NOT FOR PUBLICATION FILED

UNITED STATES COURT OF APPEALS JUN 29 2018

MOLLY C. DWYER, CLERK

U.S. COURT OF APPEALS

FOR THE NINTH CIRCUIT

KELSIE JAMES PALMER, No. 15-55222

Petitioner-Appellant, D.C. No. 2:14-cv-03246-BRO-JPR v.

DAVE DAVEY, Warden, MEMORANDUM*

Respondent-Appellee.

Appeal from the United States District Court for the Central District of California Beverly Reid O’Connell, District Judge, Presiding

Argued and Submitted March 6, 2018 Pasadena, California

Before: GOULD and MURGUIA, Circuit Judges, and CHRISTENSEN,** Chief District Judge.

Following a jury trial with codefendants Joel Childress and Eric Allen,

Petitioner Kelsie Palmer was convicted of the murder of Rosa Gallegos, the

attempted murder of Luis Miralda and Kenneth Thomas, and for making criminal

* This disposition is not appropriate for publication and is not precedent except as provided by Ninth Circuit Rule 36-3. ** The Honorable Dana L. Christensen, Chief United States District Judge for the District of Montana, sitting by designation. threats against Yvonne Love. On direct appeal, Palmer contended that his

constitutional right to confrontation was violated when the trial court admitted

Allen’s confession without properly redacting it, in violation of Bruton v. United

States, 391 U.S. 123 (1968) and its California counterpart. The California Court of

Appeal denied Palmer’s appeal. Palmer filed a writ of habeas corpus, which the

district court denied. He now appeals. For the reasons stated below we reverse

with instructions to grant the writ.

I.

On the morning of May 8, 2009, sometime after 10:00 a.m., Love was

walking along Exposition Boulevard in Los Angeles when a black Malibu

Chevrolet pulled up beside her. The back seat passenger got out of the vehicle and

held a gun to her side. The driver also stepped out of the vehicle and threatened

her. This encounter lasted a number of minutes before the vehicle drove away.

Love took down the license plate and called 911.

Shortly after the Love encounter, Gallegos was sitting in the driver’s seat of

her parked car outside the home of her boyfriend, Miralda. Miralda lived on 11th

Avenue near Jefferson Boulevard, in a neighborhood controlled by the 18th Street

gang. Miralda and his neighbor, Thomas, were standing outside of the vehicle

when they observed a black Malibu Chevrolet pull up alongside Gallegos’s car. A

few seconds later, the back seat passenger stepped out of the vehicle and began

2 firing. Miralda escaped behind the house, uninjured. Thomas was shot in the arm.

Gallegos was struck four times and died from her wounds.

The following day Miralda and Love were asked to identify Palmer in a

“six-pack” photographic spread. Love’s identification was inconclusive, but

Miralda correctly picked Palmer out of the lineup. At a subsequent live lineup that

contained no fillers from the previous spread, Love identified Palmer. At a

separate lineup, Miralda was unable to identify Palmer and instead picked out one

of the fillers.

Childress, Allen, and Palmer were arrested. While in custody, Allen

admitted to the shooting. Palmer then brought a motion to be tried separately from

his codefendants. Palmer’s counsel argued that the confession was prejudicial to

Palmer. Based on the prosecutor’s representation that he would not introduce

Allen’s testimony unless Allen took the stand, the court denied the motion.

However, on the eve of trial, the prosecutor proposed to introduce a redacted

transcript of Allen’s interview. Palmer objected, arguing that it violated Palmer’s

right to confrontation because the interview mentioned Palmer by name. The trial

court ordered that Palmer’s name be redacted and overruled the objection.

At trial, evidence showed that Palmer was a member of the Black P-Stones

street gang. The jury also heard testimony from Love, including her 911 call in

which she identified the individuals in the vehicle by their clothing. The jury heard

3 from Miralda who described getting a brief look at the front seat passenger before

the shooting began.

The gun was never recovered, but fourteen shell casings at the scene

matched casings found at Childress’s residence. Thirty-two prints were lifted from

the vehicle and both Childress’s and Allen’s prints were positively identified. No

prints were ever matched to Palmer. Evidence of phone records indicated that

Palmer received two phone calls placing him in the approximate location of the

shooting near the time of the shooting.

Over Palmer’s renewed objection, the prosecution introduced an audio

recording of Allen’s testimony accompanied by a 45-page transcript. Both the

transcript and the audio-recording were redacted to eliminate the specific names of

those in the vehicle, but indicated their existence by use of the pronouns “we,”

“they,” and “them two.” After the transcript of Allen’s testimony was provided to

the jury, the court instructed them that the contents of the testimony were admitted

“against Defendant Allen only not to any other Defendant.” At the conclusion of

trial, the jury convicted Palmer of all crimes charged.

Palmer appealed, claiming his right to confrontation was violated when the

trial court admitted Allen’s testimony. The California Court of Appeal found there

was no Bruton error, and that, in any event, the error was harmless beyond a

reasonable doubt. The California Supreme Court denied review.

4 Palmer filed a pro se habeas petition in federal court. Although the federal

court found that admission of Allen’s confession was a Bruton error and violation

of Palmer’s right to confrontation, it found that the admission did not have a

“substantial and injurious effect” on Palmer’s verdict. This appeal followed.

II.

We review the district court’s denial of a habeas petition de novo. Jones v.

Harrington, 829 F.3d 1128, 1135 (9th Cir. 2016). Palmer’s appeal is governed by

the Anti-Terrorism and Effective Death Penalty Act (AEDPA). We must give

deference to the California Court of Appeal’s conclusion on the merits, unless we

conclude that the court’s decision “was contrary to, or involved an unreasonable

application of, clearly established Federal law, as determined by the Supreme

Court of the United States” or “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)(1)–(2).

III.

A. Violation of Clearly Established Federal Law

A defendant has a right of cross-examination secured by the Confrontation

Clause of the Sixth Amendment. Bruton, 391 U.S. at 126–27. Admission of a

codefendant’s confession that inculpates the defendant violates this right. Id. at

126.

5 The California Court of Appeal concluded that there was no confrontation

error because the neutral pronouns used in Allen’s confession did not render it “so

powerfully incriminating against Palmer, that the jurors could not be expected to

follow” the limiting instruction. The court concluded that “[t]he only inference

that reasonable jurors could not avoid drawing from . . . [the] testimony and the use

of the pronouns ‘we’ and ‘they’ was that Allen and Palmer belonged to the same

gang and were acquainted.” This conclusion is an unreasonable application of

Gray v. Maryland, 523 U.S. 185, 192–95 (1998) and contrary to the Supreme

Court’s holding in Bruton.

Allen’s testimony violated Palmer’s right to confrontation because the

statement plainly indicated the existence of two other participants and Allen did

not take the stand. Allen’s statement indicated others in the shooting through use

of pronouns and references (“you three guys” “we” “they” “them” and “them

two”). A recording of Allen’s interview was played for the jury accompanied by

its transcript. While the recording was not preserved for appeal, the recording was

necessarily spliced, rendering the testimony “obviously redacted” in violation of

Gray. See Gray, 523 U.S. at 194.

The California Court of Appeal’s decision was contrary to clearly

established federal law where it concluded that any incriminating inferences drawn

by the jury were cured by the trial court’s limiting instruction. This conclusion

6 contravenes Bruton’s clear holding that a limiting instruction does not preserve a

defendant’s right to confrontation. Bruton, 391 U.S. at 137.

Although we give considerable deference to the state court’s decision, we

must set deference aside when the conclusion reached is “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). Because we

conclude that the California Court of Appeal’s decision was both, we hold that 28

U.S.C. § 2254(d) does not bar habeas review of Palmer’s claim. We conclude that

Palmer’s rights were violated when the trial court admitted Allen’s testimony.

B. Harmlessness

Palmer is not entitled to relief if the confrontation error was harmless. In

AEDPA proceedings, we apply the actual-prejudice standard under Brecht v.

Abrahamson, 507 U.S. 619, 637 (1993).1 Davis v. Ayala, 135 S. Ct. 2187, 2198–

99 (2015). We afford relief only if the error had a “substantial and injurious

effect” on the jury’s verdict. Id. at 2197–98. This standard is satisfied when the

record reveals “grave doubt” as to whether the error influenced the jury’s decision.

1 Because the Brecht standard is “less onerous” for the state than the AEDPA/Chapman standard applied on direct review, the Supreme Court has explained that “it certainly makes no sense to require formal application of both tests . . . when the latter obviously subsumes the former.” Fry v. Pliler, 551 U.S. 112, 120 (2007) (emphasis in original). Accordingly, we apply the Brecht standard.

7 Id. at 2197–98 (quoting O’Neal v. McAninch, 513 U.S. 432, 436 (1995)). We give

deference to a state court’s harmlessness determination. Jones, 829 F.3d at 1141.

In considering whether the error resulted in actual prejudice, we consider

several factors as described in Delaware v. Van Arsdall, 475 U.S. 673, 684 (1986).

Whelchel v. Washington, 232 F.3d 1197, 1206 (9th Cir. 2000). We address these

factors below.

1. The importance of the witnesses’ testimony in the prosecution’s case

Allen’s testimony was essential to the prosecution’s case. Although the state

presented evidence that supported the general motive for the shooting—this was a

turf-war between rival gangs—the prosecution’s evidence against Palmer was thin.

Allen’s testimony provided the link to put the pieces of the state’s case together.

While other evidence tended to show that Palmer was in the same gang with

Childress and Allen, Allen’s testimony, linking the three defendants together in the

car, was damning to Palmer.

The prosecution’s main evidence consisted of two eye witness

identifications from Love and Miralda, and two pieces of circumstantial evidence.

While over thirty finger prints were found in the vehicle and prints were matched

to both Childress and Allen, no prints were ever matched to Palmer.2

2 The dissent notes the two eyewitnesses’ testimony as key facts that persuade it

8 Love’s identifications of Palmer were inconsistent and as to her testimony at

trial, as the trial judge put it, “[t]he jury is turned off, her credibility is shot.” Love

first was unable to identify Palmer in a lineup. She did, however, identify Palmer

at a subsequent lineup. Love testified as to the clothing of the passengers in the

vehicle, but provided three different descriptions. First, in her initial 911 call, Love

described three separate items of clothing worn by the individuals: a black hoodie,

a red and gray sweatshirt, and a “gray shirt by itself.” Second, in her interview

with the police, Love is clear that the front seat passenger was wearing a gray,

long-sleeved, collared button-up shirt with a red print all over. Third, at the live

lineup of suspects ten days after the shooting, Love identifies a red and gray

sweatshirt that was found at Palmer’s residence and states the front seat passenger

wore the sweatshirt. Then, at trial, Love was adamant that the sweatshirt belonged

to Childress, the driver, not the front seat passenger. Love’s inconsistent

identification of Palmer and testimony as to the description of the clothing and who

wore it, renders Love’s testimony of little probative value.

Love’s behavior at trial also put her credibility and the reliability of her

that the state court decision of harmless error was a reasonable one. The dissent also contends that although one might dispute the strength of the eyewitness identifications, the jury could weigh the witnesses’ credibility, and the jury convicted Palmer. This conclusion, however, fails to consider the significant issues with Love’s identification of Palmer, and that the jury may not have relied on Love at all, and rather focused on the key evidence of Allen’s admission.

9 testimony at issue. She was combative and inappropriate on the stand. For

example, she accused the defense attorney of having dementia and asked the judge

if his wife was cheating on him. Twice she physically tried to leave the witness

stand. At one point, during a sidebar with counsel, the judge instructed he would

not allow counsel to impeach her, stating, “The jury is turned off, her credibility is

shot.” It was evident to the trial judge that there was no need for impeachment

because Love already lacked credibility as the state’s witness. As the dissent

notes, the jury had the ability to weigh Love’s testimony. In light of the significant

issues with her testimony, it is highly probable that the jury did not rely on it, and

focused on stronger evidence against Palmer, including Allen’s confession that

linked Palmer to the crime.

Miralda was the state’s second eyewitness. Although Miralda was credible,

he admitted that his recollection of the events was spotty. He got a single look at

two black males in the front seat of the car, but once the shooting began, Miralda

dropped to the ground and ran behind the house. On the day of the shooting,

Miralda gave no description of anyone in the vehicle; he only recalled that the

front passenger was black. Miralda was also inconsistent in his identifications of

Palmer. He first correctly picked Palmer out of a “six-pack” photographic spread,

but then was unable to identify Palmer and picked out a “filler” in a live lineup.

The potential importance of Miralda’s one positive identification is problematic

10 given the inconsistencies in his identification and his limited observation during

the shooting.

In sum, there were significant issues with the state’s two eyewitnesses that

could have led the jury to question the probative value of their testimony. In light

of the weaknesses of the eyewitness testimony, which was a key piece of the

state’s case, the record reveals “grave doubt” as to whether the error of admitting

Allen’s confession influenced the jury’s decision. See Davis, 135 S. Ct. at 2198.

As to the prosecution’s cell phone evidence, the evidence is, admittedly,

probative against Palmer. The prosecution provided Palmer’s cell phone records

that indicate that within a few minutes of the time the shooting is believed to have

occurred, two calls were made to Palmer’s cell phone. A Metro PCS custodian of

records testified about how cell phone towers function and other evidence showed

that one of the cell phone towers was likely .3 miles from the crime scene. The

testimony elicited showed that a cell phone does not necessarily utilize the closest

tower and that a cell phone could connect to a tower as far six miles away, but the

testimony generally supported that a cell phone would utilize a tower within a mile

to a mile and a half from where the cell phone was located.3 We acknowledge the

probative value of the cell phone evidence, but note that this evidence had its

3 The dissent contends that we do not give adequate weight to the probative value of this cell phone evidence.

11 weaknesses, and in conjunction with the weaknesses of the eyewitness testimony

detailed above, we are still left with “grave doubt” as to whether the error of

Allen’s admission influenced the jury’s decision. See id.

Moreover, the prosecution’s case is further compromised by the lack of

fingerprint evidence placing Palmer in the vehicle. Although an absence of

fingerprint evidence might not have been notable in a case where no fingerprints

were found, here, over thirty prints were found in the car, and from those prints,

Childress’s and Allen’s prints were identified, but no prints were ever matched to

Palmer.

Admission of Allen’s confession, which violated Palmer’s constitutional

right to confrontation, filled in all the holes of the state’s case. Allen’s confession

shored up the inconsistencies in the critical eye witness testimony and resolved

doubts relative to the cell phone evidence. Allen also established that there were

three people in the car with his use of pronouns. From this, the jury only had to

look at the three defendants sitting in the courtroom to find Palmer guilty as well.

On this basis, we conclude that the first factor strongly favors finding that Allen’s

testimony had a substantial and injurious effect on the jury’s verdict.

2. Whether the testimony was cumulative

This factor likely tilts in favor of the state. If the jury believed Love and

Miralda’s identifications, Allen’s testimony was duplicative.

12 3. The presence or absence of evidence corroborating or contradicting the testimony of the witness on material points Proceeding to the third factor, the cell phone evidence is useful evidence

corroborating Allen’s testimony. But again, the issues with the eyewitness

testimony indicate the absence of probative corroborating evidence. Accordingly,

we conclude this factor is neutral for the reasons discussed above.

4. The extent of cross-examination otherwise permitted

This factor weighs heavily in Palmer’s favor. Because Allen did not take the

stand, cross-examination did not occur. Palmer was denied the opportunity to

expose the highly unreliable nature of a codefendant’s testimony to the jury

through cross-examination. Thus, we find this factor weighs in Palmer’s favor.

5. The overall strength of the prosecution’s case

On the last factor, we also conclude it weighs in favor of Palmer. As

assessed under factor one, the prosecution’s case against Palmer had significant

weaknesses. The case consisted of two eye witness identifications, the sweatshirt,

and cell phone evidence. While the cell phone evidence was probative, we have

discussed the weaknesses of both the eye witness identifications and the sweatshirt

evidence, and conclude that those weaknesses were significant. Moreover, Palmer

offered expert testimony explaining that eye witness identifications are particularly

unreliable under high-stress and life-threatening circumstances. Both Miralda and

Love’s identifications were made under stress and at gun point.

13 We conclude that the prosecution’s case against Palmer heavily relied upon

Allen’s admission. In light of the significant issues with the rest of the

prosecution’s evidence against Palmer, we hold that the record reveals “grave

doubt” as to whether the error in admitting Allen’s admission influenced the jury’s

decision to find Palmer guilty. See id. Because of this “grave doubt,” we conclude

that the error here had a “substantial and injurious effect” on the jury’s verdict.

See id. at 2199. Under Brecht and the actual prejudice standard, we afford relief

where we conclude that error had such an effect. See id.

In light of this conclusion, we turn to the California Court of Appeal’s

conclusion that the introduction of Allen’s testimony was harmless beyond a

reasonable doubt because the evidence against Palmer was “overwhelming.” The

record shows that the state court’s finding was objectively unreasonable. The

evidence was not overwhelming, but rather, was weak in critical areas. Allen’s

testimony was damning to Palmer and shored up the weaknesses in the

prosecution’s case. Accordingly, we conclude that the California Court of

Appeal’s decision was unreasonable under § 2245(d)(2). We reverse the district

court with instructions to grant the writ of habeas corpus.

REVERSED.

14

FILED

Palmer v. Davey, 15-55222

JUN 29 2018

GOULD, Circuit Judge, concurring in part and dissenting in part: MOLLY C. DWYER, CLERK

U.S. COURT OF APPEALS

I concur in so far as the majority concludes that Palmer's right to

confrontation was violated when the trial court admitted the confession of Allen

without properly redacting it. As a realistic matter I believe the jury had to know

that Palmer was being incriminated by testimony on which he lost his ability to

cross examine because Allen never testified at Palmer's trial. This was what the

confrontation clause was designed to prohibit. On the other hand, I dissent in part

insofar as the majority concludes that the state court was objectively unreasonable

in finding that the error was harmless beyond a reasonable doubt. I believe the

majority's reasoning on the harmless error issue is at odds with the requirements of

the AEDPA and controlling precedents of the United States Supreme Court.

It is settled beyond any doubt that when a state court has made a harmless

error determination that is a decision on the merits. Davis v. Ayala, 135 S. Ct.

2187, 2198 (2015). We must affirm the decision that the constitutional error was

harmless unless that decision itself was objectively unreasonable. Id. at 2199. That

is the cost of our system recognizing the values of federalism and giving some

sway to a state appellate court decision such as was made in this case. The overall

standard under AEDPA is quite forgiving of state appellate courts. As the Supreme

1 Court has reminded us, time and again, a decision is not unreasonable if

"fairminded jurists could disagree on the correctness of the state court’s decision."

Harrington v. Richter, 562 US 86, 101 (2011) (citation and internal quotation

marks omitted).

Here, while someone might conclude that the evidence against Palmer was

not overwhelming, as the majority does here in characterizing the government's

case as weak, I do not see how it can be said that fair-minded jurists could not

disagree about whether the confrontation clause error was harmless beyond a

reasonable doubt. To mention just the key facts that persuade me that the state

court decision of harmless error was a reasonable one, here there were two

eyewitnesses who identified Palmer as being in the car that performed the drive-by

killing. Although one might dispute the strength of these identifications, the jury

certainly had the ability to weigh the facts that persuade our majority, and the jury

convicted Palmer. Not only were there two witnesses who described Palmer as

being a shooter in the car, there were also two pieces of hard evidence that confirm

this. One was that witnesses identified someone in the car as having a shirt with the

same color pattern as one found in Palmer's apartment. The other was that Palmer’s

cell phone was called twice at approximately the time of the murder, and this was

weighty evidence against him, because it placed him near the scene of the crime.

2 In my view, by arguing that this cell phone evidence most likely only put

Palmer within a mile to a mile and a half of the cell tower, and theoretically only

within six miles of the cell tower, the majority does not give adequate weight to the

probative value of the cell phone evidence. This evidence is more probative than

the majority’s theory would have it. Testimony at trial supported the conclusion

that Palmer would have been within a 60 degree northward facing arc of the cell

towers, which included the particular shooting site that was the focus of the

charged crime. There was also testimony supporting the conclusion that much of

the area within one to one and a half miles of the towers and within the 60 degree

northward facing arc of the cell tower would have been territory of the rival 18th

Street gang. And there was testimony that no members of the P-Stones would have

been in that territory unless they were looking for a confrontation. The Baldwin

Village area where the P-Stones congregate was not within the 60 degree

northward facing arc of the cell tower.

The majority has misread and minimized the incriminating evidence in a

way that the state appellate court did not. I do not see how the state court’s

contrary decision could be viewed as objectively unreasonable. The cell phone

evidence does not specifically place Palmer in the car at the time of the shooting,

but it supports the conclusion that he was in rival gang territory at that time, which

3 in turn supports that he was seeking confrontation of the type that occurred in the

shooting. When combined with the eye-witness identifications and the red and gray

sweater evidence, the case against Palmer is stronger than the majority describes it,

and I conclude that the state presented a strong case wholly apart from the poorly

redacted transcript of Allen’s statement.

In summary, fair-minded jurists on the state appellate court could conclude

any confrontation clause error was harmless beyond a reasonable doubt under the

Chapman test. See Chapman v. California, 386 U.S. 18, 24 (1967). Alternatively,

under the Brecht test used on collateral review of state court harmlessness

determinations, the confrontation clause error was harmless because, given the

strength of the evidence against Palmer, it does not justify having grave doubts

about whether the error caused a substantial injurious effect or influence on the

jury’s decision.1 See Jones v. Harrington, 829 F.3d 1128, 1141 (9th Cir. 2016);

1 The Brecht test is more deferential to the state court decision than the AEDPA/Chapman test. Under the AEDPA/Chapman test a federal habeas court asks whether the state court’s determination that an error was harmless beyond a reasonable doubt was objectively unreasonable. Under Brecht a federal habeas court asks whether there are grave doubts as to whether the error had a substantial and injurious effect or influence on the jury’s decision. Harrington, 829 F.3d at 1141. As the Supreme Court has explained, the AEDPA/Chapman test is subsumed under the Brecht test. Fry v. Pliler, 551 U.S. 112, 120 (2007), and that Brecht test was applied by the majority here, though I disagree with its application. 4 Brecht v. Abrahamson, 507 U.S. 619, 623 (1993). For these reasons I conclude that

we should affirm and I respectfully dissent.

5