LAZAR v. COLEMAN et al

Filing 25

MEMORANDUM AND/OR OPINION. SIGNED BY HONORABLE GERALD A. MCHUGH ON 3/1/17. 3/1/17 ENTERED AND COPIES E-MAILED. (jpd)

IN THE UNITED STATES DISTRICT COURT
FOR THE EASTERN DISTRICT OF PENNSYLVANIA

STEVEN LAZAR,
Petitioner,
v.
BRIAN V. COLEMAN,
Respondent.

:
:
:
:
:

CIVIL ACTION
No. 14-6907

MCHUGH, J.

MARCH 1, 2017
MEMORANDUM

Judicial recognition of a confession’s power as evidence dates back centuries: “[A] free
and voluntary confession is deserving of the highest credit, because it is presumed to flow from
the strongest sense of guilt; and therefore it is admitted as proof of the crime to which it refers.”
King v. Warickshall (1783) 168 Eng. Rep. 234, 234–35 (KB). Modern courts still recognize that
a “confession is like no other evidence.” Arizona v. Fulminante, 499 U.S. 279, 296 (1991). In
this habeas case, Petitioner Steven Lazar was interrogated by police over a 30-hour period about
a homicide, giving two different accounts of the crime. The second account, which was critical
to Lazar’s ultimate conviction for felony murder, was given while Lazar was in acute methadone
withdrawal. At trial, both the jury and court were unaware of this, because, unfortunately,
defense counsel stipulated that Lazar had not taken a daily dose of methadone for a month prior
to the interrogation. In fact Lazar had taken methadone a mere 24 hours before.
Given that incorrect stipulation, it is hardly surprising that on collateral review the state
courts found Lazar’s trial counsel’s performance deficient, satisfying the first of two elements of
an ineffective-assistance claim under Strickland v. Washington, 466 U.S. 668 (1984). But the
courts decided that the second element—prejudice—had not been met, and therefore denied
relief. The well-reasoned Report and Recommendation of United States Magistrate Judge
1

M. Faith Angell also concludes that federal relief should be denied. I am constrained to agree
with Judge Angell and will adopt that part of her Recommendation. But I base my denial of
relief solely on the exceedingly deferential framework imposed on me by 28 U.S.C. § 2254(d).
And I write separately to stress my serious concern about the state courts’ prejudice analysis and
to explain why I will grant a certificate of appealability.
I.

Background
A. Facts at Trial
In January 2007, an elderly man was found in his North Philadelphia home hacked to

death. His wallet and keys were missing and his drawers ransacked. The case went unsolved at
first, but a few months later, the victim’s daughter, while cleaning out the back porch of her
father’s home, found a bag with several items linked to Lazar: a letter to him from a methadone
clinic, a yearbook with his photograph, his social security card, and an access card in his name.
When police brought Lazar in for questioning, he told them his friend had taken his bag and
given it to another man. Lazar was released.
But over the next several months, Lazar made statements to others that seemed to
implicate him in the murder. For example, he told two of his neighbors that he would be going
to jail because he had “killed somebody with an axe,” and he told one of his friends that “if you
are going to kill somebody,” the “best weapon to use . . . [is] a hatchet.” He made similar
statements to two other individuals. All five testified at trial. (As discussed in more detail
below, each was impeached to some degree, and none would be considered a pristine witness for
the prosecution.) That November, after the friend contacted police about what Lazar had told
him, homicide detectives interrogated Lazar over a 30-hour period.

2

During the interrogation, Lazar gave two statements. First, at 7:20 PM on November 19
(the first day of the interrogation) he said that on the night of the murder, he and a friend
(“John”) had been doing drugs in an abandoned house near the victim’s. John had a hatchet and
talked about robbing the victim. After police drove by, the two briefly separated, and Lazar then
went into the victim’s home to look for John. He found John covered in blood, ransacking the
drawers, and the victim dead. But Lazar’s second statement, given at 2:45 PM on November 20,
told a different story. In this version, he and John had gone to the victim’s house to perform oral
sex on him—which, after John pulled a hatchet out to threaten the victim, they did. When the
victim tried to reciprocate, Lazar hit the victim with his hand, and John hit the victim with the
hatchet. John then took the victim’s wallet and went through his drawers.
The jury convicted Lazar of second-degree murder, robbery, and possession of an
instrument of crime.
B. Essential Facts of Lazar’s Strickland Claim
At trial, Lazar’s counsel stipulated that Lazar, a known methadone user, had last taken a
daily dose of methadone on October 24, nearly a month before the interrogation. That
stipulation was wrong, in a significant way: Lazar had actually last taken methadone at
10:34 AM on November 18, the day before the interrogation began. 1 And shortly after giving

1

Methadone is categorized as an “essential” medicine by the World Health Organization.
Though one might assume that withdrawal from a medication used to treat addiction is less
traumatic than withdrawal from an abused substance itself, that assumption is problematic.
Highly respected addiction specialists first questioned it over 25 years ago. See generally
Michael Gossop & John Strang, A Comparison of the Withdrawal Responses of Heroin and
Methadone Addicts During Detoxification, 158 Brit. J. Psychiatry 697 (1991). And a recent
empirical study surveying 215 inmates reflected that 70 percent would rather endure unsupported
withdrawal from heroin as compared to methadone, and that the trauma of methadone
withdrawal following arrest was so severe as to discourage future participation in methadone
treatment. See generally Jeannia J. Fu et al., Forced Withdrawal from Methadone Maintenance
3

the second statement, Lazar complained of painful urination and was taken to the hospital, where
he reported leg pain, shaking, dizziness, nausea, and diarrhea, and blood tests were positive for
multiple anti-withdrawal benzodiazepines, cocaine, and marijuana. Notations in the hospital
records, however, painted a slight contrast, describing Lazar’s appearance as “neat and
appropriate” and “oriented,” his demeanor “irritable but cooperative,” and his speech “normal in
tone, rate, and rhythm.” PCRA Hr’g Tr. (Tr.) 12:23–13:5. He was diagnosed with depression
and hypertension and discharged that night after being treated with anti-withdrawal drugs. 2
C. State-Court Collateral Review
On post-conviction review of Lazar’s Strickland claim, the state trial court agreed with
Lazar that his counsel was deficient in making the methadone-dosage-date stipulation, but found
there was no prejudice. In addition to considering the evidence before the jury, the court heard
competing expert testimony on the effects of methadone withdrawal and observational testimony
from the interrogating detectives.
Both experts agreed that an individual suffering from withdrawal could still give a
voluntary statement, but they disagreed about how fast withdrawal symptoms (such as nausea,
runny eyes and nose, and restlessness) set in: Lazar’s expert said it happened within a 24–36
hour window, while the Commonwealth’s expert said it took longer, up to 50 hours. The
detectives, for their part, testified that Lazar did not appear sick just before he was taken to the
hospital. The court, however, gave the most weight to the notations in the hospital records.
The court concluded that if the jury were armed with all of that new evidence and the
correct methadone-dosage date, there was still no reasonable probability that the outcome of the
Therapy in Criminal Justice Settings: A Critical Treatment Barrier in the United States, 44 J.
Substance Abuse Treatment 502 (2013).
2

The jury heard the evidence relating to Lazar’s hospital visit, see Trial Tr. (May 10,
2010) 180:19–183:21, but without the context of Lazar being in acute withdrawal.
4

trial would have been different. As to Lazar’s confession, the court stated: “At best, [the jury]
would have heard from competing experts that the defendant was experiencing an unknown
quantum of withdrawal symptoms during the taking of his statements.” Tr. 15:8–11. Moreover,
the court found that even disregarding Lazar’s confession, there was “overwhelming evidence of
guilt,” Tr. 15:20–21—namely, the implicating statements he made to the five individuals who
testified at trial.
The Superior Court affirmed, and the Pennsylvania Supreme Court denied Lazar’s
petition for review.
II.

Standard of Review
Among its other restrictions on habeas corpus, the Antiterrorism and Effective Death

Penalty Act of 1996 (AEDPA) severely limits a federal court’s ability to grant habeas relief as to
“any claim that was adjudicated on the merits in State court.” 28 U.S.C. § 2254(d). When
reviewing such a claim, a federal court may not grant relief unless, as relevant here, the state
court’s decision “involved an unreasonable application of” Supreme Court precedent, id.
§ 2254(d)(1)—in this case, Strickland and its progeny.
The Supreme Court has concluded that § 2254(d) “imposes a ‘highly deferential standard
for evaluating state-court rulings,’ and ‘demands that state-court decisions be given the benefit of
the doubt.’” Renico v. Lett, 559 U.S. 766, 773 (2010) (citations omitted). For my purposes,
then, the question is not whether the state court decisions were right or wrong. See White v.
Woodall, 134 S. Ct. 1697, 1702 (2014). Rather, the question is whether they were “so lacking in
justification that there was an error well understood and comprehended in existing law beyond
any possibility for fairminded disagreement.” Harrington v. Richter, 562 U.S. 86, 103 (2011).
Only then can I decline to defer to them.

5

III.

Discussion
Because I cannot say that the state courts here—in deciding that, despite Lazar’s

counsel’s deficiencies, there was no prejudice—made an error clear “beyond any possibility for
fairminded disagreement,” I must deny federal relief. But I remain troubled by that result.
To show Strickland prejudice, the “defendant must show that there is a reasonable
probability that, but for counsel’s unprofessional errors, the result of the proceeding would have
been different. A reasonable probability is a probability sufficient to undermine confidence in
the outcome.” 466 U.S. at 694. Here, unlike the state courts that denied relief, I would have
found that such a “reasonable probability” exists—because the tainted evidence is a confession.
Embracing the insight of eighteenth-century British jurists, American courts in an
unbroken line of cases have recognized that a voluntary confession has “always ranked high in
the scale of incriminating evidence.” Bram v. United States, 168 U.S. 532, 544 (1897) (quoting
Brown v. Walker, 161 U.S. 591, 596 (1896)). It is considered “evidence of the most satisfactory
character” and “among the most effectual proofs in the law.” Hopt v. Utah, 110 U.S. 574, 584–
85 (1884). Even as other rules of law have changed dramatically over time, that core principle—
that a defendant’s voluntary confession has a “profound impact on the jury” and “is probably the
most probative and damaging evidence that can be admitted against him,” Fulminante, 499 U.S.
at 296—has gone unchanged.
That probative force, of course, applies only to voluntary confessions, for the
presumption that “one who is innocent will not imperil his safety or prejudice his interests by
making an untrue statement” does not lie if he has been “deprive[d] . . . of that freedom of will or
self-control essential to make his confession voluntary.” Hopt, 110 U.S. at 585. For many years,
however, voluntariness was the only touchstone, because it was thought that a voluntary yet false

6

confession was “scarcely conceivable.” Brandon L. Garrett, The Substance of False
Confessions, 62 Stan. L. Rev. 1051, 1052 (2010) (citation omitted). But since “the experience of
the courts, the police, and the medical profession recounts a number of false confessions
voluntarily made,” Smith v. United States, 348 U.S. 147, 153 (1954), it has become clear that
relying solely on voluntariness is not enough: a confession must also be tested for reliability.
Indeed, in a case like this, where it seems probable that Lazar’s confession was voluntary
but withdrawal-tainted, that reliability backstop becomes even more important. It is not a new
idea that someone whose mental and emotional faculties are impaired could freely give a false
confession. Nearly a century ago, Dean Wigmore noted that the usefulness of a confession
largely “depends upon the mental and emotional traits of the accused”: “We may believe that
rationally a false confession is not to be apprehended from the normal person . . . [,] but we have
here perhaps a person not to be tested by a normal or rational standard.” 3 Wigmore, Evidence
§§ 820a–b (Chadbourn rev. 1970). The Supreme Court has similarly observed that “under
certain stresses a person, especially one of defective mentality or peculiar temperament, may
falsely acknowledge guilt,” In re Gault, 387 U.S. 1, 44–45 (1967) (quoting 3 Wigmore, Evidence
§ 822 (3d ed. 1940)), and that “the physical and psychological environment that yielded the
confession can . . . be of substantial relevance to the defendant’s guilt or innocence,” Crane v.
Kentucky, 476 U.S. 683, 689 (1986). Further, where (as here) the confession was given during a
“prolonged and persistent” interrogation, that “count[s] heavily against the confession.” Ashcraft
v. Tennessee, 322 U.S. 143, 161–62 (1944).
Consider then the state post-conviction courts’ decisions in this case. Those courts found
no Strickland prejudice—that is, no “reasonable probability” that, if the new methadonewithdrawal evidence had been before the jury, “the result of the proceeding would have been

7

different.” Given what is now known about false confessions, and the fundamental precept that a
confession’s “probative weight . . . [is] a matter that is exclusively for the jury to assess,” Crane,
476 U.S. at 688, that seems like a difficult conclusion to support.
As in every case, only the members of the jury that convicted Lazar could say on what
evidence they placed the most weight. And the confession hardly stood alone, as there was also
the testimony of the five witnesses to whom Lazar made (seemingly implicating) statements,
plus his bag at the scene. But three items stand in counterbalance. First, juries place significant,
often dispositive, emphasis on confessions. As a result, where the evidence tainted by deficient
counsel performance is a confession, courts should hesitate—or, to borrow a phrase, “think hard,
and then think hard again,” Camreta v. Greene, 563 U.S. 692, 707 (2011)—before finding there
was no prejudice. Second, that general principle has specific application in this case, where the
five civilian witnesses all had various credibility problems: one, Lazar’s friend, had a theft
conviction and had contacted the police about Lazar’s statements only after getting into a dispute
with Lazar; two others had pending robbery or theft cases against them at the time of trial; and
the last two changed their stories on the stand. Thus to the extent courts should ever feel
comfortable finding the untainted evidence to be “overwhelming” and sufficient grounds for
finding no prejudice, this does not seem like the right case for it. Finally, the Commonwealth’s
heavy reliance on the confession, throughout trial and specifically in closing argument, speaks
volumes about the confession’s importance to the case. Cf. United States v. Brownlee, 454 F.3d
131, 148 (3d Cir. 2006) (“[I]t is difficult for the Government to argue with effect that the
admission of the confession did not contribute to Brownlee’s conviction when it submitted just
the opposite view to the jury during the trial.”). In my view, these factors strongly suggest that
the deficient performance of Lazar’s trial counsel was prejudicial.

8

More generally, I believe Strickland prejudice is an area where courts should tread with
special care. An important distinction needs to be drawn between Strickland’s two prongs—
counsel’s performance and prejudice. When a federal court reviews a claim regarding counsel’s
performance that was decided on the merits in state court, thus triggering deference under
§ 2254(d), review is said to be “doubly” deferential. See Harrington, 562 U.S. at 105 (quoting
Knowles v. Mirzayance, 556 U.S. 111, 123 (2009)). This is because § 2254(d) requires a federal
court to give the state court deference, and Strickland required the state court to give counsel
deference.
But there is an inherent tension between this “layered reasonableness” standard of review
and the legal precepts that define Strickland’s second prong: prejudice. That tension exists
because of the test for prejudice: to find no prejudice, a court must find there is no reasonable
probability that, without counsel’s deficiencies, the result of the proceeding would have been
different—a proceeding where the prosecution must prove its case beyond a reasonable doubt.
In other words, a collateral-review court must conclude there is no “reasonable probability that at
least one juror,” Wiggins v. Smith, 539 U.S. 510, 537 (2003), if given the new information
withheld because of counsel’s errors, would have reached anything less than a “subjective state
of certitude of the facts in issue,” In re Winship, 397 U.S. 358, 364 (1970) (citation omitted).
The Supreme Court recently reaffirmed this principle in its latest Strickland decision, with Chief
Justice Roberts identifying the prejudice inquiry as whether there is a “reasonable probability
that . . . at least one juror would have harbored a reasonable doubt.” Buck v. Davis, No. 15-8049,
2017 WL 685534, slip op. at 17 (U.S. Feb. 22, 2017).
Taken seriously, this is a high standard to meet. At a conceptual level, it reinforces the
principle that a jury cannot convict someone of a crime unless all twelve jurors are persuaded

9

beyond a reasonable doubt. In practical terms, it underscores the importance of a jury deciding a
criminal case only after reviewing all relevant evidence. When that evidence relates to a tainted
confession, the defendant’s inability to put it before the jury “strips [him] of the power to
describe to the jury the circumstances that prompted his confession” and “effectively disable[s
him] from answering the one question every rational juror needs answered: If [he] is innocent,
why did he previously admit his guilt?” Crane, 476 U.S. at 689. So I hesitate to conclude that if
Lazar’s jury had known he was suffering from methadone withdrawal while under interrogation,
there is “no reasonable probability that at least one juror” might have had a reasonable doubt.
I recognize that some courts of appeals have applied the double-deference standard to
both counsel’s performance and prejudice. See, e.g., Frazier v. Bouchard, 661 F.3d 519, 534
(11th Cir. 2011); Foust v. Houk, 655 F.3d 524, 534 (6th Cir. 2011). The Supreme Court, without
explicitly holding so, also seems to assume double deference applies to prejudice. See, e.g.,
Cullen v. Pinholster, 563 U.S. 170, 202 (2011).
But applying extreme deference to state-court determinations of prejudice seems to not
only be inconsistent with Supreme Court precedent otherwise defining prejudice—it actually
undermines its vitality. The test for prejudice as defined by Wiggins—is there a “reasonable
probability that at least one juror” would have had a reasonable doubt?—has a broad reach, and
errs on the side of finding prejudice when a defendant did not receive a fair trial. But the
protection conferred by Wiggins (and other cases defining prejudice broadly) evaporates once a
state court answers that question “no,” because of the need to show not just that the state court’s
conclusion is wrong, but wrong “beyond any possibility of fairminded disagreement.” In a
disturbing irony, then, a defendant deprived of the right to have the entirety of his case

10

examined through the prism of reasonable doubt is then deprived of collateral relief because he
cannot meet a similarly onerous standard. 3
IV.

Conclusion
Because of these profound concerns, although I feel constrained to deny Lazar’s habeas

petition, I will issue a certificate of appealability. 4 An appropriate order follows.

/s/ Gerald Austin McHugh
United States District Judge

3

I would adopt the position advanced by Judge Jordan of the Eleventh Circuit in his
concurring opinion in Evans v. Secretary, Department of Corrections, 703 F.3d 1316, 1333–36
(11th Cir. 2013) (en banc), and not apply “double deference” to the issue of prejudice.
4

“A certificate of appealability may issue . . . if the applicant has made a substantial
showing of the denial of a constitutional right.” 28 U.S.C. § 2253(c)(2); Fed. R. App. P. 22(a)
(district judge may issue one). “That standard is met when ‘reasonable jurists could debate
whether . . . the petition should have been resolved in a different manner.’” Welch v. United
States, 136 S. Ct. 1257, 1263 (2016) (quoting Slack v. McDaniel, 529 U.S. 473, 484 (2000)). For
reasons already given I find that standard met here, and so will grant a certificate of appealability
on the issue whether the state courts in this case unreasonably applied Strickland prejudice.
11