Edwards v. Lamarque

Opinion

FOR PUBLICATION UNITED STATES COURT OF APPEALS FOR THE NINTH CIRCUIT

KRISTOPHER C. EDWARDS,  No. 04-55752 Petitioner-Appellee, v.  D.C. No. CV-01-10401-RGK A. LAMARQUE, Warden, OPINION Respondent-Appellant.  Appeal from the United States District Court for the Central District of California R. Gary Klausner, District Judge, Presiding

Argued and Submitted June 7, 2005—Pasadena, California

Filed December 12, 2005

Before: Betty B. Fletcher, Pamela A. Rymer and Raymond C. Fisher, Circuit Judges.

Opinion by Judge Fisher; Dissent by Judge Rymer

16145 EDWARDS v. LAMARQUE 16149

COUNSEL

David C. Cook, Deputy Attorney General, Los Angeles, Cali- fornia, for the respondent-appellant.

Steven S. Lubliner, Petaluma, California, for the petitioner- appellee.

OPINION

FISHER, Circuit Judge:

Petitioner-appellee Kristopher C. Edwards (“Edwards”) was convicted of murder in state court and filed a federal habeas corpus petition alleging, among other things, ineffec- tive assistance of counsel. The district court granted the peti- tion, finding that trial counsel’s actions leading to Edwards’ waiver of his marital privilege constituted ineffective assis- tance of counsel and prejudiced Edwards. The state appeals; we affirm. 16150 EDWARDS v. LAMARQUE I. BACKGROUND

On July 17, 1991, the body of Don Thomas was found in an alley behind a barbershop where Edwards worked. Thomas had died of multiple gunshot wounds from a .38 caliber gun and 9-mm gun. Five years later, after a tip from Edwards’ wife, police arrested Edwards in Detroit, and he was tried in California for the murder and for insurance fraud.

Before trial, and in anticipation of Edwards’ wife Kemet Gaines (“Gaines”) appearing as a witness, Edwards’ counsel John Meyers (“Meyers”) raised the issue of Edwards’ marital privilege under California Evidence Code § 980 and the pre- sumption of confidentiality of marital communications under California Evidence Code § 917. The trial court ruled that Gaines could testify about any discussions between them regarding the insurance fraud scheme as falling under the “fu- ture crimes” exception to the privilege, and also about her observations of Edwards’ behavior on the nights in question, but not about any other conversations between them.

The following facts emerged from Edwards’ and Gaines’ testimony at trial, not including any information from their testimony about their conversations with one another. Edwards and Thomas conspired to defraud Edwards’ automo- bile insurer. They stripped his car of parts, and Edwards cal- led 911 to report that he had been carjacked; the idea was to collect $10,000 from the insurance company for the “stolen” parts, and then replace them. The scheme went awry, how- ever, because Thomas stripped so many parts that the insur- ance company deemed the car “totaled,” resulting in a smaller payout (and loss of the car). Thomas and Edwards exchanged threats and angry calls, including a call from Edwards to Thomas on the night of his death, July 17. Bullets recovered at the scene of the crime and from Thomas’ body were from guns registered to Edwards. The night of the murder, Gaines returned home to find Edwards in a “weird” mood and scrub- bing his hands with laundry detergent. The next evening, July EDWARDS v. LAMARQUE 16151 18, Edwards received a phone call from Thomas’ cousin, threatening Edwards’ and Gaines’ lives. The cousin testified that Edwards threatened to “fuck [him] up too.” Edwards brandished a shotgun and paced around his apartment, staring out the window, and eventually took Gaines first to a nearby motel, then to Florida and finally to Michigan, where they resettled. At trial, Edwards insisted that he had neither com- mitted nor had anything to do with the murder, and that he had given the guns to Thomas as collateral until he was able to pay Thomas for his role in the insurance scam.

Before trial, the police questioned Gaines, who told them that she was prepared to testify that, during an exchange with Edwards on July 17 about why he was washing his hands, he had confessed to committing the murder. Meyers was aware of this proposed testimony.

Despite the trial court’s ruling as to the limited nature of questioning that would be allowed regarding Edwards’ and Gaines’ conversations, Meyers on several occasions asked both Edwards and Gaines questions about their conversations with one another on the nights of July 17 and 18. The collo- quies that followed the prosecutor’s objections revealed that Meyers did not understand what sorts of questions could lead to a waiver of the privilege.1

The waiver issue first surfaced during Meyers’ cross- examination of Gaines. In an apparent attempt to fortify the defense theory that Edwards was vigorously washing his 1 The dissent argues that Meyers’ “questions did not seek to elicit any confidential communications Edwards had with his wife,” Dissent at 16171-72, n.3, yet the very interchange it quotes shows Meyers com- pounding Edwards’ non-responsive answer by asking, “Did you tell her anything?” Meyers may not have sought to elicit confidential communica- tions, but his essentially open-ended question surely violated one of the basic rules of witness examination (whether on direct or cross): “Certainly no lawyer should ask a critical question unless he is sure of the answer.” Francis L. Wellman, The Art of Cross-Examination 39 (1904). 16152 EDWARDS v. LAMARQUE hands on the night of the murder (July 17) because the cou- ple’s dog had messed the carpet and Edwards had cleaned it up, Meyers asked Gaines what reason Edwards had offered for washing his hands on the night of July 17. The prosecutor objected that allowing Gaines to testify about Edwards’ state- ments would waive Edwards’ marital privilege. At a sidebar, Meyers asserted that the question regarding washing hands was “not a confidential communication,” and that Gaines had waived her marital privilege by taking the stand — an argu- ment wholly irrelevant to whether her testimony would waive Edwards’ privilege. The court ruled that Meyers was at risk of allowing a waiver of the privilege and sustained the objec- tion. Meyers withdrew the question.

Later, however, during direct examination of Edwards, Meyers asked Edwards if he had suggested to Gaines that she go to her brother’s house on the night of the murder, and Edwards answered, “Yes, I did.” Then, Meyers asked Edwards about the exchange with Gaines about washing his hands: “Did you tell her about what happened with the dog?” Edwards answered, “About the dogs, yes.” The prosecutor did not object to either of these questions. Thereafter, Edwards testified about the cousin’s threatening phone call the next night, and Meyers asked Edwards what he had told Gaines upon receiving the call. Edwards responded, “I told her, ‘Somebody killed Don and they think I had something to do with it, and they just threatened to come and kill us.’ ” The prosecutor objected for a second time. The court told Meyers at sidebar that his question effectively waived Edwards’ privi- lege and thus had “opened the door” to further testimony from Gaines about her discussions with Edwards on the nights in question. Meyers again stressed his view that he was allowed to ask Edwards about any communication that he did not deem to be “confidential,” and that such questioning did not open the door to the prosecution eliciting any statements from Gaines as to her conversations with Edwards. The court invited Meyers to provide relevant authorities supporting his position the next morning. EDWARDS v. LAMARQUE 16153 Meyers arrived the next morning without any such authori- ties — indeed, without any authorities at all. The judge directed Meyers’ attention to a California appellate court case, People v. Worthington, 38 Cal. App. 3d 359 (3d Dist. 1974), concerning the marital privilege and waiver. It became clear that Meyers had not read it. When the court said, “Somebody gave me some cases,” the prosecutor replied, “That was me,” and Meyers added, “I’m reading those now.” When the court offered Meyers additional time to review Worthington and other cases, he replied, “I’ve pretty well looked at the general parts. Both are attorney-client privilege, that I can see.”

In the discussion that followed, Meyers continued to argue his version of the marital privilege, in which he could elicit “nonconfidential” testimony from Edwards about his discus- sions with Gaines, but prevent Gaines from testifying about her version of the conversations. Meyers also attempted to distinguish Worthington. The court ultimately ruled that the privilege had been waived, and that the prosecution could elicit testimony from Gaines about her discussions with Edwards on the nights of July 17 and 18.

Meyers then stated, “Well, one thing for the record, that raises a significant I.A.C. claim if he is convicted on appeal, in my judgment . . . . [W]hat you’re saying is I erred in asking that question and I, quote, ‘opened the door,’ close quote, to all the other stuff coming in. To me, that’s an inadequate rep- resentation of counsel.” The trial court disagreed, calling it a “tactical decision.” Meyers responded, “It wasn’t a tactic. It was a mistake.”

Given the waiver of the marital privilege, the prosecution was able to elicit the following additional damaging testimony from Gaines: When she came home on the night of the mur- der, July 17, she asked Edwards why he was scrubbing his hands, and, upon receiving no response, whether he had killed Thomas. Edwards responded, “I’ll put it to you like this: you don’t have to worry about hearing from him again.” 16154 EDWARDS v. LAMARQUE The jury returned convictions on the insurance fraud counts, but deadlocked on the murder charge. One juror, angered by another juror’s racial slur, changed her vote and refused to convict.

Before retrial, Edwards successfully moved to have Meyers replaced. See People v. Marsden, 2 Cal. 3d 118 (1970) (hold- ing that defendant has the right to have counsel discharged upon appropriate showing). Edwards argued that Meyers had provided ineffective assistance by leading Edwards to open the door to Gaines’ testimony. The court granted the Marsden motion based on “discord” between Edwards and Meyers. Edwards’ new counsel argued that there had been no waiver of the privilege, that any waiver had been limited in scope and that any waiver was the result of Meyers’ ineffective assis- tance. The court rejected all three arguments.

At retrial, Gaines testified as she had in the first trial, adding that when she asked Edwards whether he had killed Thomas, he replied, “Yeah.” Edwards was convicted of mur- der, and filed a motion for a new trial, again arguing ineffec- tive assistance of counsel. The trial court denied the motion, ruling that Meyers had made a tactical decision. Edwards appealed to the California Court of Appeal, which concluded that Meyers had made a “reasonable tactical decision” and affirmed the conviction. The California Supreme Court denied a petition for review.

Edwards filed a federal petition for habeas corpus, alleging several alternate claims for relief, including ineffective assis- tance of counsel. The district court, adopting the recommen- dation of a magistrate judge, denied all grounds except the ineffective assistance claim. The court concluded that although it owed deference to the state court determination that Meyers’ decision was “tactical,” that finding was clearly erroneous. The court further concluded that Meyers was inef- fective under clearly established Supreme Court precedent, and that Edwards had been prejudiced. The court issued a EDWARDS v. LAMARQUE 16155 judgment granting the petition as to the ineffective assistance claim. The state timely appealed. We have jurisdiction under 28 U.S.C. § 2253.

II. STANDARD OF REVIEW

This appeal is governed by the Antiterrorism and Effective Death Penalty Act of 1996 (“AEDPA”), 28 U.S.C. § 2254, which provides that a federal court may not grant a writ of habeas corpus on behalf of a person in state custody “with respect to any claim that was adjudicated on the merits in state court proceedings unless the adjudication of the claim (1) resulted in a decision that was contrary to, or involved an unreasonable application of, clearly established federal law, as determined by the Supreme Court of the United States; or (2) resulted in a decision that was based on an unreasonable determination of the facts in light of the evidence presented in the state court proceeding.” 28 U.S.C. § 2254(d).

Under the ‘contrary to’ clause, a federal habeas court may grant the writ if the state court arrives at a con- clusion opposite to that reached by [the Supreme Court] on a question of law or if the state court decides a case differently than [the Supreme Court] has on a set of materially indistinguishable facts. Under the ‘unreasonable application’ prong, a fed- eral habeas court may grant the writ 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.

Williams v. Taylor, 529 U.S. 362, 412-13 (2000). State court findings of fact are presumed to be correct unless the peti- tioner rebuts the presumption by clear and convincing evi- dence. 28 U.S.C. § 2254(e)(1).2 2 Our dissenting colleague believes that § 2254(e)’s standard for review- ing state court factual determinations, which creates a presumption of 16156 EDWARDS v. LAMARQUE We review the district court’s decision to grant the writ de novo, and its findings of fact for clear error. Houston v. Roe, 177 F.3d 901, 905 (9th Cir. 1995).

III. DISCUSSION

A. The Right To Effective Assistance of Counsel Has Been Clearly Established by the Supreme Court

“It is past question that the [right to effective assistance of counsel] qualifies as ‘clearly established Federal law, as determined by the Supreme Court of the United States.’ That the [governing] test . . . requires a case-by-case examination of the evidence obviates neither the clarity of the rule nor the extent to which the rule must be seen as ‘established’ by this Court.” Williams, 529 U.S. at 391 (internal citations and quo- tations omitted). The parties do not dispute that the California Court of Appeal applied the correct clearly established Supreme Court law, and we so hold.

[1] We summarize the relevant principles that govern here. “[T]he [Sixth Amendment] right to counsel is the right to

accuracy that can be rebutted only by “clear and convincing evidence,” may be at odds with § 2254(d)’s separate “unreasonable determination of the facts” standard for granting petitions. In the dissent’s estimation, employing the “unreasonable determination” language of § 2254(d), the state courts’ determinations were not unreasonable. However, the Supreme Court recently suggested that the standards merge: “Under the Antiterro- rism and Effective Death Penalty Act of 1996, [petitioner] may obtain relief only by showing the [state] conclusion to be ‘an unreasonable deter- mination of the facts in light of the evidence presented in the State court proceeding.’ Thus we presume the [state] court’s factual findings to be sound unless [petitioner] rebuts the ‘presumption of correctness by clear and convincing evidence.’ ” Miller-El v. Dretke, 125 S.Ct. 2317, 2325 (2005) (emphasis added). Even if the Supreme Court had not recognized the merger of the two standards and we based our analysis solely upon Taylor v. Maddox, 366 F.3d 992 (9th Cir. 2004), we would reach the same conclusion: the state courts’ determinations were not reasonable. EDWARDS v. LAMARQUE 16157 effective assistance of counsel.” McMann v. Richardson, 397 U.S. 759, 771 n. 14 (1970). “Counsel . . . can . . . deprive a defendant of the right to effective assistance, simply by failing to render adequate legal assistance.” Strickland v. Washing- ton, 466 U.S. 668, 686 (1984) (internal citation and quotations omitted). “The benchmark for judging any claim of ineffec- tiveness must be whether counsel’s conduct so undermined the proper functioning of the adversarial process that the trial cannot be relied on as having produced a just result.” Id.

[2] “[A] defendant must [first] show that counsel’s perfor- mance was deficient[,] [which] requires showing that counsel made errors so serious that counsel was not functioning as the ‘counsel’ guaranteed the defendant by the Sixth Amendment.” Id. at 687. Further, “the proper standard for attorney perfor- mance is that of reasonably effective assistance.” Id. Put another way, counsel must perform within “the range of com- petence demanded of attorneys in criminal cases.” McMann, 397 U.S. at 771.

The Court has also emphasized that “[j]udicial scrutiny of counsel’s performance must be highly deferential,” and that “[a] fair assessment of attorney performance requires that every effort be made to . . . reconstruct the circumstances of counsel’s challenged conduct, and to evaluate the conduct from counsel’s perspective at the time.” Strickland, 466 U.S. at 689. This means that “a court must indulge a strong pre- sumption that counsel’s conduct falls within the wide range of reasonable professional assistance; that is, the defendant must overcome the presumption that, under the circumstances, the challenged action might be considered sound trial strategy.” Id. (internal quotation omitted).

[3] If a defendant can show that counsel’s actions consti- tuted ineffective assistance, he must then show that “there is a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different. 16158 EDWARDS v. LAMARQUE A reasonable probability is a probability sufficient to under- mine confidence in the outcome.” Id. at 694.

If Edwards successfully satisfied the Strickland two-part test, he is entitled to habeas relief. However, AEDPA and recent case law require significant deference to the reasoning of the state court.

B. The First Strickland Prong: Ineffective Assistance

1. The District Court Correctly Rejected the State Court’s Finding that Meyers’ Acts Reflected “Tactical” Decisions

The California Court of Appeal discussed the substance of Edwards’ ineffective assistance claim in the following pas- sage:

In view of the strong evidence against appellant, defense counsel could reasonably have decided to introduce appellant’s testimony regarding his version of the conversation he had with Gaines on the night of July 17, 1991. In fact, as previously noted, the first jury deadlocked on the murder charge. Since defense counsel was aware that the trial court had ruled that defense counsel could not introduce evi- dence of the defense’s version of a conversation between appellant and Gaines without opening the door to the prosecution’s version of that conversa- tion, the trial court reasonably determined that elicit- ing appellant’s testimony regarding his conversation with Gaines on the night of July 17, 1991, was a rea- sonable tactical decision rather than ineffective assis- tance of counsel.

The district court held that the state court’s conclusion that Meyers made a “tactical” decision to allow Edwards to waive the privilege was a “factual finding entitled to deference EDWARDS v. LAMARQUE 16159 under the AEDPA.” Edwards contests this conclusion, argu- ing that whether Meyers’ acts were tactical or not is a mixed question of law and fact, which can be distinguished from a purely factual question.

It is academic how the state court’s determination is cate- gorized, because either way it is entitled to significant defer- ence — either as a factual determination to be reviewed under the “unreasonable determination of the facts” prong of AEDPA, or as a legal conclusion, to be reviewed under the “unreasonable application” prong. In either case, the question is whether the state court’s determination can be viewed as reasonable. And either way, the district court was correct that the state court’s ruling cannot stand.

California Evidence Code § 980 establishes a marital privi- lege as follows:

[A] spouse . . . has a privilege during the marital relationship and afterwards to refuse to disclose, and to prevent another from disclosing, a communication if he claims the privilege and the communication was made in confidence between him and the other spouse while they were husband and wife.

The privilege, however, may be waived:

[T]he right of any person to claim [the marital privi- lege] . . . is waived with respect to a communication protected by the privilege if any holder of the privi- lege, without coercion, has disclosed a significant part of the communication or has consented to dis- closure made by anyone.

Cal. Evid. Code § 912.

[4] Meyers’ responses to the prosecutor’s objections and the trial judge’s comments reveal that he fundamentally mis- 16160 EDWARDS v. LAMARQUE understood the marital privilege, and thus lacked the legal understanding necessary for a competent tactical decision. Meyers plainly believed that he could prevent Gaines from testifying as to certain “confidential” conversations between her and Edwards (such as Edwards’ alleged confession to her), but that he could pick and choose other parts of the con- versations they had and elicit testimony as to Edwards’ ver- sions of those exchanges to bolster his defense. Thus, when the prosecutor first objected to the court that Meyers was “going into the area of confidential marital communication” by having Edwards report the conversation he had with Gaines about the washing of his hands, Meyers’ revealing response was, “No, no. That is not a confidential communica- tion.” Later, after causing Edwards to waive his privilege, Meyers argued to the court, “My feeling is every statement made between a husband and wife is not a confidentiality privilege.” Indeed, Meyers’ questioning (and arguments to the judge) demonstrate that he believed that he could simulta- neously prevent Gaines from testifying as to the specific exchange about the dog mess (the exchange where, in Gaines’ version, Edwards confessed to her), and yet draw from Edwards his version of that very same exchange.3 Meyers thus sought to introduce Edwards’ explanation for cleaning his hands (that their puppy had messed the floor) as well as the communication in which Edwards relayed to Gaines the threat from Thomas’ cousin to kill the two of them. When the prose- cution objected, Meyers insisted upon his view of the law, and, as the record indicates, appeared astonished by the trial court’s explanation of how waiver actually worked. After the court advised him that he could not characterize “certain com- munications as confidential” merely because they might con- 3 Meyers also directly asked Gaines what Edwards told her when she asked why he was washing his hands — the exchange that Gaines later would testify led to Edwards’ confession. Without the prosecutor’s objec- tion regarding the marital privilege, Meyers might well have elicited testi- mony about the confession at this moment. What tactical approach, if any, this represented is impossible to discern. EDWARDS v. LAMARQUE 16161 cern a particular subject, Meyers’ surprised response was, “That’s a pretty good issue.” He even stated that Gaines had waived the spousal privilege “by being here and taking the stand” — a comment with no bearing on whether Edwards, the defendant in the trial, had waived his privilege of prevent- ing his wife from testifying. Gaines’ own marital privilege was irrelevant, and Meyers’ invocation of it further demon- strates his lack of understanding of the doctrine.

[5] Fundamentally, Meyers had no conception of the most basic premise of the spousal privilege and could therefore not make competent tactical decisions regarding it. He did not recognize that all private communications between spouses are “presumed to have been made in confidence.” North v. Superior Court of Riverside County, 8 Cal. 3d 301, 310 (Cal. 1972) (citing 8 Wigmore, Evidence, § 2336 [McNaughton rev. ed. 1961]); cf. Blau v. United States, 340 U.S. 332, 333 (1951) (“[M]arital communications are presumptively confi- dential”); Wolfle v. United States, 291 U.S. 7, 14 (1934); United States v. Strobehn, 421 F.3d 1017, 1021 (9th Cir. 2005) (Rymer, J.) (noting that “[f]ederal common law assumes that private communications between spouses are intended to be confidential, and thus privileged”). Corre- spondingly, Meyers showed no understanding that courts nar- rowly construe the privilege because it “prevent[s] the admission of relevant and otherwise admissible evidence” and impedes the search for truth. People v. Sinohui, 28 Cal. 4th 205 (Cal. 2002); cf. Strobehn, 421 F.3d at 1021. Given Mey- ers’ evident misconceptions about the nature and scope of the marital privilege, Meyers was incapable of making competent tactical decisions of the sort the state court (and the dissent) imagines. It is no surprise, therefore, that when the trial court invited Meyers to provide authorities for his view, after he already opened the door to Gaines’ damaging testimony, he had none to put forward.4 4 The dissent relies on Meyers’ extemporaneous attempt to distinguish Worthington to argue that he could reasonably have thought that waiver 16162 EDWARDS v. LAMARQUE [6] Finally, upon grasping that he had led his client to waive the privilege, Meyers immediately asserted that he had made a mistake.5 His confession of error is, of course, not itself dispositive. See Morris v. California, 966 F.2d 448, 456-457 (9th Cir. 1991) (determining whether a decision was tactical based on counsel’s statement and additional facts). Nonetheless, when considered in conjunction with the evi- dence reviewed above, it is further evidence of his confusion.6

[7] The district court correctly found that the court of appeal’s conclusion that Meyers made a tactical decision to waive the privilege is rebutted by clear and convincing evi- dence in the record. Such a finding means that the state court determination was based on an “unreasonable determination of the facts in light of the evidence presented in the State court proceeding.” 28 U.S.C. § 2254(d)(2); Miller-El v. Dretke, 125 S.Ct. 2317, 2325 (2005).7

of the privilege as to one communication did not open the door to disclo- sure of the other marital communications since Worthington dealt with dueling versions of a single conversation. To so argue posits that Meyers made a tactical decision ignorant of the key marital privilege case, which the trial judge had to show to him; and credits his post-hoc creativity in trying to undo the harm he had already incurred by opening the door. Nothing in the record shows that Meyers made a decision to ask his high risk questions informed by a minimally competent understanding of the relevant law of marital privilege. It shows a misguided blunder. 5 According to Edwards’ second attorney, Meyers later offered to so tes- tify. 6 Edwards highlights several other instances in the trial where Meyers betrayed confusion about the rules of evidence and needed to be assisted by the trial judge. For example, he was confused about: the admissibility of evidence of a defendant’s character as opposed to evidence of a vic- tim’s character; hearsay; laying the foundation for testimony; and hearsay and expert testimony. None of these episodes is directly relevant, but they do reinforce the picture of Meyers as an incompetent trial attorney. 7 The second trial judge had a slightly different theory, that Meyers made a tactical decision to risk waiving the privilege. Our foregoing anal- ysis applies equally to this theory. We further note that if Meyers really wanted to pursue a strategy of nearly waiving the privilege without actu- EDWARDS v. LAMARQUE 16163 2. Alternatively, the District Court Correctly Concluded that Meyers’ Actions, if Tactical, Were Objectively Unreasonable

[8] The district court alternatively found that even if the state court reasonably concluded that Meyers had made a tac- tical decision to waive the privilege, the state court unreason- ably applied Strickland in finding such tactics themselves to be reasonable. The district court recognized the “wide lati- tude” given to counsel in determining trial strategy, and the importance of considering counsel’s behavior from the per- spective counsel himself had, rather than with hindsight. Nev- ertheless, it concluded that “[e]ven if it was a conscious strategy, Meyers’s decision to waive petitioner’s right to assert the marital privilege was so completely uninformed and misguided that a reasonably competent lawyer would not have done the same thing in like circumstances.” We agree.

The district court recognized that the Supreme Court and this circuit have applied a cost-benefit analysis to such tactical decisions. See, e.g., Darden v. Wainwright, 477 U.S. 168, 186 (1986) (“Any attempt to portray petitioner as a nonviolent man would have opened the door for the State to rebut with evidence of petitioner’s prior convictions.”); Mak v. Blodgett, 970 F.2d 614, 619 (9th Cir. 1992) (“[N]o risk would have been incurred by presenting the proferred evidence.”).

As the district court found, if we apply a cost-benefit analy- sis to Meyers’ decision to waive the privilege, the ledger is one-sided. Gaines was openly hostile to Edwards; she had tes- tified to the police that he had confessed to the murder to her,

ally doing so, the reasonable course would have been to seek, through in limine motions, to define the exact scope of the privilege and what would and would not waive it. See, e.g., People v. Dorsey, 46 Cal.App.3d 706, 718 (1975) (“The [marital privilege] claims should have been asserted and ruled upon in camera.”). 16164 EDWARDS v. LAMARQUE and Meyers was aware of that information. Indeed, it was Gaines who, five years after the murder, tipped off the police that her husband was the murderer. To offset this near-certain and extremely damaging potential testimony, Meyers could hope to demonstrate that Gaines knew that she too had been threatened by Thomas’ cousin. But as the district court observed, the jury would know either way that Thomas’ cou- sin had threatened Edwards and Gaines, and that would be reason enough to explain why Edwards and his wife fled the state.8 The additional testimony that Meyers apparently hoped to elicit was not even likely to come: Meyers appeared to hope, for example, that Gaines would testify that Edwards had said he was cleaning his hands because the dog had messed the rug, but when allowed to testify as to the exchange, she did no such thing, and instead — consistent with her state- ments to the police — testified that her question led quickly to his confession to the murder.

Nevertheless, if we accept that Meyers made a tactical deci- sion to waive, the question does become closer. Given the extremely strong circumstantial evidence against Edwards, an attorney in Meyers’ position would be forced to consider numerous different strategies, all of which could be risky or prone to failure, in an effort to win an acquittal. However, nearly all of the supposed advantages of Gaines’ testimony could be reaped without her testimony. Edwards could testify himself that he was cleaning up a dog’s mess, and it is diffi- cult to see how Meyers could hope to supplement that testi- mony with Gaines’ perspective. It comes down to the difference between Edwards testifying that Thomas’ cousin threatened his and his wife’s life, and Edwards giving that tes- 8 The dissent implausibly argues that Meyers could counter the inculpat- ing circumstantial evidence only by contending that Edwards and his wife fled because they had been threatened. However, even if Meyers reason- ably viewed the threat as an important piece of evidence in favor of his client, there was no need for him to waive Edwards’ privilege to provide an alternative explanation for his and Gaines’ sudden departure. EDWARDS v. LAMARQUE 16165 timony supplemented by Gaines’ testimony that she was aware of the threat. When balanced against the near-certainty of Gaines’ testifying as to Edwards’ confession to the murder, the tactical decision to allow the privilege to be waived was objectively unreasonable.

The California Court of Appeal’s opinion gives little expla- nation for why it concluded that Meyers made a “reasonable tactical decision” to allow the privilege to be waived. The main reason offered is that there was “strong evidence” against Edwards — apparently suggesting that desperate straits call for desperate measures. However, for the adversary process to function effectively, there must be some limits on counsel’s willingness to take enormous risks, and Meyers’ decision falls outside those limits.

The state seeks to support its view by reference to Davis v. Woodford, 384 F.3d 628 (9th Cir. 2004). In that case, we con- cluded that defense counsel’s decision to allow an expert to testify despite the expert’s statement that he might offer “pos- sibly damaging information” was a reasonable tactical deci- sion. Id. at 648. Upon learning that the expert might do so, counsel sought a continuance, and when it was denied, made the decision to allow the expert to testify. We ruled that “[w]hen the effort to postpone the [proceedings] and bring in a new psychiatric expert failed, defense counsel quite reason- ably decided that the risk of a slight unknown was preferable to presenting no psychiatric testimony . . . . [Defendant] needed someone who could . . . put together the pieces to demonstrate, as [the expert] did, that [defendant] was not entirely a creature of his own making, that he had psychologi- cal problems, and that hope existed for some type of rehabili- tation.” Id.

Contrary to the state’s position, Davis actually illustrates how a decision to allow the admission of damaging testimony, depending on numerous factors, can be either eminently rea- sonable or clearly unreasonable. See Strickland, 466 U.S. at 16166 EDWARDS v. LAMARQUE 693 (“[A]n act or omission that is unprofessional in one case may be sound or even brilliant in another.”). In Davis, counsel had a strong incentive to put on an expert who could help illustrate why his client “was not a creature of his own mak- ing,” but was faced with a situation where his only option was an expert who had warned that he would be entirely candid and, possibly, say things damaging to the defendant’s case. In such a situation, the benefit was clearly strong, because with- out any expert testimony about the defendant’s psychology, there was a gaping hole in the defense; whereas the risk of harmful statements was, if not remote, at least uncertain — counsel was not aware what the expert might say.

In this case, the balance tilts the other way. The benefits to be gained from Gaines’ testimony (or Edwards’ statements as to what he told her) are few. It is on this side of the equation that there is uncertainty: Meyers could not know whether Gaines would provide the sort of exculpatory testimony he apparently hoped she would. On the other side of the equation there is no uncertainty (or at least very little): Meyers knew that Gaines, a hostile witness, was prepared to testify that her husband had confessed to the murder to her. Davis thus illus- trates how in this instance, given these facts, Meyers acted outside the bounds of professional competence.

[10] Regardless of whether Meyers’ acts are viewed as reflecting a tactical decision to waive the privilege or as a plain mistake, the California Court of Appeal’s conclusion that his performance was adequate was an unreasonable appli- cation of clearly established Supreme Court precedent, or was based on an unreasonable determination of the facts in light of the evidence presented at trial.

C. The Second Strickland Prong: Prejudice

We must now consider whether Edwards was prejudiced by Meyers’ ineffective assistance. Because the court of appeal found no ineffective assistance, it never reached the prejudice EDWARDS v. LAMARQUE 16167 inquiry.9 The first question, then, is whether there is anything to defer to in making this determination.

[10] The district court found that the court of appeal had not addressed the prejudice issue, but that the trial court had. The second trial judge, upon denying the motion for a new trial, stated, “I am convinced that without the testimony that I allowed in, that the verdict may well have been a different one. So that I want to make it clear that if I’m wrong on this [the first Strickland prong], I do think it’s reversible from my standpoint.” The district court found this perspective “highly persuasive” and “entitled to deference” because “it is the last (and only) reasoned opinion of the state court on the issue of prejudice.” We agree on both counts. Because the trial court’s view was the “last reasoned decision” by a state court on the issue of prejudice, we owe it deference, Campbell v. Rice, 408 F.3d 1166, 1170 (9th Cir. 2005); and we are independently persuaded of the existence of prejudice in any event. As the district court noted, a defendant’s own confession is the most damning sort of evidence available against him. See, e.g., Ari- zona v. Fulimante, 499 U.S. 279, 296 (1991) (quoting Bruton v. United States, 391 U.S. 123, 139-40 (1968)) (“A confession is like no other evidence. Indeed, the defendant’s own confes- sion is probably the most probative and damaging evidence that can be admitted against him . . . . Certainly, confessions have a profound impact on the jury, so much so that we may justifiably doubt its ability to put them out of mind even if told to do so.”) (internal quotations omitted). It is true that the circumstantial evidence against Edwards was extremely strong here, including testimony from the victim’s cousin that Edwards threatened to “fuck you up too.” Upon retrial with- out any of Gaines’ testimony, he might well be convicted; but, as the second trial judge himself noted, it is impossible to be 9 The state’s argument that the California Court of Appeal did conduct a prejudice inquiry is meritless; the court of appeal had no reason to reach the prejudice inquiry after finding no ineffective assistance, and its opin- ion cannot plausibly be read to contain an alternative holding. 16168 EDWARDS v. LAMARQUE free from doubt in the verdict given the improper admission of Gaines’ testimony. For example, the prosecutor empha- sized the confession during closing arguments, even reading Gaines’ statement to the police verbatim. It is clear at least that the prosecutor believed that this evidence was of central importance.10

[11] If we defer to the trial court’s conclusion of prejudice, it is established. Alternatively, our independent inquiry leads us to conclude that prejudice is established here. The proba- bility of a different result is “sufficient to undermine confi- dence in the outcome.” Strickland, 466 U.S. at 694.

IV. CONCLUSION

We affirm the district court. Meyers’ acts were not tactical but simple incompetence; alternatively, if tactical, they were objectively unreasonable; and there is a reasonable probability that but for Meyers’ ineffective assistance, Edwards would have received a different verdict.

AFFIRMED.

10 The state argues that the fact that the first jury hung proves no preju- dice. The dissent argues further that the hung jury is probative of the effec- tiveness of Meyers’ “tactical” decision. Both arguments are irrelevant. Edwards was convicted upon retrial, and it is this conviction from which he seeks relief. Further, as the prosecution recognized at the time, the mis- trial was almost certainly the result of nullification. EDWARDS v. LAMARQUE 16169 RYMER, Circuit Judge, dissenting:

I part company because in my view, the state courts did not make an unreasonable determination of the facts in light of the record before them when they found that John Meyers, Edwards’s counsel at his first trial, made a tactical decision in asking Edwards questions on the stand that waived the marital communications privilege. Nor did the state courts apply Strickland v. Washington, 466 U.S. 668 (1984), unreasonably in concluding that this was not an unreasonable strategy for Meyers to pursue, given the substantial amount of circumstan- tial evidence implicating Edwards that had already been pre- sented, an implicit confession by Edwards to the victim’s cou- sin that he was the murderer, Edwards’s unexplained flight with his wife two days after the murder, and Meyers’s reason- able belief, in light of the state of California law at the time, that he could elicit testimony from Edwards about his conver- sations with his wife on the night after the murder without opening the door to his wife’s version of their conversation on the night of the murder. Therefore, I disagree with the deci- sion of the district court granting Edwards’s habeas petition, and dissent from the majority’s contrary disposition.

At the outset, the district court reviewed the state courts’ finding that Meyers made a tactical decision, not a mistake, under the “clear and convincing evidence” standard of 28 U.S.C. § 2254(e)(1) rather than, as we have since clarified in Taylor v. Maddox, 366 F.3d 992, 999-1000 (9th Cir. 2004), for whether that determination was unreasonable under § 2254(d)(2).1 The district court held that Meyers’s declara- tion, his stipulated testimony that he had made a mistake dur- ing the first trial, his statements to the trial judge at sidebar during the first trial, and his apparent ignorance of the nature 1 As the majority notes, the Supreme Court in Miller-El v. Dretke, 125 S.Ct. 2317, 2325 (2005), intimates that the standards under §§ 2254(d)(2) and (e)(1) may merge. However, the Court’s passing recitation of the AEDPA standard does not undermine Taylor, which is binding on us. 16170 EDWARDS v. LAMARQUE and scope of the marital communications privilege, amounted to clear and convincing evidence that the state courts’ finding was incorrect. See 28 U.S.C. § 2254(e)(1). However, the state courts’ finding should not have been reviewed under § 2254(e)(1) because Edwards introduced no new evidence in the district court. Instead, as we explained in Taylor, when a federal habeas petitioner introduces no new evidence and relies solely on the state court record in challenging a state court factual determination, the proper inquiry is whether the state court made “an unreasonable determination of the facts in light of the evidence presented in the State court proceed- ing.” 28 U.S.C. § 2254(d)(2). Only if a state court factual determination survives this intrinsic review and the petitioner presents new evidence in district court that was not in the state court record may the district court proceed to the extrin- sic review contemplated by § 2254(e)(1) and ask whether the new evidence clearly and convincingly shows that the state court factual determination was incorrect. Edwards presented no new evidence in the district court that was not in the state court record, and did not seek an evidentiary hearing in the district court to elicit any such evidence; accordingly, the proper question is whether the California Court of Appeal’s determination that Meyers was not ineffective was based on an unreasonable determination of the facts in light of the evi- dence before it.

Here, two state trial judges and three state appellate judges found that Meyers’s decisions were tactical. While this is a close call and we may or may not have made the same judg- ment ourselves, that is not the question. The question is whether the state courts’ determination was unreasonable. I can’t say that it was.

Unlike the situation in People v. Dorsey, 46 Cal.App.3d 706 (1975), in which the defendant’s trial counsel failed to raise the marital communications privilege at all, Meyers clearly knew about the privilege and timely raised it before trial began. Edwards argues that Meyers was ineffective EDWARDS v. LAMARQUE 16171 because he was ignorant about the operation and scope of the privilege and of how waiver operates under California law. However, the state courts concluded that Meyers took a calcu- lated risk.2 It appears from the record that Meyers believed that he could question Edwards about his conversation with his wife on July 18, the night after the murder, without waiv- ing the privilege as to Edwards’s different conversation with her on July 17, the night of the murder. The confession occurred during the conversation on July 17. Meyers with- drew an earlier question about Edwards’s July 17 conversa- tion with his wife regarding the reason Edwards was washing his hands, and thereby avoided opening the door to Edwards’s wife’s testimony about Edwards’s confession to her during that same conversation. The question Meyers did pursue with Edwards was about the telephone call he received the next evening and what he told his wife about it, which elicited from Edwards evidence that both he and his wife had been threat- ened.3 At the point when Meyers asked this question, he was 2 Our review of the state courts’ decisions should not be influenced by the fact that Meyers fell on his sword during the first trial. After the trial judge ruled that his questions had opened the door, Meyers immediately asserted “that’s inadequate representation of counsel.” If anything is tacti- cal, that claim surely is, and it demonstrates how quickly Meyers could think on his feet. 3 In fact, the question Meyers asked Edwards did not call for Edwards to reveal any confidential communications he had with his wife. Meyers asked Edwards: Q: Where was Kemet [Edwards’s wife] at that time [the night after the murder]? A: She started hollering, “What’s going on? What’s going on? What’s happening? What’s wrong? What’s going on?” Q: Did you tell her anything? A: I told her, “Somebody killed Don and they think I had some- thing to do with it, and they just threatened to come and kill us.” (RT 545-46.) As this excerpt demonstrates, Edwards’s answers were not responsive to Meyers’s questions. Meyers’s questions did not seek to elicit 16172 EDWARDS v. LAMARQUE faced with a considerable amount of circumstantial evidence of Edwards’s guilt, an implicit confession related by the vic- tim’s cousin — who testified that Edwards told him “And I’ll fuck you up too” — and Edwards’s flight with his wife two days after the murder for no apparent reason other than con- sciousness of guilt. The only way Meyers could combat the inevitable inferences to be drawn from this evidence was to defuse the flight issue by making it look like Edwards and his wife fled because they had been threatened, not because he was guilty or because his wife was scared of him instead of a threat by a third party.

Meyers also argued in state court that eliciting evidence from Edwards about a different conversation with his wife that occurred on a different day about a different topic — threats to Edwards and his wife — did not waive the privilege as to all conversations between Edwards and his wife that took place after the murder. Meyers’s position on this point was far from unreasonable. When the trial judge directed the prosecutor’s and Meyers’s attention to People v. Worthington, 38 Cal.App.3d 359 (1974), Meyers argued that Worthington was distinguishable because the dueling versions of a commu- nication between a husband and wife in that case all had to do with a single conversation. Meyers was quite correct on this point, and was not unreasonable in maintaining that Wor- thington could not possibly stand for the proposition that a waiver of the marital communications privilege as to one con- versation on one day about one subject opens the door to

any confidential communications Edwards had with his wife. Thus, noth- ing Meyers asked Edwards on the stand indicates that he expected Edwards to waive the privilege through his responses, and it is not clear how Meyers acted unreasonably in asking the questions that he did. The question “Did you tell her anything?” calls for a “yes” or “no” answer, which is not a privileged communication. The fact that Edwards spontane- ously waived the privilege by giving non-responsive answers to Meyers’s questions does not mean that Meyers was unreasonable in asking those questions. EDWARDS v. LAMARQUE 16173 admission of confidential marital communications that occurred during a different conversation on a different day about a completely different subject. That Meyers turned out to be wrong in his position because the trial judge disagreed with him about the scope of the waiver does not make the cal- culated risk he took unreasonable given that, had he suc- ceeded, he would have been able to explain Edwards’s flight and maybe create reasonable doubt by suggesting that the per- son who made the threat was the murderer instead of Edwards.

Finally, Edwards was not convicted on the first-degree murder charge in the first trial. The jury may have hung due to unrelated racial tension among the jurors, as Edwards con- tends, or it may have hung because of reasonable doubt about whether Edwards was the murderer. The point is that no one knows why Edwards was not convicted in the first trial, but that he was not is at least some indication that Meyers’s tacti- cal decision may have worked.

I would reverse, and part company with my colleagues, because I cannot conclude that the state courts applied Strick- land in an objectively unreasonable manner or made an unrea- sonable determination of the facts in light of the state court record in deciding that Meyers made a reasonable tactical decision that resulted in waiver of the marital communications privilege.