

The petition for panel rehearing and rehearing en banc is DENIED. A concurrence in the denial by Judges W. Fletcher and Friedland and a dissent from the denial by Judge Bea are filed concurrently with this order.

W. FLETCHER and FRIEDLAND, Circuit Judges, concurring in the denial of rehearing en banc:

Our opinion in this capital case speaks for itself. See Kayer v. Ryan, 923 F.3d 692 (9th Cir. 2019). However, our colleagues’ dissent from the denial of en banc review makes new and unfounded arguments to which we feel it appropriate to respond.

George Kayer shot and killed his friend Delbert Haas in Arizona while returning from a gambling trip to Nevada. Kayer, Lisa Kester (Kayer’s girlfriend), and Haas were in Haas’s van. Kayer was driving. Kayer had already indicated to Kester that he would kill Haas. The three of them had consumed a case of beer during the several-hour drive. Kayer took a back road and stopped the van. When Haas went to the back of the van to urinate, Kayer shot him. Kayer and Kester drove away, but returned when they realized Kayer had not gotten Haas’s house keys. When they returned, Haas did not appear to be dead. Kayer shot him again, killing him. Ten days later, when Kayer and Kester returned to Nevada, Kester approached a security guard at a Las Vegas hotel and told him what had happened. Kayer and Kester were both charged with capital murder. Kester testified against Kayer in return for a reduced sentence of three years probation. Id. at 695–96. KAYER V. RYAN 5

Our dissenting colleagues do not dispute that Kayer’s counsel performed deficiently. Kayer’s first lawyer, Linda Williamson, was inexperienced and incompetent. She represented Kayer for a year and a half. During that time, she did no work to prepare for the penalty phase of Kayer’s trial. Id. at 702–03. Kayer’s second lawyer, David Stoller, was experienced but incompetent. He represented Kayer for eleven months. During that time, he, like Williamson, did no work to prepare for the penalty phase. Id. at 703–04. The jury returned a guilty verdict on March 26, 1997. Stoller’s mitigation expert first interviewed Kayer on May 21, 1997, almost two months later, six days before the date originally set for the sentencing hearing. Id. at 704.

As a result of counsels’ deficient preparation, the mitigation evidence at the sentencing hearing was meager. It took only part of a morning. There were five witnesses: (1) a detention officer who testified that Kayer was well behaved in the jail law library; (2) Kayer’s mother, who testified that, to her knowledge, Kayer had never killed anything or anyone since shooting jackrabbits as a teenager; (3) Kayer’s half- sister, who testified that Kayer had “highs and lows,” had drinking and gambling problems, and had, “I guess,” been diagnosed “as a bipolar manic-depressive, or something like that”; (4) the mitigation expert, who testified she had not had enough time to gather information that would support “a medical opinion about a diagnosis of a psychiatric condition”; and (5) Kayer’s mentally impaired son, who gave eleven lines of testimony. Id. at 696–98.

In Arizona at the time, capital sentences were imposed by judges rather than juries. The Supreme Court would not decide Ring v. Arizona, 536 U.S. 584 (2002), until five years later. Under Arizona law, a sentencing judge balanced

aggravating and mitigating circumstances. There were specified statutory aggravating circumstances, but no non- statutory aggravating circumstances. There were specified statutory mitigating circumstances, but any other mitigating circumstances could be considered as well. Statutory mitigators were given greater weight than non-statutory mitigators.

The sentencing judge found two statutory aggravating factors under Arizona law: (1) that Kayer had previously been convicted of a “serious offense”; and (2) that the murder had been committed for “pecuniary gain.” ARIZ. REV. STAT. § 13-703(F)(2), (F)(5) (1977). (All references are to the 1997 version of Arizona Revised Statutes.) The judge explicitly refused to find as an additional aggravating circumstance that the murder had been committed in “an especially heinous, cruel or depraved manner.” Id. at § 13-703(F)(6); Kayer, 923 F.3d at 698. The judge found one non-statutory mitigating factor—that Kayer had “become an important figure in the life of his son.” The judge sentenced Kayer to death. Id. at 698.

During this pre-Ring period, the Arizona Supreme Court resentenced de novo in capital cases on direct appeal, giving no deference to a sentencing decision of the trial judge. In its de novo resentencing of Kayer in 1999, the Arizona Supreme Court found the same two statutory aggravating factors and the same single non-statutory mitigating factor. Like the sentencing judge, it did not find the additional statutory aggravating circumstance that the murder had been committed in “an especially heinous, cruel or depraved manner.” It sentenced Kayer to death. State v. Kayer, 194 Ariz. 423, 984 P.2d 31 (1999). KAYER V. RYAN 7

On state post-conviction review (“PCR”), Kayer’s lawyers claimed that he had received ineffective assistance of counsel (“IAC”) at the sentencing phase. His lawyers presented extensive evidence of Kayer’s mental illness and of mental illness in Kayer’s family, none of which had been presented at the sentencing hearing. We describe that evidence at length in our opinion. To recapitulate the main points:

Kayer’s father was an alcoholic and obsessive gambler. Kayer’s Aunt Opal on his mother’s side was schizophrenic (“I have [heard voices] all my life. . . . It runs in the family”). She testified that Kayer had told her, “I thought it was normal[.] I hear voices, too.” Kayer, 923 F.3d at 711. Kayer’s Aunt Ona Mae on his mother’s side was an alcoholic with severe mood swings. Kayer’s Aunt Tomi on his mother’s side was an alcoholic and a severe depressive. Kayer’s cousin on his mother’s side was schizophrenic and bipolar. Id.

Kayer himself was slow to walk and fell often. As a small boy, he had so many bruises on his body that his mother would not take him out in public. He was dyslexic and got very poor grades in school. He enlisted in the Navy after high school but was quickly discharged with a mental “impairment” described in the discharge papers as “severe.” Id. at 709. He had two unsuccessful marriages in his early twenties. He began committing property crimes in his mid- twenties, and became a heavy drinker and compulsive gambler. He checked himself into a VA hospital in his late twenties, saying “I just want to know what’s wrong.” Id. at 710. Six years later, he again checked himself into a VA hospital, where a doctor wrote that he “showed bipolar traits” and prescribed lithium (a standard medication for bipolar

disorder). He was given a “provisional diagnosis” of “Personality Disorder/Bipolar.” Id. at 710–11. Kayer told a probation officer a year later that until the second stay in the VA hospital, “he had no idea what was wrong with him.” Id. When Kayer was forty, he suffered a severe heart attack and was admitted to a VA hospital. He checked himself out of the hospital “against medical advice.” Id. He killed Haas six weeks later.

Three doctors testified in the PCR court without contradiction. Dr. Anne Herring testified that Kayer “demonstrated significant difficulty when required to execute complex problem solving,” and that “similar deficits have been associated with chronic heavy substance abuse, traumatic brain injury, and with bipolar disorder.” Id. at 712. Dr. Michael Sucher, an addiction specialist, testified to his “untreated alcoholism and untreated pathological gambling.” Id. Dr. Barry Morenz, a psychiatrist, characterized Kayer’s beliefs as “really delusional.” Among other things, Kayer had believed ever since he was a boy, and continued to believe as an adult, that he was a reincarnated being from another planet. Id. Dr. Morenz diagnosed Kayer’s mental state at the time of the murder: “He was having problems with bipolar disorder symptoms and may have been manic or hypomanic, he was having difficulties with out of control pathological gambling and he had difficulty with extensive alcohol abuse.” Id. at 713.

The Arizona judge who presided over Kayer’s trial and sentenced him to death also presided over his state PCR proceeding. In a very brief order, the state PCR judge denied Kayer’s IAC claim. He held that Kayer’s trial attorneys, Williamson and Stoller, had provided professionally competent service, despite the fact that Williamson did no KAYER V. RYAN 9

mitigation work whatsoever, and Stoller’s mitigation expert did not even begin work until six days before the originally scheduled sentencing hearing. Alternatively, the state PCR judge held that Kayer had not shown prejudice: “This court further concludes that if there had been a finding that the performance prong of the Strickland standard had been met, that no prejudice to the defendant can be found.” Id. at 714 (emphasis in the judge’s order). The Arizona Supreme Court denied Kayer’s petition for review without explanation. Id. at 700. The state PCR judge’s decision was therefore the last reasoned state court decision.

We held that there had been deficient performance by counsel at the penalty phase, and that the state PCR judge had been objectively unreasonable, within the meaning of AEDPA, in concluding otherwise. Our colleagues have not disputed this holding. Counsels’ failure to prepare for the penalty phase hearing was egregious, and the mitigation evidence presented at the hearing was pathetically inadequate. See Rompilla v. Beard, 545 U.S. 374 (2005).

We also held that the no-prejudice decision by the state PCR judge was an objectively unreasonable decision within the meaning of AEDPA. Our dissenting colleagues object to this holding.

I. Our Reasoning

There were three steps in our reasoning:

A. Step One

First, we compared the aggravators and mitigators at the two different stages in state court:

1. Sentencing Phase and Direct Appeal

In the trial court and in the Arizona Supreme Court on direct de novo review, there were two statutory aggravators and one non-statutory mitigator. No mitigating factor—either statutory or non-statutory—was found based on mental impairment. Given the meager evidence presented at sentencing, we held that the Arizona Supreme Court had “made a reasonable determination of the facts in concluding that Kayer suffered from no mental impairment.” Kayer, 923 F.3d at 702.

The first statutory aggravator was a prior conviction for a “serious offense.” ARIZ. REV. STAT. § 13-703(F)(2). Kayer’s prior conviction was for first degree burglary. This conviction is the least serious of the “serious offenses” under the aggravator. Serious offenses range from burglary to first degree murder, second degree murder, manslaughter, aggravated assault resulting in serious physical injury, sexual assault, and any dangerous crime against children. See ARIZ. REV. STAT. § 13-703(H)(1)–(6). The second statutory aggravator was commission of a crime for “pecuniary gain.” See ARIZ. REV. STAT. § 13-703(F)(5). The gain in Kayer’s case was relatively modest: avoiding repayment of a $100 loan from Haas, and stealing money and jewelry from Haas’s person and personal property from his house. Neither the sentencing judge nor the Arizona Supreme Court found the proposed statutory aggravator of killing in “an especially heinous, cruel or depraved manner.” ARIZ. REV. STAT. § 13- 703(F)(6).

The one non-statutory mitigator was Kayer’s importance in the life of his son. KAYER V. RYAN 11

2. State PCR Proceeding

Based on the extensive evidence presented during the state PCR proceeding, we concluded that Kayer had established the statutory mitigator of mental impairment under Arizona law: “The defendant’s capacity to appreciate the wrongfulness of his conduct or to conform his conduct to the requirements of law was significantly impaired, but not so impaired as to constitute a defense to prosecution.” ARIZ. REV. STAT. § 13-703(G)(1). In order to reach that conclusion, we analyzed Arizona Supreme Court cases in which that statutory mitigator had been found. Kayer, 923 F.3d at 718 (providing as examples State v. Stevens, 158 Ariz. 595, 764 P.2d 724, 727–29 (1988) (long-term alcohol and drug use); State v. Gretzler, 135 Ariz. 42, 659 P.2d 1, 16–17 (1983) (long-term drug use)). The state PCR judge made no finding, one way or the other, whether Kayer had established the statutory mitigator of mental impairment. If he had made a finding that Kayer had not established this statutory mitigator, the finding would have been objectively unreasonable, given the clear case law of the Arizona Supreme Court.

The Strickland prejudice question in the PCR court was the effect of the addition of the new statutory mitigator of “mental impairment” to the relatively weak non-statutory mitigator of “importance in the life of his son,” balanced against the same two statutory aggravators.

B. Step Two

Second, we recited the established law for determining prejudice in a Strickland IAC case under AEDPA. Under that law, we do not look to what the initial sentencing judge

would have done if the later-presented evidence had been presented at the sentencing hearing. Instead, we look to the probability of a different outcome in the Arizona Supreme Court, which sentences de novo in capital cases. We filter the Strickland standard through the lens of AEDPA to give appropriate deference to the decision of the state PCR judge. Kayer, 923 F.3d at 719–20.

The prejudice standard under Strickland is not whether the newly introduced evidence would “more likely than not have produced a different outcome.” Rather, the Strickland prejudice standard is the less demanding standard of “reasonable probability”:

The defendant must show that there is a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different. A reasonable probability is a probability sufficient to undermine confidence in the outcome.

Strickland v. Washington, 466 U.S. 668, 694 (1984). When filtered through the lens of AEDPA, the standard is that articulated by the Supreme Court in Porter v. McCollum. The Strickland prejudice question for a federal habeas court under AEDPA is whether

it was objectively unreasonable [for the state habeas court] to conclude there was no reasonable probability the sentence would have been different if the sentencing judge . . . had heard the significant mitigation evidence KAYER V. RYAN 13

that [defendant’s trial] counsel neither uncovered nor presented.

Porter v. McCollum, 558 U.S. 30, 31 (2009) (per curiam).

C. Step Three

Third, we compared the facts of Kayer’s case to the facts of other Arizona capital cases to determine prejudice. We discussed several Arizona Supreme Court cases and concluded that one case in particular predicted what that court would likely have done if the information presented during Kayer’s state PCR proceeding had been presented at the original sentencing hearing. Id. at 721–23.

In State v. Brookover, 124 Ariz. 38, 601 P.2d 1322 (1979), defendant Brookover had agreed to buy 750 pounds of marijuana from the victim. When the marijuana was delivered, Brookover shot the victim in order to avoid paying for it. “The victim fell to the floor moaning and asked the defendant what he had done. The defendant said ‘Don’t worry . . . it will be over soon’ and shot him once more in the back,” killing him. Id. at 1323. There were essentially the same two statutory aggravators in Brookover’s case as in Kayer’s case: (1) conviction for a prior “serious offense,” though this aggravator, at the time of Brookover’s sentencing, required the crime be one for which the death penalty could be imposed; and (2) killing for pecuniary gain (recognized a year later, retroactively, as a statutory aggravator). As in Kayer’s case, the Brookover court rejected a statutory aggravator of killing in “an especially heinous, cruel, or depraved manner.” There was also the same mitigating factor that in Kayer’s case had been established only after he obtained competent counsel during the state court PCR

proceedings: “mental impairment.” Unlike in Kayer’s case, there was no additional mitigator in Brookover’s case. In its de novo sentencing determination in Brookover, the Arizona Supreme Court held that a death sentence could not be imposed. It held, “Under the circumstances, leniency is mandated.” Id. at 1326 (emphasis added).

The comparison between Kayer’s case and Brookover is striking. To summarize: Both shot their victims twice, wounding them with the first shot and, after time for deliberation, killing them with the second shot. Both men shot and killed their victims for “pecuniary gain.” In neither case was the pecuniary gain great. Both men had prior convictions for “serious crimes,” though Kayer’s was a much less serious crime than Brookover’s. Both men had the statutory mitigator of “mental impairment.” Kayer had an additional mitigator, the non-statutory mitigator of importance in the life of his son. Our dissenting colleagues call Kayer’s crime a “brutal and venal murder.” Dissent at 43. But it was no worse than the murder in Brookover. Indeed, the courts in both Kayer’s case and in Brookover specifically rejected the proposed statutory aggravator that the murder had been committed in “an especially, heinous, cruel or depraved manner.”

Given the striking similarity between the facts of Brookover and the facts of Kayer’s case, and given that the Arizona Supreme Court had held in Brookover that a non- capital sentence was “mandated,” we held that the state court judge was “objectively unreasonable” in holding that there was “no reasonable probability” that Kayer’s sentence would have been different if the evidence presented to the PCR court had been presented in the original sentencing hearing. We wrote: KAYER V. RYAN 15

In determining prejudice, we need not go so far as Brookover. We need not decide that leniency was “mandated” and that the state PCR court was unreasonable in concluding otherwise. We need only decide whether “it was objectively unreasonable” for the state court to conclude that there was “no reasonable probability” that Kayer’s sentence would have been different if Kayer’s attorneys had presented to the sentencing court the mitigating evidence later presented to the PCR court. Porter [v. McCollum], 558 U.S. at 31 . . . . In light of the foregoing, and particularly in light of the Arizona Supreme Court’s decision in Brookover, we hold that there is a reasonable probability Kayer’s sentence would have been less than death, and that the state PCR court was unreasonable in concluding otherwise.

Kayer, 923 F.3d at 723.

II. Our Colleagues’ Dissent

Our dissenting colleagues make two arguments based on mistakes of law.

First, our colleagues argue that we were required to give deference to the prejudice decision of the state PCR judge on the ground that he made the initial sentencing decision. They write:

[W]ho better to determine whether the new evidence would have made a difference at

sentencing than the judge who sentenced Kayer to death. Judge Kiger presided over both sentencing and the PCR proceedings, and he concluded the new evidence would have made no difference. His “unique knowledge of the trial court proceedings”—including his front-row seat to the presentation of evidence showing Kayer’s brutal and venal murder—“render[ed] him ‘ideally situated’” to evaluate Kayer’s claim that the introduction of new evidence would have changed the sentencing outcome. Murray v. Schriro, 882 F.3d 778, 821 (9th Cir. 2018) (quoting Landrigan, 550 U.S. at 476). This is not to say that Judge Kiger is entitled to some sort of super-deference simply because he sentenced Kayer to death. But there is something particularly troubling about the panel majority affording no deference whatsoever to Judge Kiger’s PCR court decision, as the last reasoned state court opinion.

Dissent at 42–43 (emphasis added). They also write:

All this evidence was before the state PCR court, which concluded that Kayer had not been prejudiced by his trial counsel’s failure to introduce this evidence in mitigation and before sentencing. He should know, because in Arizona the same judge who presides over a defendant’s trial and sentencing also presides over the PCR proceeding.

Id. at 41 (emphasis added). KAYER V. RYAN 17

In making their “he-should-know” argument, our colleagues ignore clear Supreme Court law to the contrary. Strickland itself—the foundation case—tells us not to give deference to the prejudice determination of the state PCR judge on the ground that he or she was also the sentencing judge. Under AEDPA, we give deference to the decision of the last reasoned state court decision. If that is the decision of the state PCR judge at the trial court level, we of course give deference to that decision. But the fact that the PCR judge was also the sentencing judge is irrelevant to the deference we should give to his or her prejudice determination. The Court wrote:

The assessment of prejudice should proceed on the assumption that the decisionmaker is reasonably, conscientiously, and impartially applying the standards that govern the decision. It should not depend on the idiosyncracies of the particular decisionmaker, such as unusual propensities toward harshness or leniency.

Strickland, 466 U.S. at 695.

Our colleagues’ reliance on Murray and Landrigan is misplaced. In Murray, we held only that it does not violate due process to have the same person act as both the trial judge and PCR judge. Murray, 882 F.3d at 820–821. In Landrigan, the Supreme Court held that the PCR judge, who had also been the trial judge, was “ideally situated” to evaluate a factual claim about what had been said during trial in a colloquy between the judge and the defendant. Schriro v. Landrigan, 550 U.S. 465, 476 (2007). Neither Murray nor Landrigan even remotely support the proposition that we owe

deference to a Strickland prejudice determination by the PCR judge on the ground that he or she was also the trial judge.

Second, our colleagues argue that in determining prejudice we should not have looked to precedential decisions of the Arizona Supreme Court. Calling our approach a “stunning error,” they write that

the panel majority . . . proposes that the yardstick for whether there is a reasonable probability Kayer would not have been sentenced to death if the new evidence were presented to the sentencing court is whether this case is more like cases in which the Arizona Supreme Court at one point affirmed a death penalty imposed by the trial court on direct de novo review or more like cases in which the Arizona Supreme Court reversed. . . . [This] mode of habeas review of a Strickland claim [] is quite literally unprecedented.

Dissent at 44–45 (emphasis in original).

As a factual matter, our colleagues are mistaken in saying that this mode of analysis is “quite literally unprecedented.” In White v. Ryan, 895 F.3d 641 (9th Cir. 2018)—another Arizona capital case, which we discussed and applied in our opinion—we spoke directly to this issue. Kayer, 923 F.3d at 720. We wrote in White that an analysis of prejudice at sentencing must look to what the Arizona Supreme Court would likely have done if the evidence has been presented to it on direct appellate review. We faulted the state habeas court for failing to perform this analysis. We wrote, “The KAYER V. RYAN 19

PCR court erred by . . . fail[ing] to consider the probability of a different outcome in the Arizona Supreme Court.” Id. at 671.

As we pointed out in our en banc opinion in McKinney v. Ryan, 813 F.3d 798 (9th Cir. 2015) (en banc), the Arizona Supreme Court is conscientious in following its own precedents. There, we wrote:

[T]he Arizona Supreme Court has a strong view of stare decisis. The Court wrote in White v. Bateman, 89 Ariz. 110, 358 P.2d 712, 714 (1961), for example, that its prior caselaw “should be adhered to unless the reasons of the prior decisions have ceased to exist or the prior decision was clearly erroneous or manifestly wrong.” See also Young v. Beck, 227 Ariz. 1, 251 P.3d 380, 385 (2011) (“[S]tare decisis commands that ‘precedents of the court should not be lightly overruled,’ and mere disagreement with those who preceded us is not enough.” (quoting State v. Salazar, 173 Ariz. 399, 416, 844 P.2d 566 (1992))); State ex rel. Woods v. Cohen, 173 Ariz. 497, 844 P.2d 1147, 1148 (1993) (referring to a “healthy respect for stare decisis”); State v. Williker, 107 Ariz. 611, 491 P.2d 465, 468 (1971) (referring to a “proper respect for the theory of stare decisis”).

Id. at 826.

As a matter of law, our colleagues are also mistaken. At all times relevant to our decision, the Arizona Supreme Court

reviewed de novo on direct appeal all sentencing decisions in capital cases. The prejudice question is necessarily the following: Is there a reasonable possibility that there would have been a different decision by the Arizona Supreme Court if that court had seen the newly presented evidence on direct appeal? The only way to answer that question is to compare the evidence—including the newly presented evidence—to the evidence in other cases reviewed by the Arizona Supreme Court on direct appeal.

Our colleagues write further:

The [panel majority’s] rule is as misguided as it is novel. For starters, [its] approach would make federal habeas review of every Strickland claim turn on the state in which the petitioner was sentenced. So U.S. Supreme Court habeas precedents that involve California apparently could be distinguished away in habeas appeals from Arizona, on the sole ground that “we ask what an Arizona rather than a California sentencing court would have done.” [Kayer, 923 F.3d at 724.] The panel majority appears untroubled by this point, but its implications are striking: Their approach—at least for Strickland prejudice—transmutes “clearly established Federal law, as determined by the Supreme Court of the United States” into law as determined by state supreme courts. 28 U.S.C. § 2254(d)(1).

Dissent at 45–46. KAYER V. RYAN 21

Our colleagues misunderstand the nature of an IAC claim. IAC claims in § 2254 habeas petitions are often—even usually—premised on the law of the particular state in which the petitioner was convicted. If an attorney fails to make what would have been a winning claim under state law, a federal habeas court determines prejudice by asking what the decision under that state law would likely have been if the claim had been made. We do not look to the law of another state or to federal law when the state court would never have applied that law. For example, in an IAC claim where a petitioner argues that counsel should have raised a claim in Arizona state court under Arizona law, we do not ask what California or federal law exists on the point, or what a California or federal court would have done. The IAC claim is based on what the Arizona court would have done under Arizona law had the claim been presented. Our colleagues are right that our approach would often have us look to state law in addressing petitions raising IAC claims. But they are mistaken in contending that the approach is “novel.” On the contrary, it is the normal and uncontroversial approach.

III. Brookover

Our colleagues do not want to confront the Arizona Supreme Court’s decision in Brookover. Both arguments just reviewed are designed to persuade the reader that decisions of the Arizona Supreme Court in capital cases, and Brookover in particular, are irrelevant to the Strickland prejudice question. Our colleagues would prefer to regard as the controlling case a habeas challenge to a decision by the California Supreme Court. Dissent at 54–56.

Our colleagues discuss Brookover only briefly, and only at the very end of their long dissent. They try to avoid the effect of Brookover in two ways.

First, our colleagues point out that Brookover’s mental impairment came from an “organic brain injury.” Id. at 53. They compare Brookover’s impairment to what they characterize as Kayer’s “self-administered ‘untreated alcoholism and untreated pathological gambling.’” Id. In thus referring to Kayer’s mental state, they ignore his “severe” “mental impairment” when he was discharged from the Navy as a very young man; his two stays in VA hospitals, resulting in a bipolar diagnosis and lithium prescription; his hearing voices, as described by his aunt; his delusional beliefs, including the belief that he came from another planet; and the extensive mental illness in his family. More important, the Arizona Supreme Court concluded that Brookover’s “mental impairment” was a statutory mitigator because his “mental condition was . . . a major and contributing cause of his conduct . . . .” Brookover, 601 P.2d at 1326. There is nothing in the Court’s opinion specifying that the cause of the impairment is relevant. The relevant fact is the impairment itself.

Second, our colleagues dismiss Brookover as a case decided “forty years ago.” Dissent at 44–45. When Kayer’s case was decided on direct appeal by the Arizona Supreme Court, Brookover was twenty (not forty) years old. We wrote in our opinion:

The Arizona Supreme Court in capital cases routinely cites and treats as binding precedent its own decisions from twenty years (and more) before. See, e.g., State v. Hedlund, 245 KAYER V. RYAN 23

Ariz. 467, 431 P.3d 181, 190 (2018) (discussing and distinguishing State v. Graham, 135 Ariz. 209, 660 P.2d 460 (1983); State v. Trostle, 191 Ariz. 4, 951 P.2d 869, 885 (1997) (discussing and relying on State v. Richmond, 114 Ariz. 186, 560 P.2d 41, 52–53 (1976))). See also State v. Stuard, 176 Ariz. 589, 863 P.2d 881, 902 (1993) (citing, inter alia, State v. Doss, 116 Ariz. 156, 568 P.2d 1054, 1060 (1977), and writing, “Leniency is therefore required.”). Nothing in the practice of the Arizona Supreme Court suggests that when it sentenced Kayer de novo in 1999, it would have treated as less-than-binding a twenty-year-old precedent. In that precedent—Brookover—the Arizona Supreme Court had held, on facts less favorable to the defendant than those in Kayer’s case, that a non-capital sentence was “mandated.”

Kayer, 923 F.3d at 725.

IV. Summary

There are two things that differentiate this case from run- of-the-mill IAC habeas cases under AEDPA.

First, this is not the usual case in which the evidence presented in the state PCR proceeding was merely cumulative of evidence already presented at the sentencing phase, establishing more firmly an already established proposition. Instead, this is a case in which new evidence established for the first time the existence of a new and important mitigating factor. The effect of the new evidence was to change the

evidence in favor of mitigation, from one weak non-statutory mitigator (importance in the life of Kayer’s son) to two mitigators—the continuing non-statutory mitigator, plus the new statutory mitigator of mental impairment. The two mitigators must now be weighed against two existing, relatively weak aggravators.

Second, this is an unusual case in that there is a state supreme court decision in a capital case with strikingly similar facts, in which the Court held that a non-capital sentence was “mandated.” We did not hold, based on Brookover, that the Arizona Supreme Court would necessarily have held that a non-capital sentence was “mandated.” But we did hold, based on Brookover, that it was “objectively unreasonable” for the state PCR judge to conclude that there was “no reasonable probability” of a different sentence. Porter, 558 U.S. at 31.

Contrary to the contention of our dissenting colleagues, we are acutely aware of the deference required under AEDPA. Even after giving all appropriate deference to the decision of the PCR judge, we concluded that habeas relief is warranted. KAYER V. RYAN 25

BEA, Circuit Judge, joined by BYBEE, CALLAHAN, M. SMITH, IKUTA, OWENS, BENNETT, R. NELSON, BADE, COLLINS, LEE, and BRESS, Circuit Judges, dissenting from the denial of rehearing en banc:

Like clockwork, practically on a yearly basis since the Millennium, we have forced the Supreme Court to correct our inability to apply the proper legal standards under the Antiterrorism and Effective Death Penalty Act (“AEDPA”).1

1 See, e.g., Sexton v. Beaudreaux, 138 S. Ct. 2555, 2558 (2018) (per curiam) (finding “[t]he Ninth Circuit failed to [] apply” the proper standard and instead “spent most of its opinion conducting a de novo analysis”); Kernan v. Cuero, 138 S. Ct. 4, 9 (2017) (per curiam) (finding “several problems with the Ninth Circuit’s reasoning,” including that it failed to recognize that “fairminded jurists could disagree” about how to construe Supreme Court precedent); Davis v. Ayala, 135 S. Ct. 2187, 2193 (2015) (“The Ninth Circuit’s decision was based on the misapplication of basic rules regarding harmless error.”); Lopez v. Smith, 574 U.S. 1, 6 (2014) (per curiam) (criticizing “the Ninth Circuit in particular” for applying a legal standard nowhere found in AEDPA); Johnson v. Williams, 568 U.S. 289, 297 (2013) (holding that “the Ninth Circuit declined to apply the deferential standard of review” mandated by AEDPA); Cavazos v. Smith, 565 U.S. 1, 7 (2011) (per curiam) (citation omitted) (“When the deference to state court decisions required by § 2254(d) is applied to the state court’s already deferential review, there can be no doubt of the Ninth Circuit’s error below.”); Felkner v. Jackson, 562 U.S. 594, 598 (2011) (per curiam) (explaining that “[t]here was simply no basis for the Ninth Circuit” to grant habeas relief under AEDPA’s highly deferential standard, “particularly in such a dismissive manner”); Premo v. Moore, 562 U.S. 115, 123 (2011) (“The [Ninth Circuit] was wrong to accord scant deference to counsel’s judgment, and doubly wrong to conclude it would have been unreasonable to find that the defense attorney qualified as counsel for Sixth Amendment purposes.”); Harrington v. Richter, 562 U.S. 86, 92 (2011) (“[J]udicial disregard [for the sound and established principles of when to issue a writ of habeas corpus] is inherent in the opinion of the Court of Appeals for the Ninth Circuit here under review.”); Knowles v. Mirzayance, 556 U.S. 111, 121

A divided panel in this case took that tradition one step further, though, by re-writing AEDPA entirely: to institute the federal habeas court as a mere second state appellate court of state law error review.

A divided panel in this federal habeas appeal granted petitioner George Russell Kayer relief. Kayer v. Ryan, 923 F.3d 692, 726 (9th Cir. 2019). Kayer claimed that the Arizona Superior Court erred in holding, on post-conviction review (“PCR”), that the failure of Kayer’s trial counsel to conduct an adequate penalty phase investigation did not violate his Sixth Amendment right to counsel under Strickland v. Washington, 466 U.S. 668 (1984).

Kayer was sentenced to death for the first-degree, premeditated murder of an acquaintance over a minor debt that Kayer owed the victim. Kayer shot the victim in the head, stripped the victim’s body of any valuables, returned to steal the victim’s house keys, shot the victim again for good measure, ransacked the victim’s home, and pawned off the loot. Kayer’s attorneys by all accounts did little investigation

(2009) (holding the Ninth Circuit’s erroneous issuance of a writ was “based, in large measure, on its application of an improper standard of review”); Uttecht v. Brown, 551 U.S. 1, 22 (2007) (finding “[t]he Court of Appeals neglected to accord” the proper deference to the state trial court); Schriro v. Landrigan, 550 U.S. 465, 473 (2007) (“The question under AEDPA is not whether a federal court believes the state court’s determination was incorrect but whether that determination was unreasonable—a substantially higher threshold.”); Rice v. Collins, 546 U.S. 333, 342 (2006) (“[The Ninth Circuit’s] attempt to use a set of debatable inferences to set aside the conclusion reached by the state court does not satisfy AEDPA’s requirements for granting a writ of habeas corpus.”); Woodford v. Visciotti, 537 U.S. 19, 25 (2002) (per curiam) (criticizing the Ninth Circuit for “substitut[ing] its own judgment for that of the state court, in contravention of 28 U.S.C. § 2254(d)”). KAYER V. RYAN 27

for the penalty phase, and the panel majority concluded that an adequate penalty phase investigation would have uncovered evidence of Kayer’s “mental illness, and gambling and alcohol addiction.” Kayer, 923 F.3d at 727 (Owens, J., concurring in part and dissenting in part). But even assuming Kayer’s penalty-phase counsel was ineffective, the state PCR court reasonably determined that Kayer’s counsel’s failure to investigate did not prejudice Kayer. The state supreme court denied review.

Reviewing the PCR court’s decision, the panel majority cast aside (albeit with some lip service) AEDPA’s highly deferential standard of review. By any fair reading of the panel majority’s opinion, it reviewed the PCR court’s decision de novo as to whether an Arizona court, applying Arizona precedent, would have granted relief—a radical approach unwarranted under AEDPA. In short, the panel majority reasoned that because it believed there was a reasonable probability Kayer’s sentence would have been less than death if the evidence of mental impairment produced to the PCR court were presented to the sentencing court, the PCR court’s contrary finding was objectively unreasonable. Taking the panel majority at its word, it views as objectively unreasonable—and thus meritorious of a federal writ of habeas corpus—that the PCR court reached a different conclusion about prejudice than did the panel majority. That is de novo review, plain and simple. As noted, and making matters worse, the panel majority evaluated whether the state court’s no-prejudice finding adhered to Arizona’s inapplicable state law—not federal law.

Beyond the legal errors, Kayer’s proposed mitigating evidence—relating mostly to his “untreated alcoholism and untreated pathological gambling,” Kayer, 923 F.3d at 719,

and absent any findings of organic brain damage—is hardly overwhelming, and reasonable jurists could find that it did not undermine confidence in the death sentence. As such, it provides no basis for relief under AEDPA’s deferential standard. As Judge Owens convincingly observed, given the “brutal” manner in which Kayer killed the victim and the “hardly overwhelming” mitigating evidence, ample room remains for fairminded disagreement whether the failure of Kayer’s counsel to investigate prejudiced him. Id.at 726–27 (Owens, J., concurring in part and dissenting in part).

Contrary to the panel majority’s opinion, AEDPA as interpreted by the Supreme Court nowhere instructs that entitlement to federal habeas relief turns on a de novo review of whether an Arizona court in PCR proceedings adhered to Arizona precedent regarding de novo review of death penalty sentences. AEDPA instead requires that Kayer show that the Arizona Superior Court’s PCR determination was “so lacking in justification that there was an error well understood and comprehended in existing law beyond any possibility for fairminded disagreement.” Harrington, 562 U.S. at 103.

From our position, the issue is not what we think the state PCR court should have done to conform to Arizona law. The issue is whether what the state PCR court in fact did (its decision, not how it arrived at its decision) was objectively unreasonable under the standard articulated in Harrington. The Supreme Court has told us—specifically us—not to “ignore[]” that this is literally “the only question that matters.” Id. at 102 (quoting Lockyer v. Andrade, 538 U.S. 63, 71 (2003)). How the panel majority’s opinion could outright ignore (and replace) this standard is incomprehensible. We should have taken this case en banc KAYER V. RYAN 29

to correct the panel majority’s opinion’s errors before the Supreme Court (again) does it for us.

I

George Russell Kayer was convicted of first-degree murder for the death of Delbert Haas and sentenced to death in Arizona Superior Court in 1997.

The series of events that led to Kayer’s conviction all began with a gambling trip gone awry. In 1994, Kayer, his girlfriend Lisa Kester, and pal Delbert Haas hopped in Haas’s van to travel from Arizona to Nevada for a gambling trip. The three spent their first night sharing a room at a hotel in Laughlin, Nevada. Kayer that night told Haas that he had “won big” during the day using a special gambling system. This was apparently a lie, but Kayer knew Haas had recently received money from an insurance settlement, and Kayer used the lie to convince Haas to lend him about $100 in gambling money.

The next day, Kayer of course lost Haas’s money gambling. But Kayer lied to Haas again, this time fabricating a story about how he had in fact “won big” but that someone stole the winnings. In private, Kester asked Kayer what he planned to do now that he was out of cash. Kayer replied that he would rob Haas. Kester pointed out the obvious fact that Haas would identify Kayer as the thief. According to Kester, Kayer responded, “I guess I’ll just have to kill him.”

The following day, the trio drove back to Arizona, consuming a case of beer between the three during the several-hour drive. Haas and Kayer argued about Kayer’s debt. During a stop to buy snacks and use the bathroom,

Kayer pulled a gun from beneath a seat in the van and put it in his pants. Kayer asked Kester if she was “going to be all right with this.” Kester asked Kayer to warn her before he pulled the trigger.

The three continued on their way. Kayer, who was driving, left the main highway, telling his companions he was taking a shortcut. Kayer stopped the van by the side of the road, at which point Haas exited and walked toward the back to urinate. Kester went to exit the van as well, but Kayer stopped her, gesturing to the gun. Kester received her warning. Through the back window of the van, Kester saw Kayer walk up behind Haas and—as Kayer had planned to do—shoot Haas in the head while he was urinating.

Kayer dragged Haas’s body into the bushes; took Haas’s wallet, watch and jewelry; got back in the van; and drove away with Kester. Back on the road, Kayer realized he forgot to take Haas’s house keys and drove back to where he dumped the body. Kayer exited the van to retrieve the keys from Haas’s body, but then returned and asked for the gun. Haas, Kayer said, did not appear to be dead. Kayer went back to Haas’s body, and Kester heard a second shot. Kayer and Kester then drove to Haas’s Arizona home and looted it. They spent the next week pawning and selling items from Haas’s home and gambling with the proceeds.

Ten days after the murder, Kester got cold feet and approached a security guard at a Las Vegas hotel to report Kayer’s murder of Haas. Kayer was indicted for, and eventually convicted of, first-degree murder. During the penalty phase hearing, Kayer’s counsel argued as mitigating circumstances that Kayer suffered from mental illness and substance abuse. But Kayer’s counsel adduced virtually no KAYER V. RYAN 31

evidence to support that argument. The judge2 held that Kayer had not established any mental impairment and sentenced him to death.

On direct appeal, the Arizona Supreme Court independently reviewed and affirmed Kayer’s death sentence. In Arizona, mitigating evidence can serve either as a statutory or non-statutory mitigating factor, with greater weight due to statutory factors. The Court refused to find a mitigating circumstance based on mental impairment, either as a statutory or non-statutory factor. The Court did find one non- statutory mitigating circumstance (Kayer’s importance in his son’s life), but held it was outweighed by two statutory aggravating circumstances—a previous conviction of a “serious offense” for first-degree burglary in 1981 and “pecuniary gain” as motivation for the murder.

On state habeas review, Kayer argued that his trial counsel provided ineffective assistance of counsel at the penalty phase. Kayer presented evidence in his PCR proceeding that his trial counsel performed little investigation of mitigating circumstances. Had counsel properly investigated, Kayer argued, trial counsel would have discovered that: Kayer suffered from bipolar disorder and “personality disorders”; Kayer had “a family history of problems with alcohol, gambling and bipolar disorder that increased his risk of developing one or more of these disorders”; Kayer’s father died when he was young, resulting in “significant instability including frequent moves”; Kayer’s “performance in school was not good”; Kayer was “having difficulties with out of control pathological gambling” and “extensive alcohol abuse” at the time of the murder; and, to

2 This ruling pre-dated Ring v. Arizona, 536 U.S. 584 (2002).

top it all off, Kayer had suffered a heart attack weeks before the murder, an “important source of emotional distress that was likely exacerbating all his other problems.” Kayer, 923 F.3d at 713. The PCR court denied relief, holding that Kayer’s counsel had not been ineffective, and that, in any event, any deficiencies did not prejudice Kayer. The Arizona Supreme Court denied review without comment.

Kayer then sought federal habeas relief. The district court denied relief. Kayer appealed, contending, as relevant here, that the Arizona PCR court erred in holding that his Sixth Amendment right to counsel was not violated by his counsel’s deficient performance at the penalty phase. A divided panel reversed, holding that no reasonable jurist could conclude that Kayer’s counsel had rendered effective representation by failing to conduct a mitigation investigation in preparation for the penalty phase, and that counsel’s failure to investigate prejudiced Kayer.3 Judge Owens dissented as to whether counsel’s failure prejudiced Kayer. A reasonable jurist could find the purported mitigation evidence would not have made a difference, Judge Owens reasoned, given the “brutal” manner in which Kayer killed Haas and the “hardly overwhelming” mitigating evidence. Id. at 726–27 (Owens, J., concurring in part and dissenting in part).

3 Apparently, the following are thus unreasonable jurists: Arizona Superior Court Judge William T. Kiger, the state PCR court judge; the five members of the Arizona Supreme Court who denied Kayer’s petition for review of Judge Kiger’s PCR decision; U.S. District Court Judge David G. Campbell, who denied federal habeas relief; Judge Owens; and me. KAYER V. RYAN 33

II

The Court of Appeals reviews the district court’s denial of a 28 U.S.C. § 2254 habeas corpus petition de novo. Deck v. Jenkins, 814 F.3d 954, 977 (9th Cir. 2016).

Because Kayer’s state conviction was entered after April 24, 1996, Kayer’s habeas petition is subject to AEDPA, under which “[w]e review the last reasoned state court opinion.” Musladin v. Lamarque, 555 F.3d 830, 834 (9th Cir. 2009). In this case, that opinion is the written order of the state PCR court.

AEDPA bars relitigation of any claim the state court decided on the merits unless the state court’s determination was “contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States.” 28 U.S.C. § 2254(d)(1). A state court unreasonably applies Supreme Court precedent “if the state court identifies the correct governing legal principle from the Supreme Court’s decisions but unreasonably applies that principle to the facts of the prisoner’s case.” Mann v. Ryan, 828 F.3d 1143, 1151 (9th Cir. 2016) (en banc) (alteration omitted) (quoting Williams v. Taylor, 529 U.S. 362, 413 (2000)).

AEDPA’s standard is “highly deferential” and “difficult to meet.” Harrington, 562 U.S. at 102, 105 (citations omitted). To meet it, a petitioner must show that the state court’s decision was “so lacking in justification that there was an error well understood and comprehended in existing law beyond any possibility for fairminded disagreement.” Id. at 103. According to the Supreme Court, this is literally “the only question that matters.” Id. at 102 (quoting Lockyer, 538

U.S. at 71). In other words, AEDPA “demands that state- court decisions be given the benefit of the doubt.” Visciotti, 537 U.S. at 24.

III

Kayer contends that the state PCR court’s review of his Sixth Amendment claim involved an unreasonable application of the Supreme Court’s decision in Strickland. See 28 U.S.C. § 2254(d)(1). Specifically, Kayer contends that he was denied his Sixth Amendment right to effective assistance of counsel due to his attorneys’ inadequate mitigation investigation in preparation for his penalty phase hearing. See Wiggins v. Smith, 539 U.S. 510, 521–22 (2003).

For the sake of argument, I assume the panel rightly concluded that Kayer’s attorneys performed deficiently by failing to conduct an adequate penalty phase investigation.4 But a habeas petitioner must establish both deficient performance and “a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different.” Strickland, 466 U.S. at 694. No matter how inadequate Kayer’s attorneys may have been, deficient performance alone is not enough to merit relief. And “[a] reasonable probability is a probability sufficient to undermine confidence in the [verdict].” Id. As is critical here, after a state PCR court finds there was no reasonable probability the result would have been different but for counsel’s unprofessional errors, that finding must stand

4 Judge Owens appears to have concurred in this conclusion. See Kayer, 923 F.3d at 727 (Owens, J., concurring in part and dissenting in part). KAYER V. RYAN 35

unless it was “objectively unreasonable.” Porter v. McCollum, 558 U.S. 30, 31 (2009) (per curiam).

IV

The discussion that follows proceeds in three steps. First, I relate the mitigating evidence Kayer presented to the PCR court. Second, I explain why the panel majority’s review of the PCR court’s opinion applied a de novo review that flouts AEDPA’s highly deferential standards. Third, I explain why the mitigation evidence was not “sufficient to undermine confidence in the [verdict],” Strickland, 466 U.S. at 694—certainly not “beyond any possibility for fairminded disagreement,” Harrington, 562 U.S. at 103.

A

Personal Background. According to some, Kayer was “slow to walk” and “slow at all developmental stages.” Kayer was dyslexic and struggled in school. His high school transcript is smattered with Cs, Ds, and Fs, though, also, with the occasional Bs (in “Speech,” “Typing,” and “Drafting”). Kayer left high school without graduating and enlisted in the Navy. There, he had two “unauthorized absences” (“UAs”) in eight months. He returned from his second UA “to see a psychiatrist.” At Bethesda Naval Hospital, in May 1973 and at the age of eighteen, Kayer was diagnosed with a “passive- aggressive personality.” On discharge, his Lieutenant Commander noted that he had a “sever[e] . . . personality disorder.”

In his twenties, Kayer didn’t fare much better. He bounced around Arizona state colleges but never graduated. Kayer also “never held a job for a sustained period.”

Throughout his twenties and thirties, he was a serial burglar (arrested twice). He married twice in his early twenties, but both ended in divorce. He then met Cindy Seitzberg, with whom he fathered a son. His son was dropped in the delivery room and suffered permanent brain damage. Seitzberg left a year later, and Kayer’s half-sister and mother became his infant son’s co-guardians.

Addictive Behavior. Throughout this period, Kayer “smoke[d] weed almost every day” (beginning at sixteen),5 drank regularly (beginning in his early twenties), and became a compulsive gambler (“[s]ometime in his twenties”). Kayer, 923 F.3d at 709–10. As to his alcohol abuse, a former accomplice to his burglaries recounted that together they would drink beer “for breakfast, lunch and dinner.” When Kayer voluntarily checked himself into a Veterans Administration (“VA”) hospital at the age of thirty-five, the observing doctor reported that Kayer “had been drinking continuously and heavily for the past seven years.” As to his gambling, Kayer’s half-sister described his obsession with a personal gambling “system” and countless trips to Las Vegas—including, once, while on house arrest after release for a burglary conviction. That time, Kayer turned himself in after he lost all his money and was sentenced to an additional nineteen months for violating parole.

Family History. Kayer came from a family of unsavory characters. His father was an alcoholic and compulsive gambler who left the family when Kayer was two and died at age thirty-nine of a heart attack. On his mother’s side, one of

5 The panel majority opinion contains passing references to Kayer “us[ing] speed on the weekends” and using LSD “sometimes” in his late teens. Kayer, 923 F.3d at 709. KAYER V. RYAN 37

his aunts was “an alcoholic with severe mood swings,” and another was an alcoholic who was “severely depressed.” One of his cousins—diagnosed first as schizophrenic and then as bipolar—“blew” her “entire retirement” in a single weekend in Vegas. And that’s not to mention his Uncle John: “a thief, a robber,” who “held his own family members at gunpoint and knifepoint a few times.” Uncle John, one of Kayer’s aunts testified, “hit his head in a creek in Oklahoma and he just never did do too good after that.”

Despite the cast of characters circling Kayer’s youth, there is no evidence Kayer was ever the subject of abuse, either by beatings or sexual molestation. Nor is there any evidence Kayer suffered organic brain damage from an accident or some traumatic childhood event.

Mental Health. Kayer’s mother and sister testified before the PCR court that Kayer experienced “severe mood swings”—for example, proposing to take a trip “out-of-the blue” when “it wasn’t a good time,” and “either work[ing] [at something] all out, or do[ing] nothing.”

In 1983 and at the age of twenty-nine, Kayer voluntarily went to a VA hospital. The doctor diagnosed him with an “adjustment disorder with depressed mood.” Six years later, he was “admitted . . . with depression and suicidal ideation” after his then-girlfriend left him. Kayer was kept at the hospital for eighteen days. The observing doctor wrote that Kayer “showed bipolar traits,” but was not “considered to be a danger to himself or others” at the time of discharge. A year later, he was referred to a VA “Day Treatment Center” with a “provisional diagnosis” of “Personality Disorder/Bipolar.”

Some evidence suggests Kayer held delusional beliefs and heard voices. In an interview with a private investigator for one of Kayer’s mitigation experts, Kayer stated that he came to believe at age seven—and continues to believe—that he came to earth from another planet. He also maintained, in the same interview, that the uncle of Kayer’s second wife—an Afghan woman—was “the deposed king of Afghanistan.” Finally, one of Kayer’s aunts testified in the PCR court that she has heard voices her entire life, and that Kayer too heard voices: “I was just telling him about my life and he said ‘I thought it was normal[.] I hear voices, too.’”

Professional Assessments. Three experts evaluated Kayer in Arizona prison, after Kayer’s murder conviction. First, Dr. Anne Herring met with Kayer in March 2005 (over a decade after the murder), and administered a battery of tests. Dr. Herring testified before the PCR court that Kayer received average scores on all tests except for “one of the more cognitively challenging” ones due to Kayer’s “persist[ence] in applying incorrect concepts despite receiving feedback.” Dr. Herring suggested that applying such “incorrect concepts” is similar to deficits that “have been associated with chronic heavy substance abuse, traumatic brain injury, and with bipolar disorder.” 6

Next, Dr. Michael Sucher, a specialist in “alcohol and drug addiction medicine,” met with Kayer in April 2005. Dr. Sucher’s notes reflect that Kayer spent “probably one-quarter to one-third” of his interview discussing his gambling “system.” Dr. Sucher testified that Kayer had “untreated

6 But again, Dr. Herring did not cite to any record evidence, nor did anything in the record reflect, that Kayer ever experienced traumatic brain injury. KAYER V. RYAN 39

alcoholism and untreated pathological gambling.” And he gave this incisive take: Gambling and drinking “often make individuals who are so impaired do things that they would not normally do.”

Finally, Dr. Barry Morenz twice interviewed Kayer in March and April 2005 for a total of five and a half hours and reviewed Kayer’s medical records. Dr. Morenz wrote in his subsequent report that Kayer spent much of the interview talking about his system for predicting winning lottery numbers. Kayer explained that he believed in reincarnation and that there is “residue in him from when Mars was populated and perhaps populations from other worlds as well.” Dr. Morenz characterized Kayer as “really delusional,” and ultimately diagnosed Kayer with “Bipolar type I disorder, hypomaniac; Alcohol dependence in a controlled environment; Polysubstance abuse in a controlled environment; Pathological gambling; Cognitive disorder not otherwise specified.” And Dr. Morenz purported to offer an account of Kayer’s conditions in 1994 that tied together the various strands of evidence discussed above:

There are a number of factors that have increased the risk of Mr. Kayer developing a number of psychiatric problems. First, there is considerable comorbidity among psychiatric diagnoses. . . . In Mr. Kayer this is relevant because people with bipolar disorders and personality disorders are at an increased risk of developing substance abuse disorders. Also, people with personality disorders have an increased risk of mood disorders. Secondly, Mr. Kayer had a family history of problems with alcohol, gambling

and bipolar disorder that increased his risk of developing one or more of these disorders. Thirdly, as a child Mr. Kayer grew up with significant instability including frequent moves and his father’s sudden death when Mr. Kayer was still very young which probably contributed to his later psychiatric difficulties. There is evidence that even as a child Mr. Kayer was showing signs of emotional problems as his performance in school was not good. This poor school performance was probably an early sign of a bipolar disorder or a personality disorder or a combination of the two. By the time Mr. Kayer washed out of the military Mr. Kayer likely had moderately severe psychiatric problems that went untreated. . . . [I]t seems clear that he has suffered from serious psychiatric problems during most of his adult life and he continues to show signs of those problems today. . . .

At the time of the murder in 1994 Mr. Kayer was probably having serious psychiatric problems. He was having problems with bipolar disorder symptoms and may have been manic or hypomanic, he was having difficulties with out of control pathological gambling and he had difficulty with extensive alcohol abuse. These difficulties were likely superimposed on his personality disorder problems and his cognitive disorder not otherwise specified. Mr. Kayer’s belief that he would not live long as a result of the heart KAYER V. RYAN 41

attack he had suffered a few weeks before the murder was another important source of emotional distress that was likely exacerbating all his other problems during this period.

All this evidence was before the state PCR court, which concluded that Kayer had not been prejudiced by his trial counsel’s failure to introduce this evidence in mitigation and before sentencing. He should know, because in Arizona the same judge who presides over a defendant’s trial and sentencing also presides over the PCR proceeding. After first noting that counsel had performed adequately, the PCR court noted: “This court further concludes that if there had been a finding that the performance prong of the Strickland standard had been met, that no prejudice to the defendant can be found.” Kayer, 923 F.3d at 714 (quoting PCR court order).

B

As detailed below, there is no ignoring the obvious conclusion that a reasonable jurist could conclude that Kayer was not in fact prejudiced by his counsel’s failings in this case, but broader legal errors permeate the panel majority’s opinion, which counseled en banc correction.

As discussed above, but it bears repeating, AEDPA’s standard of review is “highly deferential” and “difficult to meet.” Harrington, 562 U.S. at 102, 105 (citations omitted). Here, we must apply that strong deference and decide whether “it was objectively unreasonable [for the state PCR court] to conclude there was no reasonable probability the sentence would have been different if the sentencing judge” heard the mitigation evidence Kayer’s counsel presented to

the PCR court. Porter, 558 U.S. at 31. Whether the PCR court’s no-prejudice conclusion was objectively unreasonable depends on whether it was “so lacking in justification that there was an error well understood and comprehended in existing law beyond any possibility for fairminded disagreement.” Harrington, 562 U.S. at 103.

What standard did the panel majority apply? The panel majority stated that in its own view (affording no deference whatsoever), “the evidence [Kayer] presented to the PCR court was sufficient to establish [an Arizona] statutory mitigating circumstance” of mental impairment. Kayer, 923 F.3d at 718. Then (again, in its own view), the panel majority posited that the sentencing court “would have added” the Arizona statutory mitigating mental impairment to the balance of aggravating and mitigating factors, and the addition “could have changed the outcome of the sentencing proceeding.” Id. at 718–20. Thus, in the majority’s independent judgment, it pronounced “that there is a reasonable probability Kayer’s sentence would have been less than death” if the sentencing judge heard the mitigation evidence. Id. at 723. And because the panel majority believed there was such a probability, that alone meant “that the state PCR court was unreasonable in concluding otherwise.” Id. Again: “unreasonable in concluding otherwise.” Id.

The panel majority’s opinion is de novo review, plain and simple. Nothing in its review of the state PCR court’s decision included any deference whatsoever, particularly the high deference mandated by AEDPA. And the dearth of deference is particularly unnerving here because who better to determine whether the new evidence would have made a difference at sentencing than the judge who sentenced Kayer KAYER V. RYAN 43

to death. Judge Kiger presided over both sentencing and the PCR proceedings, and he concluded the new evidence would have made no difference. His “unique knowledge of the trial court proceedings”—including his front-row seat to the presentation of evidence showing Kayer’s brutal and venal murder—“render[ed] him ‘ideally situated’” to evaluate Kayer’s claim that the introduction of new evidence would have changed the sentencing outcome. Murray v. Schriro, 882 F.3d 778, 821 (9th Cir. 2018) (quoting Landrigan, 550 U.S. at 476).7 This is not to say that Judge Kiger is entitled to some sort of super-deference simply because he sentenced Kayer to death. But there is something particularly troubling about the panel majority affording no deference whatsoever to Judge Kiger’s PCR court decision, as the last reasoned state court opinion.

In failing to accord the state PCR court decision any deference whatsoever, the panel majority committed two errors. First, the majority contended the PCR court is entitled to no “special” deference—really, no deference at all— because Strickland demands that PCR courts assess whether there is a reasonable possibility that “the sentencer— including an appellate court, to the extent it independently reweighs the evidence—would have concluded that the balance of the aggravating and mitigating factors did not warrant death.” Kayer, 923 F.3d at 720 (quoting Strickland, 466 U.S. at 695). Put differently, the panel majority did not defer to Judge Kiger’s analysis of whether there is a reasonable possibility that new evidence would have resulted

7 Apparently, the panel majority here is not alone in improperly second-guessing state court judges in this regard recently. See Washington v. Ryan, 922 F.3d 419, 433–34 (9th Cir. 2019) (Callahan, J., dissenting).

in a sentence less than death because Judge Kiger in his PCR role must conduct this analysis in an objective and independent manner. And Judge Kiger in his PCR role must objectively consider what an independent reviewing court might think. But just because Judge Kiger in his PCR role here was required to “consider the probability of a different outcome in the Arizona Supreme Court,” White v. Ryan, 895 F.3d 641, 671 (9th Cir. 2018), does not mean Judge Kiger merits no deference at all in gauging whether the new evidence would have made a difference.

The panel majority’s second, and more stunning error concerns its discussion of the probability of a different outcome in the Arizona Supreme Court. After just saying that “we assess prejudice independent of the particular judge or judges” to take away deference to Judge Kiger, the panel majority then proposes that the yardstick for whether there is a reasonable probability Kayer would not have been sentenced to death if the new evidence were presented to the sentencing court is whether this case is more like cases in which the Arizona Supreme Court at one point affirmed a death penalty imposed by the trial court on direct de novo review or more like cases in which the Arizona Supreme Court reversed. Kayer, 923 F.3d at 721–23. Purportedly because Kayer’s case looks more like the reversals—and in particular, State v. Brookover, 601 P.2d 1322 (Ariz. 1979) (in banc)—the panel majority concludes that Kayer established a reasonable probability of a different outcome.

The panel majority does not explain why it is appropriate for AEDPA review to turn on Arizona Supreme Court reversal trends on de novo review of direct appeals. Just because the Arizona Supreme Court in Brookover reversed a case with some similarities, but also with a glaring KAYER V. RYAN 45

dissimilarity (as discussed further below), forty years ago on de novo review does not mean the PCR court’s conclusion that in this case the new evidence would not have made a difference is objectively unreasonable and beyond room for fairminded disagreement. Harrington, 562 U.S. at 103.

Admittedly, the panel majority’s resort to the “best evidence” of what the Arizona Supreme Court would have done—its decisions—has a certain first-blush plausibility. But by holding that the Arizona courts were objectively unreasonable in failing to adopt this court’s analysis of Arizona law, the panel majority employs a mode of habeas review of a Strickland claim that is quite literally unprecedented. The panel majority’s defense of its approach cites not a single authority for the proposition that the measure for federal habeas review of a state PCR court’s Strickland-prejudice conclusion may be evaluated by “look[ing] at de novo sentencing decisions by the [state] Supreme Court in comparable cases.” Kayer, 923 F.3d at 724.

The rule is as misguided as it is novel. For starters, the panel majority’s approach would make federal habeas review of every Strickland claim turn on the state in which the petitioner was sentenced. So U.S. Supreme Court habeas precedents that involve California apparently could be distinguished away in habeas appeals from Arizona, on the sole ground that “we ask what an Arizona rather than a California sentencing court would have done.” Id. The panel majority appears untroubled by this point, but its implications are striking: Their approach—at least for Strickland prejudice—transmutes “clearly established Federal law, as determined by the Supreme Court of the United States” into

law as determined by state supreme courts. 28 U.S.C. § 2254(d)(1).

The panel majority’s approach is also impossible to square with Judge W. Fletcher’s earlier en banc majority opinion in McKinney v. Ryan, 813 F.3d 798 (9th Cir. 2015) (en banc). There, McKinney was sentenced to death and his sentence was affirmed by the Arizona Supreme Court. Id. at 802. McKinney filed a federal habeas petition, challenging in relevant part his death sentence, because the state courts—following Arizona Supreme Court precedent— forbade consideration of certain mitigating evidence unless it bore a “causal nexus to his crimes.” Id. at 803. And the en banc panel granted McKinney’s writ with respect to his sentence because Arizona’s “causal nexus” requirement was contrary to clearly established federal law established in Eddings v. Oklahoma, 455 U.S. 104, 114 (1982), which held that a sentencer in a capital case may not refuse to consider as a matter of law relevant mitigating evidence. Id. at 810.

Unlike here, the en banc majority in McKinney not once queried whether the Arizona courts’ analysis of Arizona law was based on an objectively reasonable reading of Arizona precedent. Because unlike here, the en banc majority in McKinney cited the proper standard of review set forth in Harrington, which is that federal courts may grant relief under AEDPA “only if the state court’s application of clearly established federal law was objectively unreasonable, such that fairminded jurists could not disagree that the arguments or theories that supported the state court’s decision were inconsistent with the holding in a prior decision of the Supreme Court.” Id. at 811 (internal quotation marks, citations, and brackets omitted). KAYER V. RYAN 47

How then did the panel majority here rationalize tethering the propriety of the state PCR court’s no-prejudice finding to a comparison of the facts with “de novo sentencing decisions by the Arizona Supreme Court in comparable cases”? Kayer, 923 F.3d at 724. In short, it did not. Nor could it, unless the panel majority meant to create a rule that under AEDPA, adherence to state precedent is relevant (and mandatory) when it leads to relief, but not when it leads to denial of relief. See discussion supra of McKinney. It should go without saying that AEDPA condones no such a rule. Rather, federal courts should be concerned only with what the Supreme Court has repeatedly instructed is “the only question that matters”: whether “there is no possibility fairminded jurists could disagree” that the Arizona decision itself conflicts with federal precedent. Harrington, 562 U.S. at 102.

At a more fundamental level, the panel majority’s approach is deeply anathema to AEDPA’s basic purpose. “Section 2254(d) reflects the view that habeas corpus is a ‘guard against extreme malfunctions in the state criminal justice systems,’ not a substitute for ordinary error correction through appeal.” Id. at 102–03 (quoting Jackson v. Virginia, 443 U.S. 307, 332 n.5 (1979) (Stevens, J., concurring in judgment)). It is hard to see here how the panel majority’s analysis differs at all from de novo review—indeed, a de novo sentencing direct appeal analysis the Arizona Supreme Court itself has already done in denying relief to Kayer.

It cannot be stressed enough just what the panel did wrong. Rather than ask whether fairminded jurists could disagree about whether the new mitigation evidence was sufficient to undermine confidence in the outcome, the panel majority began its inquiry by asking from scratch: “[W]as the

mitigation evidence that was presented to the PCR court sufficient to establish a ‘reasonable probability,’ ‘sufficient to undermine confidence in the outcome,’ that the result of the sentencing hearing would have been different?” Kayer, 923 F.3d at 716. Looking to Arizona Supreme Court de novo sentencing appeals, the panel majority then concluded that those cases indeed “show[] that there is a reasonable probability that Kayer would not have been sentenced to death if the mitigating evidence presented to the PCR court had been presented to the sentencing court.” Id. at 721. The panel majority therefore held “there is a reasonable probability Kayer’s sentence would have been less than death, and”—almost as an afterthought—“that the state PCR court was unreasonable in concluding otherwise.” Id. at 723. (Here, of course, neglecting to mention that the standard is whether the state PCR court was “objectively” unreasonable: that its decision was one “beyond any possibility for fairminded disagreement,” the “only question that matters.” Harrington, 562 U.S. at 102–03.)

The Supreme Court has repeatedly condemned this de- novo-masquerading-as-deference approach. In Harrington, for example, the Court chastised the Ninth Circuit for “treat[ing] the unreasonableness question as a test of its confidence in the result it would reach under de novo review.” 562 U.S. at 102. But that is precisely what the panel majority did here. And if there were any remote doubt about the impropriety of panel majority’s analysis, one need only ask themselves if the following sounds familiar:

Here it is not apparent how the Court of Appeals’ analysis would have been any different without AEDPA. The court explicitly conducted a de novo review; and KAYER V. RYAN 49

after finding a Strickland violation, it declared, without further explanation, that the state court’s decision to the contrary constituted an unreasonable application of Strickland. AEDPA demands more. Under § 2254(d), a habeas court must determine what arguments or theories supported or, as here, could have supported, the state court’s decision; and then it must ask whether it is possible fairminded jurists could disagree that those arguments or theories are inconsistent with the holding in a prior decision of this Court. The opinion of the Court of Appeals all but ignored the only question that matters under § 2254(d)(1).

Id. at 101–02 (internal quotation marks and citations omitted).

The panel majority’s approach evinces no deference: In both the panel majority’s actual analysis and an Arizona Supreme Court de novo sentencing direct review, the inquiry begins and ends with “de novo sentencing decisions by the Arizona Supreme Court in comparable cases.” Kayer, 923 F.3d at 724. It should go without saying that AEDPA does not authorize Article III judges to role-play as super state Supreme Court justices.8

8 As should be understood, the proper approach instead requires federal courts to “presum[e] that state courts know and follow the law.” Visciotti, 537 U.S. at 24. Federal courts must determine what arguments supported or could have supported the state court’s decision, and ask “whether it is possible fairminded jurists could disagree that those arguments or theories are inconsistent with the holding in a prior decision” of the U.S. Supreme Court. Harrington, 562 U.S. at 102.

C

As just discussed, the panel majority mangled the law in reviewing de novo a state court decision and making out of whole cloth a method of review that requires idiosyncratically comparing a given case’s facts to past state supreme court cases engaged in their own de novo review. If these errors were not reason enough to take the case en banc, the panel majority’s conclusions are clearly unwarranted under the proper AEDPA framework.

Assuming that all of the mitigating evidence Kayer presented in the PCR proceeding (and summarized above) would have been introduced to the trial court, is it possible that fairminded jurists could find that evidence insufficient to establish a reasonable probability of a different outcome? See Harrington, 562 U.S. at 102. Based on the above description of the evidence alone, one would think the answer obvious. As Judge Owens notes, Kayer’s crime was “brutal” and the mitigation evidence “hardly overwhelming.” Kayer, 923 F.3d at 726–27 (Owens, J., concurring in part and dissenting in part). What is more, Kayer deliberated about committing the brutal, venal crime before (twice) pulling the trigger.

Kayer killed Haas for a few hundred dollars’ worth of cash and other items. He did so in a premeditated fashion, telling his girlfriend that, unable to repay his minor gambling debt to Haas, “I guess I’ll just have to kill him.” Kayer then deliberately took Haas to a remote location, shot Haas in the head at point-blank range, and stripped Haas’s body of valuables. But even that wasn’t enough, apparently. Once on the road, Kayer turned around to retrieve Haas’s house keys from the body, to loot Haas’s home. After shooting Haas again for good measure, Kayer took Haas’s keys and in an act KAYER V. RYAN 51

of increased venality, ransacked Haas’s home—before spending the next week gambling and pawning off the loot. And while Kayer’s girlfriend later thought better of what happened and turned herself in, Kayer never looked back.

In the face of that brutal crime, “assuring the court that genetics made him the way he is could not have been very helpful.” Landrigan, 550 U.S. at 481. And that is indeed what Kayer’s mitigation evidence amounts to—“mental illness, and gambling and alcohol addiction.” Kayer, 923 F.3d at 727 (Owens, J., concurring in part and dissenting in part). There is no evidence of childhood “severe privation and abuse,” or “physical torment,” or sexual molestation—no broken bones, concussions, hospitalization, or any kind of serious or lasting injury from childhood abuse. Cf. Wiggins, 539 U.S. at 512; Williams, 529 U.S. at 370. Despite Dr. Herring writing in her report that Kayer’s deficits were akin to those associated with traumatic brain injury, nothing in the record indicates that Kayer ever actually experienced any traumatic brain injury.

Although Kayer started smoking weed at sixteen and drinking in his twenties, his family did not appear to have introduced him to alcohol or drugs. And he was not hospitalized for drinking when he was very young. Cf. Williams, 529 U.S. at 395 n.19. Notwithstanding descriptions of Kayer as “slow,” there is no evidence Kayer suffered organic brain damage, or has an IQ in the modern-day equivalent of what was previously termed the “mentally retarded range.” Cf. Rompilla v. Beard, 545 U.S. 374, 393 (2005); Wiggins, 539 U.S. at 535. Indeed, Dr. Herring testified that Kayer’s results were “average” on all but one in an extensive battery of psychiatric tests. On direct appeal, Kayer even cited his “relatively high intelligence” as a

mitigating factor. Arizona v. Kayer, 984 P.2d 31, 48 (Ariz. 1999) (en banc) (emphasis added).

Taking at face value Kayer’s evidence, “[w]hether, and to what degree, [it] is mitigating is highly debatable.” Pinholster v. Ayers, 590 F.3d 651, 715 (9th Cir. 2009) (Kozinski, J., dissenting), rev’d sub nom. Cullen v. Pinholster, 563 U.S. 170 (2011). For instance, one ordinarily might think that evidence the defendant drinks and gambles to excess would cast his character in a particularly unfavorable light. Yet because “Kayer had been drinking heavily on the day of the killing, and Kayer killed the victim in order to obtain funds to continue gambling,” the panel majority suggested, his alcoholism and pathological gambling are especially mitigating. Kayer, 923 F.3d at 721. The panel majority of course ignored that Kayer first planned to kill Haas before the day of the killing and thus had time to change his mind. This was a planned and brutal murder. Only a jurist caught up in the throes of an “infatuation with ‘humanizing’ the defendant” could take seriously the panel majority’s conclusions to the contrary. Pinholster, 590 F.3d at 692 (Kozinski, J., dissenting).

Making matters worse, although the panel majority’s unprecedented new standard of review that requires comparing cases to state supreme court reversals is unwarranted as a matter of law, the panel majority’s analysis further errs by “overlook[ing] arguments that would otherwise justify the state court’s result.” Harrington, 562 U.S. at 102. The panel majority relies heavily on Brookover as its Arizona Supreme Court de-novo-reversal analogue. There, defendant Brookover agreed to buy 750 pounds of marijuana from the victim. Upon delivery, Brookover shot KAYER V. RYAN 53

the victim to avoid paying for it, and shot him once more in the back when the victim fell to the floor.

There, as in Kayer’s case, the Arizona Supreme Court did not find a statutory aggravating factor that the murder had been committed in “an especially heinous, cruel, or depraved manner.” Like Kayer, Brookover had been previously convicted of a serious crime, and he committed the crime for pecuniary gain. “The one mitigating circumstance was mental impairment.” See Kayer, 923 F.3d at 722. The Arizona Supreme Court set aside the death sentence imposed by the trial court.

“The only difference,” says the panel majority, between Brookover and Kayer’s case is “that one of the statutory aggravators was stronger in Brookover”: Although the Brookover opinion does not describe the prior conviction, the statutory aggravator at the time required a crime “for which the death penalty or life imprisonment could be imposed.” Id. at 723–24. (Arizona later changed the statutory aggravator to a prior conviction for a “serious crime”). Neither death nor life imprisonment could be imposed for first-degree burglary, Kayer’s prior conviction. Thus, the panel majority reasoned, the reversal of Brookover’s sentence on worse facts compels reversal of Kayer’s sentence on better facts.

But the panel majority’s mechanistic weighing of the “statutory” and “non-statutory” mitigators elided obvious distinctions between the cases. Just to take one: Brookover’s mental impairment involved an organic brain injury: a “pre existing” “neurological lesion” associated with serious “anti social” behavior. Brookover, 601 P.2d at 1325. Kayer’s claimed mitigating evidence was merely self-administered “untreated alcoholism and untreated pathological gambling.”

This is no trivial distinction. In countless cases finding Strickland prejudice on federal habeas review for failing to investigate at the penalty phase, the Supreme Court has found particularly sympathetic claims of “organic brain damage” and mental retardation, see Rompilla, 545 U.S at 392; “brain damage” and “brain abnormality,” see Porter, 558 U.S. at 36; and “frontal lobe brain damage,” with “bottom first percentile” cognitive functioning, see Sears v. Upton, 561 U.S. 945, 946 (2010). Kayer, by contrast, has adduced no evidence of such injury or functioning. A fairminded jurist could reasonably distinguish Brookover on that ground alone.

Consider as well that (at least in the panel majority’s view) Kayer’s mental impairment was so significant because he was at least provisionally diagnosed bipolar, which purportedly may have manifested “manic or hypomanic” symptoms at the time of the murder. Kayer, 923 F.3d at 713. But relying on Kayer’s bipolar disorder is not so simple. As the panel majority noted, Kayer at one point was prescribed lithium, the standard drug to treat the disorder. Id. at 719. Evidence in the record, however, reflects that Kayer intentionally refused to take the medication. See id. at 697. Why couldn’t a fairminded jurist review that evidence and reason that to the extent Kayer’s bipolar disorder in some sense influenced Kayer’s decision to murder his friend, Kayer should still be held responsible for his conduct because he refused to medicate?

Finally, as Judge Owens’s dissent carefully explains, Visciotti, 537 U.S. at 24, puts any lingering doubt to rest. Kayer, 923 F.3d at 726–27 (Owens, J., concurring in part and dissenting in part). “There, in a preplanned armed robbery, the defendant and his co-worker shot two co-workers as they all drove to a party and made a remote bathroom stop (one KAYER V. RYAN 55

victim died and one survived).” Id. The defendant was sentenced to death and, “[a]t the PCR stage, the California Supreme Court determined that the defendant had not been prejudiced by his counsel’s failure to introduce mitigating evidence” of Visciotti’s brain damage and impulse disorder. Id. at 727.

The California Supreme Court held that the mitigating evidence was outweighed by “the circumstances of the crime (a cold-blooded execution-style killing of one victim and attempted execution-style killing of another, both during the course of a preplanned armed robbery) coupled with the aggravating evidence of prior offenses (the knifing of one man, and the stabbing of a pregnant woman as she lay in bed trying to protect her unborn baby).” Visciotti, 537 U.S. at 26. The Ninth Circuit granted habeas relief. But the Supreme Court summarily reversed, faulting the Ninth Circuit for impermissibly “substitut[ing] its own judgment for that of the state court.” Id. at 25. The Supreme Court had to remind the Ninth Circuit that “[a]n unreasonable application of federal law is different from an incorrect application of federal law.” Id. (internal quotation marks omitted).

The panel majority freely admitted that the facts in Visciotti “are similar” to Kayer’s, though it halfheartedly distinguished Visciotti as a California (not Arizona) case. Kayer, 923 F.3d at 724. This admission alone should have decided this case. It strains credulity to suggest that Kayer’s case marks an “extreme malfunction[] in the state criminal justice system[],” Harrington, 562 U.S. at 102, when the facts of his case admittedly resemble so starkly the facts of a case in which the Ninth Circuit was reversed (summarily) by the Supreme Court.

Again, that this case is similar to Visciotti should be enough by itself to demonstrate that fairminded jurists could disagree whether the mitigation evidence would have outweighed the aggravating evidence.

Perhaps the most telling indication of the panel majority’s error in review by case-comparison is that the panel majority rests its holding on an Arizona Supreme Court de novo review reversal bearing “striking” parallels to this case. Kayer, 923 F.3d at 722. To reach the opposite result, Judge Owens too cites a precedent with “remarkably similar” facts. Id. at 726 (Owens, J., concurring in part and dissenting in part). The difference? The panel majority’s is a decades-old Arizona Supreme Court de novo review of a sentencing decision involving a defendant with organic brain damage, a critical fact. Judge Owens’s is a Supreme Court summary reversal of a Ninth Circuit habeas grant for one of errors the majority now repeats. I’ll side with Judge Owens.

V

No one disputes Kayer has lived an unfortunate life. But sympathy alone is not a basis to cast aside AEDPA in favor of a novel de-novo-masquerading-as-deference approach never sanctioned by the Supreme Court. And if that were not bad enough, the panel majority’s de novo review fails as well. AEDPA—as the Supreme Court has told us—does not permit such conclusions. It is likely time for the Supreme Court to remind us of AEDPA’s requirements. For these reasons, I respectfully dissent from our court’s unwillingness to rehear this case en banc.