Murillo, Edward A. v. Frank, Matthew

In the

United States Court of Appeals

For the Seventh Circuit

____________

No. 04-2202

EDWARD A. MURILLO,

Petitioner-Appellee,

v.

MATTHEW J. FRANK, Secretary, Wisconsin

Department of Corrections,

Respondent-Appellant.

____________

Appeal from the United States District Court

for the Eastern District of Wisconsin.

No. 01-C-1285—William C. Griesbach, Judge.

____________

ARGUED JANUARY 18, 2005—DECIDED APRIL 1, 2005

____________

Before POSNER, EASTERBROOK, and ROVNER, Circuit

Judges.

EASTERBROOK, Circuit Judge. Santiago Herrera was shot

and killed in 1998. He was a casualty of a conflict between

the Latin Kings (Herrera’s gang) and La Familia about

what in antitrust law would be called exclusive retail sales

territories. Zebulon Robinson, who had been negotiating to

purchase marijuana from Herrera, saw the murder and at

trial identified as the shooter Edward Murillo, one of three

members of La Familia who had approached Herrera and

2 No. 04-2202

proclaimed that La Familia owned the territory. Robinson’s

word was of uncertain reliability—he had been engaged in

an illegal drug transaction, his recollection of events

changed over time, and, worse, he had taken the murder

weapon to a friend to hide, though he testified that the

assassin had tossed him the gun while fleeing. So the

prosecutor wanted additional evidence. The state called Luis

Murillo, Edward’s brother. But Luis refused to testify, even

after receiving immunity from prosecution and being held

in contempt for his intransigence. With Luis unavailable,

the prosecutor offered—and the judge admitted, over an

objection based on the Constitution’s confrontation

clause—a statement that Luis had made during custodial

interrogation six days after Herrera’s death.

Luis told the police a series of lies. First he denied being

anywhere near the murder. He had been with his girlfriend

at the time, Luis asserted. An officer checked out the story

(or pretended to) and told Luis that his alibi did not hold

up. Moreover, the officer said, a Crime Stoppers tip had

placed Luis at the scene. Luis then said that he had indeed

been nearby and had seen the perpetrators running away,

but that he had nothing to do with the crime. During the

interrogation Luis became increasingly upset and nervous,

breaking into tears when the officers refused to accept his

version of events. After still more attempts at evasion, Luis

stated that he had seen his brother Edward approach

Herrera and pull the trigger. Luis signed an affidavit to

that effect. But Luis denied being one of the triad who had

approached Herrera; how he could have seen the events so

clearly, if he had not been part of the delegation, he did not

reveal. (Luis said in his statement and affidavit that

Robinson was the third member of the group; Robinson tes-

tified that Luis was the third.)

The jury convicted Edward of first degree murder and two

firearms offenses. He was sentenced to life imprisonment.

Wisconsin’s intermediate appellate court affirmed, rejecting

No. 04-2202 3

his argument under the confrontation clause, see 240 Wis.

2d 666, 623 N.W.2d 187 (2000), and the Supreme Court of

Wisconsin declined to hear the case. Edward fared better on

federal collateral review under 28 U.S.C. §2254. The district

court issued a writ of habeas corpus in his favor, concluding

that the state judiciary’s decision was contrary to Lilly v.

Virginia, 527 U.S. 116 (1999), a decision that the district

judge viewed as on all fours, right down to fact that the

declarant was the defendant’s brother. 316 F. Supp. 2d 744

(E.D. Wis. 2004). Wisconsin contends in this court that Lilly

is distinguishable because the statement there had been

admitted on the ground that it was against the declarant’s

penal interest, while admission of Luis’s affidavit rests on

the ground that it undercut his social interest. Wisconsin is

among those states (a distinct minority) that deem a dec-

laration against “social interest”—that is, a statement that

shames the speaker in his community—an exception to the

hearsay doctrine. See Wis. Stat. §908.045(4); Thomas J.

Imwinkelreid, Declarations Against Social Interest: The

(Still) Embarrassingly Neglected Hearsay Exemption, 69 S.

Cal. L. Rev. 1427 (1996). The idea is that Luis’s statement

violated the social norm of solidarity among gang (and family)

members, so that Luis would not have accused Edward un-

less the accusation were true.

Wisconsin’s judiciary used the framework of Ohio v.

Roberts, 448 U.S. 56, 66 (1980), which stated that hearsay

is compatible with the confrontation clause when supported

by (1) a “firmly rooted hearsay exception” or (2) “particular-

ized guarantees of trustworthiness”. The state acknowl-

edges that the “social interest” doctrine is novel rather than

established by force of history but contends that it satisfies

Roberts’s second option and insists that Lilly does not hold

otherwise.

If this question were to arise today, the governing decision

would be Crawford v. Washington, 124 S. Ct. 1354 (2004),

rather than Roberts or Lilly. The Court held in Crawford

4 No. 04-2202

that, when testimonial declarations (such as affidavits or

formal confessions) are at issue, judicial assessments of reli-

ability never suffice and that an opportunity to cross-ex-

amine the declarant is essential. Edward was unable to

cross-examine Luis, so Crawford amounts to a per se rule

that his statement and affidavit were inadmissible. But

Crawford was not issued until after Edward’s conviction

became final through the conclusion of direct review, and

the district court held that it could not be applied retroac-

tively on collateral attack. 316 F. Supp. 2d at 749-50.

Edward defends the judgment in his favor by asking us to

apply Crawford, which would avoid any need to consider

Roberts and Lilly.

The parties debate whether Crawford applies retroactively

under Teague v. Lane, 489 U.S. 288 (1989), and its succes-

sors—most recently, Schriro v. Summerlin, 124 S. Ct. 2519

(2004), and Beard v. Banks, 124 S. Ct. 2504 (2004). Section

2254(d)(1) says that the writ shall not issue unless the state

court’s decision “was contrary to, or involved an unreason-

able application of, clearly established Federal law, as

determined by the Supreme Court of the United States”. A

natural reading of §2254(d)(1) is that the only federal law

that matters is the law that had been “clearly established”

when the state court is called on to decide. See Williams v.

Taylor, 529 U.S. 362, 390, 412 (2000); Gosier v. Welborn,

175 F.3d 504, 510 (7th Cir. 1999) (Ҥ2254(d)(1) means that

only rules articulated by the Supreme Court of the United

States before the state court rendered its decision may be

applied on collateral review. Section 2254(d)(1) differs from

Teague because the new statute closes the escape hatches

in Teague”) (emphasis in original). See also Ramdass v.

Angelone, 187 F.3d 396, 406 (4th Cir. 1999); Randy Hertz &

James S. Liebman, 2 Federal Habeas Corpus Practice &

Procedure §32.3 at 1430 (4th ed. 2001).

There is a provision for giving effect to retroactive legal

changes, but it lies in §2244(b)(2)(A) and depends on the

No. 04-2202 5

Supreme Court’s own declaration that its decision meets the

Teague standard. See, e.g., Tyler v. Cain, 533 U.S. 656

(2001). The Supreme Court has not held that Crawford

applies retroactively on collateral review. Because the

Supreme Court rarely decides retroactivity as part of the

original opinion, reading §2254(d)(1) to prevent any other

court from making the retroactivity decision would mean

that the year allowed to initiate a collateral proceeding often

would expire. The Supreme Court may decide in Dodd v.

United States, cert. granted, 125 S. Ct. 607 (2004) (argued

March 22, 2005), whether prisoners bear the risk of delay

until a retroactivity decision can be made. While waiting for

the outcome of Dodd, we deem it best to address how

Crawford is classified under Teague. Unless Crawford meets

the Teague standard for retroactivity, whether §2254(d)(1)

allows a court of appeals to make its own retroactivity

assessment is unimportant.

Teague and its successors say that a new rule of con-

stitutional law is retroactive on collateral attack only if

it places certain conduct beyond the reach of the criminal

law (that is, establishes that the defendant’s acts were not

subject to punishment) or if it establishes one of the rare

“watershed rules of criminal procedure implicating the fun-

damental fairness and accuracy of the criminal proceeding.”

O’Dell v. Netherland, 521 U.S. 151, 157 (1997). Three courts

of appeals have held that Crawford is not retroactive under

this approach. Mungo v. Duncan, 393 F.3d 327 (2d Cir.

2004); Dorchy v. Jones, 398 F.3d 783, 788 (6th Cir. 2005);

Brown v. Uphoff, 381 F.3d 1219, 1226-27 (10th Cir. 2004).

One court has held that Crawford is retroactive— though

not by application of the Teague standard. Bockting v.

Bayer, 2005 U.S. App. LEXIS 3012 (9th Cir. Feb. 22, 2005).

The three judges in Bockting wrote separately. Judge Noonan

concluded that Crawford applies on collateral review be-

cause it did not change the law. Judge McKeown concluded

that Crawford did change the law, and changed it so dra-

6 No. 04-2202

matically that it established a “watershed rule” that applies

retroactively. Judge Wallace agreed with Judge McKeown

that Crawford changed the law but disagreed with her un-

derstanding of Crawford’s fundamentality. He saw it as an

ordinary development in criminal procedure that like almost

all other such changes applies prospectively. We agree with

this perspective and thus follow Mungo, Dorchy, and Brown.

It is obvious to us—as it was to a prior panel, see Owens

v. Frank, 394 F.3d 490, 501 n.8 (7th Cir. 2005)—that

Crawford establishes a new rule. It discards the framework

that Roberts had adopted. True enough, as Judge Noonan

observed, Crawford did not say that it was overruling

Roberts; it emphasized that the declarant in Roberts had

been subject to cross-examination. But it assuredly (and

explicitly) jettisoned the Roberts standard. 124 S. Ct. at 1373.

All of the Supreme Court’s decisions between Roberts and

Crawford had applied that understanding, though some of

the Justices had questioned whether it should be main-

tained. Lilly provides a good example. Seven Justices asked

and answered the Roberts questions (disagreeing four to

three about their resolution); two wrote separately to stake

out a different position. A rule is “new” for retroactivity

analysis unless it was dictated by earlier decisions. See

Banks, 124 S. Ct. at 2511. Crawford was not “dictated” by

Roberts or Lilly; it broke from them. That the break takes

the form of a return to an older, less flexible but historically

better grounded approach does not make it less a break. All

constitutional decisions find their ultimate basis in texts

adopted long ago—here in the Bill of Rights (1791) and their

application to the states via the fourteenth amendment

(1868). Judicial rhetoric routinely invokes older norms. This

does not mean that there has been no “new rule” of con-

stitutional criminal procedure since 1868.

Whether Crawford adopts a fundamental rule essential to

a fair and accurate trial is a subject that we pretermitted in

Owens. Like the second, sixth, and tenth circuits (and Judge

No. 04-2202 7

Wallace in Bockting), we think the answer a straightforward

“no.” The Supreme Court has not identified any decision,

other than Gideon v. Wainwright, 372 U.S. 335 (1963), that

would be so profound. It has repeatedly declined invitations

to treat one or another decision as a “watershed rule,”

including both Banks and Summerlin last Term. Summerlin

holds that Apprendi v. New Jersey, 530 U.S. 466 (2000),

surely a more sweeping change than Crawford (and more

important to defendants, too, because it entitles them to a

jury decision, while Crawford affects only what evidence the

jury hears), is not retroactive on collateral attack. See also

McReynolds v. United States, 397 F.3d 479 (7th Cir. 2005).

Indeed, it would be a close question whether Crawford

helps or hinders accurate decisionmaking. Live testimony

is preferable to affidavits and transcribed confessions, be-

cause cross-examination can probe its weaknesses, but re-

corded testimony may be better than silence, when death or

incapacity or threats or loyalty to one’s confederates keep

witnesses off the stand. The point of Crawford is not that

only live testimony is reliable, but that the sixth amendment

gives the accused a right to insist on live testimony, whether

that demand promotes or frustrates accuracy. Like the self-

incrimination clause and other provisions in the Bill of Rights,

the confrontation clause can be invoked to prevent the

conviction of persons who are guilty in fact. What Crawford

holds is that defendants enjoy this right even when the

hearsay is trustworthy. This is not an indispensable

innocence-protecting decision that must be applied retroac-

tively to criminal prosecutions that have already been

finally resolved on direct review.

There is another way to see the point. Violation of a truly

vital rule of criminal procedure, such as entitlement to

counsel (the holding of Gideon) leads to reversal without

inquiry into harmless error. See United States v. Cronic,

466 U.S. 648 (1984). Violations of other, less fundamental

rules are subject to harmless-error analysis. The confronta-

8 No. 04-2202

tion clause is in the latter category; courts regularly

examine evidence admitted without cross-examination to

determine whether the error was harmless. See Delaware

v. Van Arsdall, 475 U.S. 673, 682 (1986). This shows, as the

tenth circuit observed in Brown, that Crawford cannot have

established the sort of indispensable doctrine that applies

retroactively even to closed cases.

Let us return, then, to how the Court understood and

applied the Roberts framework in Lilly. Virginia charged

Benjamin Lilly with homicide. The evidence against him

included a statement by his brother Mark, who (like Luis

Murillo) refused to testify at trial. Mark’s statement con-

ceded that he had participated in the crime by helping to

steal the gun that Benjamin used in the slaying, and that

he had been present when Benjamin shot the victim. The

statement was to this extent against Mark’s penal interest.

Mark went on, however, to say that Benjamin had commit-

ted the murder without his encouragement or approval. This

portion of the statement exculpated Mark at Benjamin’s

expense, and the Court held that it could not be used against

Benjamin. (We refer to Justice Stevens’s opinion as that of

“the Court.” Because Justices Scalia and Thomas proposed

to exclude statements such as Mark’s categorically, the pos-

ition that was to prevail in Crawford, the plurality opinion

joined by Justices Stevens, Souter, Ginsburg, and Breyer was

the most narrow ground of decision and hence constitutes

the holding. See Marks v. United States, 430 U.S. 188, 193

(1977).)

The Court first concluded that declarations against penal

interest given during custodial interrogation (as opposed to,

say, declarations during and in furtherance of a conspiracy)

may be introduced by, but not against, an accused. 527 U.S.

at 131-34. This followed, Justice Stevens explained, from

many decisions, perhaps most prominently Bruton v. United

States, 391 U.S. 123 (1968), which had held that one defen-

dant’s confession that inculpates another may not be used

No. 04-2202 9

at a joint trial (unless the defendant who confessed also

testifies), because it deprives the co-defendant of an oppor-

tunity to cross-examine the confessing defendant. If the

confession, surely against the defendant’s penal interest,

must not be used when the declarant is in the courtroom

but refuses to testify, it cannot be used when the declarant

has not been charged with crime in the first place, the

Court held.

That left the question whether Mark’s statement could be

used nonetheless on the ground that its self-inculpatory

portions demonstrated its truthfulness. The Court held not,

527 U.S. at 135-39, in part because attempts to evade or

minimize blame by pointing fingers at others tend to be

unreliable and in part because “[i]t is highly unlikely that

the presumptive unreliability that attaches to accomplices’

confessions that shift or spread blame can be effectively

rebutted when the statements are given under conditions

that implicate the core concerns of the old ex parte affidavit

practice—that is, when the government is involved in the

statements’ production, and when the statements describe

past events and have not been subject to adversarial

questioning.” Id. at 137.

Everything the Court said about Mark’s statement in Lilly

is true of Luis’s statement too. The portion exculpating the

declarant (and inculpating the accused) is not a declaration

against penal interest, and at all events declarations made

during custodial interrogation cannot be used against an

accused. A statement, made during interrogation and blam-

ing someone else, also is too unreliable to supply the “parti-

cularized guarantees of trustworthiness” that until Crawford

could have supported admissibility. To the extent there is

any difference between Mark Lilly’s statement and Luis

Murillo’s, Luis’s is the less reliable, because he did not in-

culpate himself in any fashion; Luis professed to have had

nothing to do with the murder. If the confrontation clause

forbade Virginia to use Mark Lilly’s statement, how could

10 No. 04-2202

it allow Wisconsin to use Luis Murillo’s? The state court’s

decision is contrary to Lilly because it “confronts a set of

facts that are materially indistinguishable from a decision

of [the Supreme Court] and nonetheless arrives at a result

different from [that] precedent.” Williams, 529 U.S. at 406.

Instead of offering “particularized” (i.e., statement specific)

guarantees of trustworthiness, Wisconsin relied principally

on the social-interest doctrine itself, which is general rather

than particularized.

To the extent that it discussed the specifics of this situ-

ation, Wisconsin’s appellate court thought that Luis Murillo’s

statement was more reliable than Mark Lilly’s “because

Mark was up to his neck in criminal involvement” (240 Wis.

2d at 680) while Luis denied culpability. That’s backward.

Mark’s statement is the more reliable because he admitted

some crimes, which exposed him to prosecution. Luis

provided only one self-serving statement after another.

Wisconsin also maintains that Luis’s statement is more

reliable than Mark Lilly’s because Luis was emotionally

overwrought, while Mark was calculating. Again this “dis-

tinction” inverts the usual understanding of reliability.

We are not aware of any social science evidence supporting

the proposition that emotionality guarantees accuracy.

Jurors (and operators of polygraph machines) use sweating,

fidgeting, and other signs of emotional stress as cues that

the speaker is attempting deceit. Perhaps Luis was fright-

ened and crying because he feared that, unless he fingered

someone else, the state would put him on trial for the

murder, or because he feared that no matter what he said

the gang would suspect him of assisting the police and opt

for revenge.

Wisconsin’s “social interest” exception to the hearsay rule

is itself unsupported by any data of which we are aware; the

state has not cited any, either in this litigation or in the

decisions creating the exception. Gang members may boast

about their criminal exploits, or may adopt a code under

No. 04-2202 11

which they routinely testify (or confess) falsely in order to

throw the police off the scent and then refuse to follow

through at trial, exactly as Luis did. In such a subculture

false allegations of criminality, far from being shameful, are

normal. We need not pursue this subject, however. This un-

usual exception to the hearsay doctrine cannot support the

use of confessions and affidavits when the long-established,

and better supported, penal-interest exception does not.

Was the error harmless under Brecht v. Abrahamson, 507

U.S. 619 (1993)? The state says yes but does not supply

much argument. Wisconsin’s appellate brief devotes only

two pages to the subject, more than half of that to a recap

of legal rules. The brief’s single paragraph devoted to the

trial asserts that there was solid evidence independent of

Luis’s statement but does not make a serious effort to put

his statement in the context of the entire trial and assess

what effect it may have had. That makes it impossible to

say that Luis’s statement, which coming from the accused’s

own brother must have played a big role in jurors’ evalua-

tion, was unlikely to have had a substantial and injurious

effect on the verdict. See O’Neal v. McAninch, 513 U.S. 432

(1995). Why would the prosecutor have introduced Luis’s

statement, given the considerable risk even under Roberts,

unless he expected it to have punch and doubted the

sufficiency of other evidence? The state has not furnished a

sound basis for us to deem the error harmless.

AFFIRMED

12 No. 04-2202

A true Copy:

Teste:

________________________________

Clerk of the United States Court of

Appeals for the Seventh Circuit

USCA-02-C-0072—4-1-05