Cornell Reynolds v. Randall Hepp

In the
United States Court of Appeals
For the Seventh Circuit
____________________
No. 16‐3430
CORNELL D. REYNOLDS,
Petitioner‐Appellant,
v.
RANDALL HEPP,
Respondent‐Appellee.
____________________
Appeal from the United States District Court for the
Eastern District of Wisconsin.
No. 16‐CV‐508 — William C. Griesbach, Chief Judge.
____________________
ARGUED JANUARY 4, 2018 — DECIDED AUGUST 30, 2018
____________________
Before  WOOD,  Chief  Judge,  and  HAMILTON  and  BARRETT,
Circuit Judges.
HAMILTON,  Circuit  Judge.  Petitioner‐appellant  Cornell
Reynolds  seeks  a  writ  of  habeas  corpus  under  28  U.S.C.
§ 2254. In 2002, a Wisconsin jury convicted Reynolds in a fatal
carjacking. He seeks habeas relief based on two alleged viola‐
tions of his Sixth and Fourteenth Amendment right to coun‐
sel. First, he argues that Wisconsin violated his right to coun‐
sel when it stopped paying his state‐appointed lawyer during
2  No. 16‐3430
his direct appeal. In the alternative, he argues that he received
ineffective  assistance  from  his  counsel  during  his  direct  ap‐
peal and trial. The district court denied relief. We affirm.
I.  Factual & Procedural Background
In May 2001, two young men approached a group of teen‐
agers in a parking lot in Milwaukee. One of the men shot two
of the teenagers, killing one of them. The two men drove away
in the car the two teenage victims had been driving. Reynolds
was  arrested  two  days  later  after  he  was  identified  as  the
shooter.  Wisconsin  indicted  him  for  carjacking  resulting  in
death, carjacking resulting in bodily harm, and possessing a
firearm  as  a  felon.  A  jury  convicted  Reynolds  on  all  three
charges.  Reynolds  sought  relief  through  a  direct  appeal,  a
state  post‐conviction  proceeding,  and  a  state  habeas  corpus
petition. Finding no relief in the state courts, he filed his fed‐
eral habeas corpus petition. We start by summarizing the rel‐
evant facts at each stage of Reynolds’s case.
A.  Trial & Direct Appeal
Before trial, Reynolds raised an equal protection challenge
to  Wisconsin’s  felony  carjacking  statute.  The  trial  court  re‐
jected the challenge, and the case proceeded to trial. The main
issue  at  trial  was  identification.  The  prosecution  called  four
eyewitnesses,  including  the  surviving  gunshot  victim.  Each
identified  Reynolds  as  one  of  the  two  carjackers  and  as  the
shooter of at least one victim. Reynolds did not testify and did
not  call  any  witnesses.  The  jury  convicted  Reynolds  of  car‐
jacking  resulting  in  death,  carjacking  resulting  in  bodily
harm, and possession of a firearm by a felon.
The  Wisconsin  State  Public  Defender’s  Office  appointed
attorney  Terry  Williams  as  Reynolds’s  post‐conviction  and
No. 16‐3430  3
appellate counsel. With his counsel’s help, Reynolds moved
for a new trial on the basis that his trial counsel provided in‐
effective assistance and should have raised an alibi defense.1
Reynolds  submitted  two  affidavits  describing  his  wherea‐
bouts on the night of the shooting. The trial court found that
the affidavits did not help Reynolds since they placed him just
a  few  minutes  from  the  crime  scene  around  the  time  of  the
shooting. The court denied Reynolds’s motion without an ev‐
identiary hearing, concluding that he had failed to raise facts
showing ineffective assistance or prejudice to his defense.
On direct appeal, Reynolds argued that he was entitled to
an evidentiary hearing on his ineffective assistance claim and
added  an  argument  that  his  trial  counsel  failed  to  cross‐
examine  the  eyewitnesses  adequately.  The  Wisconsin  Court
of Appeals reversed and remanded with instructions to hold
what  Wisconsin  courts  call  a  Machner  hearing.  See  State  v.
Machner,  285  N.W.2d  905  (Wis. App.  1979).  On  remand,  the
trial  court  held  an  evidentiary  hearing,  considered  both
arguments, and denied Reynolds’s motion for a new trial.
Reynolds appealed again. At some point after the eviden‐
tiary hearing—and while Reynolds’s appeal was pending—
staff at the Wisconsin State Public Defender’s Office told at‐
torney Williams that he had spent too much time on his cases.
The  State  Public  Defender’s  Office  informed  Williams  that
they would not pay him for some of the work he had com‐
pleted on Reynolds’s appeal and that they would not pay him
for any more work on the case. They also told Williams that
they would no longer assign cases to him. Handling such a
1  In  Wisconsin,  defendants  generally  begin  their  direct  appeals  by
moving for a new trial in the trial court. See Wis. Stat. § 809.30(2)(h).
4  No. 16‐3430
management  issue  by  cutting  off  attorney  Williams  without
taking concrete steps to ensure that his clients would continue
to be represented would seem highly unusual and troubling.
The record before us simply does not show what efforts the
staff might have made to protect Reynolds’s interests.
Reynolds did not immediately learn that the State Public
Defenders  had  stopped  paying  his  attorney.  But  when  he
asked Williams to pursue in the second appeal the equal pro‐
tection challenge to the carjacking statute—which trial coun‐
sel had unsuccessfully raised before trial—Williams told him
that the State was not paying him anymore and that he could
not  afford  to  investigate  that  additional  question.  Reynolds
could not afford to pay for additional work, even assuming
that Williams could have accepted private payments on top of
the earlier public defender payments. The result was that Wil‐
liams did not investigate the equal protection challenge as a
possible  argument  on  appeal.  Nevertheless,  Williams  com‐
pleted briefing in the Wisconsin Court of Appeals on the inef‐
fective‐assistance claims and later filed a petition for review
with the Wisconsin Supreme Court. The Court of Appeals af‐
firmed the denial of a new trial, and the Wisconsin Supreme
Court denied review.
B.  Collateral Post‐Conviction Relief
In  October  2010,  Reynolds  filed  a  pro  se  collateral  post‐
conviction motion in the trial court under Wis. Stat. § 974.06.
As relevant here, Reynolds argued that he received ineffective
assistance  of  appellate  counsel  because  attorney  Williams
failed to argue that his trial counsel should have challenged
the  jury  instructions  on  carjacking.  The  formal  charging  in‐
strument had said that Reynolds took the vehicle “by use of”
a dangerous weapon, but the jury was instructed to convict
No. 16‐3430  5
even if it found that he took the car by only “threat of the use
of  a  dangerous  weapon.”  Reynolds  argued  that  the  jury
should have been instructed using the language in the formal
charges and that Williams was ineffective for not raising this
argument on appeal.
The trial court denied the motion without a hearing, and
the Wisconsin Court of Appeals affirmed. The appellate court
first  found  that  Reynolds’s  claim  for  ineffective  appellate
counsel  was  conclusory  and,  therefore,  procedurally  barred
under State v. Escalona‐Naranjo, 517 N.W.2d 157 (Wis. 1994). In
the alternative, the court found that the jury instructions were
sufficient  to  support  a  conviction  under  Wisconsin  law  be‐
cause  the  carjacking  statute  establishes  a  single  offense  that
can be committed by using or threatening to use a weapon.
See Wis. Stat. § 943.23(1g) (defining carjacking as intentionally
taking “any vehicle without the consent of the owner” by “the
use  of,  or  the  threat  of  the  use  of,  force”  or  a  dangerous
weapon). The court also found that even if the jury instruc‐
tions  were  improper,  Reynolds  could  not  prove  prejudice.
The central issue at trial was identification, not the difference
between using a gun and threatening to use a gun—after all,
the carjacker shot two victims, and one died. The Wisconsin
Supreme Court again denied review.
C.  State Habeas Corpus Proceedings
Reynolds next filed a state habeas corpus petition  in the
Wisconsin Court of Appeals under State v. Knight, 484 N.W.2d
540 (Wis. 1992). In it, he alleged that Wisconsin violated his
Sixth and Fourteenth Amendment right to counsel when the
State  Public  Defender’s  Office  stopped  paying  his  appellate
attorney. Reynolds claimed that the State interfered with his
representation, created a conflict of interest between him and
6  No. 16‐3430
his lawyer, and caused his lawyer to forgo investigating the
equal protection challenge. The Court of Appeals once again
denied relief. The court relied on two cases that applied the
conflict‐of‐interest standard in Cuyler v. Sullivan, 446 U.S. 335
(1980)—Freeman v. Chandler, 645 F.3d 863, 869 (7th Cir. 2011),
and  State  v.  Love,  594  N.W.2d  806,  810–11  (Wis.  1999)—and
found that Reynolds had not shown that the alleged conflict
adversely  affected  his  lawyer’s  representation  of  him.  The
court  found  as  a  matter  of  state  law  that  Reynolds  had  for‐
feited the equal protection challenge by failing to raise it be‐
fore the Machner evidentiary hearing. That means the forfei‐
ture occurred before the State stopped paying his lawyer. Any
effort  to  bring  the  equal  protection  claim  after  that  point
would have been frivolous, the Wisconsin Court of Appeals
found, so the conflict of financial interest could not have ad‐
versely affected Williams’s representation of Reynolds.
D.  Federal Habeas Corpus Proceedings
Reynolds then filed his federal habeas corpus petition un‐
der 28 U.S.C. § 2254. He argued that the State deprived him of
counsel when it created a conflict of interest during his direct
appeal, and that his appellate counsel was constitutionally in‐
effective for failing to object to the jury instructions.2 The dis‐
trict court denied relief. The court found that there was suffi‐
cient  evidence  to  convict  Reynolds  of  either  using  force  or
threatening to use force, and thus any mistake in the jury in‐
structions  could  not  have  prejudiced  him.  The  court  also
found that Reynolds had not shown that a conflict adversely
2  Reynolds  raised  several  additional  arguments  about  how  his  trial
and appellate attorneys were ineffective, but he does not pursue those ar‐
guments in this appeal.
No. 16‐3430  7
affected  his  appellate  representation  because  Reynolds  ar‐
gued only that the conflict prevented his lawyer from pursu‐
ing a procedurally barred claim. The district court declined to
grant a certificate of appealability, but we granted a certificate
to  review Reynolds’s  claims surrounding  his representation
after  the  State  Public  Defender’s  Office  stopped  paying  his
lawyer.
II.  Denial of Effective Counsel
The Sixth and Fourteenth Amendments guarantee crimi‐
nal defendants the effective assistance of counsel during their
first appeals as of right. Evitts v. Lucey, 469 U.S. 387, 393–94,
396–97  (1985);  Douglas  v.  California,  372  U.S.  353,  356–57
(1963). The two‐prong analysis in Strickland v. Washington, 466
U.S.  668  (1984),  governs  most  ineffective  assistance  claims.
Strickland requires a petitioner to show: (1) that the attorney
provided constitutionally deficient performance; and (2) that
the deficient performance prejudiced the defense. Id. at 687.
Prejudice  under  Strickland  means  “a  reasonable  probability
that, but for counsel’s unprofessional errors, the result of the
proceeding would have been different.” Id. at 694.
Although  Strickland  states  the  general  rule,  the  Supreme
Court has recognized some exceptions for situations in which
prejudice  is either  presumed or easier  to show. Bell  v.  Cone,
535  U.S.  685,  695–96  (2002)  (explaining  three  circumstances
when  prejudice  is  presumed).  Reynolds  argues  that  two  of
these exceptions apply here: a complete denial of counsel and
an actual conflict of interest.
A.  Complete Denial of Counsel
Reynolds  first  argues  that  the  State  caused  Williams  to
abandon his role as an advocate when it stopped paying him,
8  No. 16‐3430
causing  a  complete  denial  of  counsel  under  United  States  v.
Cronic, 466 U.S. 648 (1984). Cronic recognized three “circum‐
stances that are so likely to prejudice the accused” that they
amount to a complete denial of counsel and prejudice is pre‐
sumed. The first is when a defendant is denied the presence
of counsel at a critical stage of the proceedings. The other two
involve performance along those lines: counsel “entirely” fail‐
ing “to subject the prosecution’s case to meaningful adversar‐
ial testing,” and circumstances that are likely to prevent any
lawyer, “even a fully competent one,” from providing effec‐
tive assistance. Id. at 658–60; see also Bell, 535 U.S. at 695–97
(summarizing the three Cronic scenarios in which prejudice is
presumed  and  emphasizing  that  counsel’s  failure  “must  be
complete” to trigger presumption). Reynolds refers to the lat‐
ter two situations as “constructive” denial of counsel and ar‐
gues that one occurred here. He also likens his case to Penson
v. Ohio, 488 U.S. 75 (1988), in which the Court found a com‐
plete denial of counsel when appellate counsel withdrew after
submitting only a bare assertion that the appeal was meritless.
Id. at 88–89.
The Wisconsin Court of Appeals did not reach this denial‐
of‐counsel  question,  which  leads  Reynolds  to  argue  that  de
novo review applies. But the Wisconsin court failed to reach
this issue for a reason: Reynolds did not raise it in state court.
He procedurally defaulted this claim, and we cannot consider
it on the merits.
Before pursuing habeas relief in federal court, a state pris‐
oner  must  exhaust  available  state‐court  remedies.  28  U.S.C.
§ 2254(b)(1).  In  addition,  the  federal  claims  “must  be  fairly
presented to the state courts” so that the state has a “fair op‐
portunity to consider” the issues “and to correct that asserted
No. 16‐3430  9
constitutional defect.” Picard v. Connor, 404 U.S. 270, 275–76
(1971).  Fair  presentation  generally  requires  the  petitioner  to
raise both the operative facts and controlling legal principles
in the state courts. Ellsworth v. Levenhagen, 248 F.3d 634, 639
(7th Cir. 2001). If the petitioner fails to present a claim to the
state courts, that claim is procedurally defaulted, and a fed‐
eral  habeas  court  cannot  review  it  absent  “cause  and preju‐
dice” or a miscarriage of justice (meaning conviction of an in‐
nocent person). E.g., Thomas v. Williams, 822 F.3d 378, 384, 386
(7th Cir. 2016), citing Coleman v. Thompson, 501 U.S. 722, 750
(1991).
Even liberally construed, Reynolds’s pro se petition to the
Wisconsin Court of Appeals did not fairly present this claim
for denial of counsel. He argued only that Wisconsin denied
him the “effective assistance of counsel” by creating “an ac‐
tual conflict of interest” between him and his counsel. The ar‐
gument that he was denied counsel altogether is a different
legal theory governed by a different legal rule—one that the
Wisconsin courts did not have the chance to apply. By refor‐
mulating  his claim in this  appeal, Reynolds seeks to benefit
from a less stringent standard for deciding prejudice: Cronic’s
presumption of prejudice rather than Sullivan’s adverse‐effect
standard.  This  tactic  does  not  work.  See  Boyko  v.  Parke,  259
F.3d 781, 788 (7th Cir. 2001) (reformulation “should not place
the claim in a significantly different legal posture by making
the claim stronger or more substantial”). A petitioner is not
bound to his state court presentation word‐for‐word, and he
may provide further support for his claim. But he still must
have presented the governing legal principles and “the sub‐
stance of a federal habeas corpus claim” to the state court. Pi‐
card, 404 U.S. at 278; see also Duncan v. Henry, 513 U.S. 364,
10  No. 16‐3430
366 (1995) (per curiam) (“mere similarity of claims is insuffi‐
cient to exhaust”).
Even  though  Reynolds  did  not  describe  the  issue  in  so
many words as a complete denial of counsel, he might have
fairly presented this claim if he had cited relevant cases and
alleged facts that were “well within the mainstream of consti‐
tutional  litigation”  of  that  right.  Ellsworth,  248  F.3d  at  639.
Reynolds did not do so. He did not rely on cases addressing
complete denials of counsel. Instead, he cited Strickland, Sulli‐
van, and Walberg v. Israel, 766 F.2d 1071 (7th Cir. 1985), a case
applying  the  Sullivan  conflict‐of‐interest  standard.  Reynolds
also cited Douglas v. California, 372 U.S. 353 (1963), and Evitts
v. Lucey, 469 U.S. 387 (1985)—two cases that established the
right to effective assistance of counsel on direct appeal. But a
single reference to these cases was insufficient, especially in
light of the facts Reynolds described. Nowhere did Reynolds
assert that his lawyer abandoned his case. Although he said
that his lawyer “had in essence been fired,” he also wrote that
“Williams  felt  that  it  was  his  professional  responsibility  to
complete the current briefing process” and that he petitioned
the  Wisconsin  Supreme  Court  for  further  review  as  well.
When  describing  his  lawyer’s  deficient  performance,  Rey‐
nolds argued only that his lawyer declined to investigate one
argument: the equal protection claim. Asserting that his law‐
yer failed to make that single argument did not fairly notify
the  state  court  that  Reynolds  claimed  a  complete  denial  of
counsel. See Penson, 488 U.S. at 88–89 (distinguishing between
failing “to press a particular argument” and being “entirely
without the assistance of counsel on appeal”).
The fair presentment rule is not so rigid that a pro se peti‐
tioner  needed  to  cite  Cronic  or  any  other  denial‐of‐counsel
No. 16‐3430  11
case by name. For example, a petitioner could exhaust a de‐
nial of counsel claim by describing facts that amount to a com‐
plete  denial  of  counsel  and  citing  Strickland,  which  summa‐
rized: “Actual or constructive denial of the assistance of coun‐
sel altogether is legally presumed to result in prejudice.” 466
U.S. at 692. Still, the petitioner must describe facts and legal
principles that fairly notify the court of what it should decide.
Because  Reynolds  (1)  did  not  rely  on  cases  addressing  the
complete denial of counsel, (2) framed the violation as a con‐
flict of interest under Sullivan, and (3) did not allege that he
lacked  an  attorney,  this  abandonment  claim  was  not  fairly
presented to the state courts. Moreover, even if the denial‐of‐
counsel claim were not procedurally barred or if we were in‐
clined to consider it as sufficiently related to the claims Rey‐
nolds did raise, the record makes clear that attorney Williams
simply did not abandon Reynolds. Even after being told by
the State Public Defender that he would not be paid any more
for working on Reynolds’s case, Williams continued to repre‐
sent Reynolds for free in the state court of appeals, completing
the briefing there, and then filing a petition for review with
the Wisconsin Supreme Court. That fact does not immunize
his work from scrutiny, but this was not an abandonment that
could lead to the presumption of prejudice.
B.  Conflict of Interest
In the alternative, Reynolds argues that he was denied the
effective  assistance  of  counsel  when  the  State  Public  De‐
fender’s Office created a conflict of interest by ceasing to pay
his lawyer for work on Reynolds’s case while also informing
him that he would no longer receive new case assignments.
Reynolds  properly  exhausted  this  claim,  and  the  Wisconsin
Court of Appeals decided it on the merits. We cannot issue a
12  No. 16‐3430
writ of habeas corpus on this basis unless the decision “was
contrary  to,  or  involved  an  unreasonable  application  of,
clearly  established  Federal  law,  as  determined  by  the  Su‐
preme Court of the United States.” 28 U.S.C. § 2254(d)(1).
The first question under § 2254(d)(1) is the scope of clearly
established  federal  law  on  a  defendant’s  Sixth Amendment
right  to  conflict‐free  counsel.  The  Supreme  Court  has  held
that concurrent multiple representation—for example, repre‐
senting  co‐defendants  in  a  single  trial—violates  the  Sixth
Amendment if it “gives rise to a conflict of interest.” Sullivan,
446 U.S. at 348, citing Holloway v. Arkansas, 435 U.S. 475, 482
(1978).
Reynolds  argues  that  his  lawyer’s  financial  interest  kept
the lawyer from zealously representing him. We reject this ar‐
gument  for  two  independent  reasons.  First,  the  Supreme
Court has not yet extended its multiple‐representation deci‐
sions to financial conflicts of interest between attorney and cli‐
ent, let alone provided clear guidance as to whether or under
what circumstances a financial conflict of interest between at‐
torney and client violates a defendant’s right to counsel. That
silence  presents  a  nearly  insurmountable  obstacle  to  this
claim on a federal petition for a writ of habeas corpus. Second,
the Supreme Court has not given lower courts much guidance
as to what counts as an “adverse effect” under Sullivan, as dis‐
tinct from a reasonable probability of a different outcome un‐
der  Strickland.  Without  clearer  guidance  from  the  Supreme
Court, the state court decision here was not contrary to or an
unreasonable  application  of  clearly  established  Supreme
Court law.
In cases where defense counsel represents multiple clients
with conflicting  interests,  the Supreme Court  holds that the
No. 16‐3430  13
standard for reversal depends on whether anyone objected to
the conflict during the representation: “absent objection, a de‐
fendant must demonstrate that ‘a conflict of interest actually
affected the adequacy of his representation.’” Mickens v. Tay‐
lor, 535 U.S. 162, 168 (2002), quoting Sullivan, 446 U.S. at 349.
If  no  party  raises  a  timely  objection  to  a  conflict  of  interest
arising from multiple representation, the accused must show
that “an actual conflict of interest adversely affected his law‐
yer’s performance.” Sullivan, 446 U.S. at 348. If the defendant
shows  a  conflict  and  a  lapse  in  representation,  prejudice  is
presumed. Id. at 349–50. Reynolds argues that the logic of Sul‐
livan should apply to the financial pressure on his appellate
lawyer, which caused the lawyer to decline to investigate the
equal protection challenge and adversely affected his perfor‐
mance.
The Wisconsin Court of Appeals accepted Sullivan as stat‐
ing the controlling federal rule. That court found that even if
the State interfered with the lawyer’s representation, that po‐
tential  interference  came  after  the  lawyer  had  forfeited  the
equal protection claim. After the Machner evidentiary hearing,
the scope of the appeal was limited by the lawyer’s earlier fil‐
ings, and the equal protection claim was no longer available.3
The Wisconsin court then concluded that the financial conflict
3 Reynolds argues that Wisconsin law did not actually limit the scope
of the appeal at this point. The state courts treated the failure to raise the
issue properly as a waiver. In habeas corpus litigation, federal courts or‐
dinarily  defer  to  state  courts’  applications  of  state  procedural  rules  and
doctrine to proceedings in state courts, at least as long as the relevant rules
or doctrines are firmly established and consistently followed. E.g., Mar‐
tinez v. Ryan, 566 U.S. 1, 9–10 (2012). We see no reason to question the state
courts’ routine conclusion that Reynolds failed to raise the equal protec‐
tion claim in a timely way.
14  No. 16‐3430
had no adverse effect on Williams’s representation. See Sulli‐
van, 446 U.S. at 349–50. The state court reasoned that even if
attorney Williams had investigated the equal protection the‐
ory,  he  should  have  concluded  that  the  theory  could  not
properly be presented to the court of appeals. It had been for‐
feited  because  it  had not  been  raised  in  Reynolds’s  first  ap‐
peal. Sep. App. 71–72.
The Wisconsin court’s conclusion may or may not be cor‐
rect in the end, but it did not unreasonably apply Sullivan. The
Supreme Court itself simply has not extended Sullivan, or any
constitutional analysis of conflicts of interest, to financial con‐
flicts between attorney and client. Sullivan dealt with one team
of lawyers who represented three men who were accused of
the same murders but were tried in three separate trials. The
Court  vacated  habeas  relief  for  Sullivan,  the  one  client  who
had been convicted. The Court held that, in the absence of a
timely objection, the accused was required to show that the
potential conflict of interest (a) had ripened into an actual con‐
flict of interest that (b) adversely affected the adequacy of his
lawyers’ representation of him. 446 U.S. at 348.
The Supreme Court has not yet held that Sullivan extends
even  from  multiple,  conflicted  concurrent  representations  to
multiple,  conflicted  successive  representations,  Mickens,  535
U.S. at 174–76, let alone to financial conflicts of interest. Mick‐
ens affirmed the denial of habeas relief where the lawyer rep‐
resenting the accused in a murder case had been representing
the murder victim on unrelated, undisclosed (and confiden‐
tial) juvenile charges at the time he was murdered. The Mick‐
ens  opinion  noted  that  the  case  had  been  argued  on  the  as‐
sumption that the Sullivan would extend to conflicted succes‐
No. 16‐3430  15
sive representations, but that the Court was not actually en‐
dorsing that assumption. Id. at 175. Mickens expressly left that
successive representation question open. Id. at 176.
Under 28 U.S.C. § 2254(d)(1), the scope of the Sullivan rule
is a critical issue here. Mickens acknowledged that the federal
courts of appeals had applied Sullivan to “all kinds of alleged
attorney  ethical  conflicts,”  including  financial  conflicts,  as
well as attorney’s interests in getting a job with the prosecutor
or romance with the prosecutor. 535 U.S. at 174. But Mickens
pointedly  did  not  endorse  those  extensions  of  Sullivan.  Id.
Cautioning the  lower  courts more broadly,  the  Court  wrote
that  “Not  all  attorney  conflicts  present  comparable  difficul‐
ties.” 535 U.S. at 175.4
This  circuit  has  said  that  the  “precise  scope  of  claims  to
which the  Cuyler [v. Sullivan] standard applies has not been
definitively stated by the Supreme Court.” Spreitzer v. Peters,
114 F.3d 1435, 1451 n.7 (7th Cir. 1997) (denying habeas relief).
Since before Mickens, we have at least assumed that Sullivan
extends to financial conflicts of interests. United States v. Mar‐
rera, 768 F.2d 201, 206–07 (7th Cir. 1985) (attorney and defend‐
ant  agreed  to share  movie rights to  story  of crime). Mickens
makes it very difficult, though, to take that step in a habeas
corpus  challenge  to  a  state  conviction  governed  by
4 The cited financial conflict of interest case was United States v. Hearst,
638 F.2d 1190, 1193 (9th Cir. 1980) (remanding for hearing on whether at‐
torney F. Lee Bailey’s book contract regarding the defense of Patty Hearst
adversely affected his representation of her). But Mickens also noted the
Fifth Circuit decision in Beets v. Scott, 65 F.3d 1258, 1269–72 (5th Cir. 1995)
(en banc), which declined to extend Sullivan to a financial conflict of inter‐
est (also involving media contracts) and instead applied the Strickland v.
Washington standard to the claim of ineffective assistance of counsel.
16  No. 16‐3430
§ 2254(d)(1). To prevail here, Reynolds needs to show that it
would be unreasonable for the state court to fail to extend Sul‐
livan to financial conflicts between attorney and client and that
the state court applied Sullivan unreasonably. That extension
would  be  a  significant  new  development,  with  effects  far
more sweeping than its extension from concurrent to succes‐
sive representations. Extension to financial conflicts of inter‐
ests would pose a range of challenging problems. It would not
be unreasonable within the meaning of § 2254(d)(1) for lower
courts to refrain from making that extension themselves.
A wide range of fee arrangements can create at least some
tension between attorneys and clients. For example, it is not
unusual in criminal defense for a lawyer to insist on a sizable
up‐front retainer against which the lawyer can bill, often with
little hope of replenishing the retainer if the case takes more
work than the lawyer initially expected. Nor is it unusual for
clients  to  fall  behind  in  paying  their  lawyers.  In  those  and
many other circumstances, lawyers may face more or less se‐
rious financial conflicts of interest as the client and case may
call for additional work that will be essentially unpaid. View‐
ing the lawyer’s incentives through a narrow and short‐term
economic  lens,  these  situations  give  a  lawyer  a  financial  in‐
centive  to  minimize  non‐essential  work,  much  like  attorney
Williams’s  incentives  here.  Nor  is  it  unusual  for  underpaid
and overworked public defenders to have to allocate a finite
number of hours in a month or year among different clients.
If such routine incentives, combined with an admission by
a  lawyer  that  his  or  her  judgments  about  which  additional
work to carry out were affected at the margin by those incen‐
tives, were enough to find an actual conflict and “adverse ef‐
fect”  that  demonstrated  ineffective  assistance  of  counsel,
No. 16‐3430  17
many convictions would become vulnerable to collateral at‐
tack.  Yet  we  know  that  lawyers  facing  such  incentives  and
choices are also motivated by their sense of professional duty
and pride, and an interest in their professional reputations, to
continue to do good work for a client even if no more money
will  be  paid  in  the  particular  case.  Those  realities  underlie
Rule of Professional Conduct 1.16. A client’s failure to pay as
agreed  can  provide  good  cause  for  a  lawyer  to  terminate  a
representation.  But  Rule  1.16  provides  that  a  lawyer  in  that
situation must continue the representation if ordered to do so
by a court. When that happens, there is a financial conflict of
interest not unlike what has happened here. We have not been
directed to any cases treating such conflicts as showing a de‐
nial  of  effective  counsel,  let  alone  to  a  controlling  Supreme
Court case holding as much.
These  considerations  help  to  show  that  how  to  apply
conflict‐of‐interest doctrine, which so far has developed in the
Supreme  Court  to  address  loyalties  to  multiple  clients,  to
financial  conflicts  raises  challenging  questions  that  the
Supreme Court simply has not addressed. Under § 2254(d)(1),
we  may  not  use  a  federal  habeas  proceeding  to  extend
Supreme  Court  precedents  into  new  fields  where  there  is
room  for  fair‐minded  disagreement,  as  there  is  here.  See
generally Harrington v. Richter, 562 U.S. 86, 101 (2011).
Still, Reynolds argues that a conflict of interest is a conflict
of interest. From his point of view, the fact that the attorney’s
loyalties or interests ran in his personal favor rather than in
favor of another client should not matter. After all, attorney
Williams acknowledged that it  was finances that led him to
refuse Reynolds’s request that he look into the equal protec‐
tion challenge to the carjacking statute, and perhaps also to a
18  No. 16‐3430
challenge to the jury instructions on carjacking. Reynolds re‐
lies on the language in Sullivan teaching, in the context of mul‐
tiple, concurrent representations, that an actual conflict of in‐
terest  that  adversely  affected  the  lawyer’s  performance  vio‐
lates the Sixth Amendment right to counsel. 446 U.S. at 348–
49.
Assuming that the Sullivan standard applies to this finan‐
cial conflict of interest, which would be consistent with both
the state appellate court’s approach here and our opinion in
Marrera, Reynolds has shown some effect on Williams’s repre‐
sentation: a failure to investigate at least the additional equal
protection claim. An adverse effect seems to be a lower stand‐
ard than the Strickland reasonable probability of a different re‐
sult. But what does it take to show an adverse effect? The Wis‐
consin court concluded there was no adverse effect from the
conflict  of  interest  because  any  new  claims  or  theories  that
Reynolds asked Williams to pursue, including the equal pro‐
tection claim, were already forfeited by the time the conflict
arose.  Yet  effective  lawyers  can sometimes overcome proce‐
dural hurdles in criminal appeals and habeas cases. And by
asking whether Williams could have raised the equal protec‐
tion claim after Reynolds asked him to, the Sullivan “adverse
effect” standard begins to slide toward the Strickland standard
of prejudice, gauged in terms of the outcome of the case. The
Sullivan analysis seems to focus more on the conflict’s effect
on the attorney’s actions than on the outcome of the case.
Yet Sullivan still requires not just an effect but an adverse
effect on attorney performance. 446 U.S. at 348. The Supreme
Court has not clarified what counts as an adverse effect. See
Beets v. Scott, 65 F.3d 1258, 1265 (5th Cir. 1995) (en banc) (“The
No. 16‐3430  19
precise nature of [Cuyler v. Sullivan’s] ‘actual conflict’ and ‘ad‐
verse effect’ elements is rather vague … .”). Nor has the Court
clarified where this standard lies between any effect on attor‐
ney  performance  and  an effect that  rises  to Strickland preju‐
dice.
We  have  written  broadly  that  a  defendant  shows  an  ad‐
verse effect “if, but for the attorney’s actual conflict of interest,
there is ‘a [reasonable] likelihood that counsel’s performance
somehow would have been different.’” Stoia v. United States,
22 F.3d 766, 771 (7th Cir. 1994), quoting Frazer v. United States,
18  F.3d  778,  787  (9th  Cir.  1994)  (Beezer,  J.,  concurring).  We
used that language in a most unusual case, and one that did
not involve the forfeiture problem that the state court found
decisive here.  The lawyer  in  Stoia who had  directed  the de‐
fense strategy had entered his own plea agreement with the
federal  government.  The  lawyer  was  probably  violating  his
plea  agreement  by  directing  the  defense  strategy  in  Stoia’s
case, and at the same time the lawyer was under investigation
for soliciting perjury and obstructing justice in his represen‐
tation  of  one  of  Stoia’s  co‐defendants.  22  F.3d  at  771–72.  In
Stoia,  the  defendant’s  other  lawyers  identified  a  number  of
specific ways in which the strategist made decisions that hurt
Stoia’s defense. Id. at 773.
That situation in Stoia is not readily comparable to a cli‐
ent’s  (or  state  public  defender  agency’s)  more  quotidian  re‐
fusal or inability to pay. And in any event, the broad Stoia lan‐
guage is not binding on state courts. Under § 2254(d)(1), we
cannot grant habeas relief based on only our own precedent
when a state court has ruled on the merits of a claim.
Even  if  our  language  in  Stoia  were  binding  precedent  in
this case, state courts interpreting it would still need to decide
20  No. 16‐3430
what it means for counsel’s performance to be “different,” let
alone  “adverse.”  “Adverse”  can  be  read  fairly  to  require  a
showing of at least some sort of negative consequence for the
client. It would not be unreasonable for a state court consid‐
ering this question to peek at the merits of the claim that the
attorney did not investigate and raise. Why would failure to
raise  a  sure  loser  of  an  issue  necessarily  amount  to  an  “ad‐
verse effect” requiring a new trial?
The Wisconsin carjacking statute does not rely on any dis‐
tinctions that implicate suspect classes or fundamental rights.
An equal protection challenge would have to show the legis‐
lature drew a line that bore no rational relationship to a legit‐
imate governmental purpose, e.g., Armour v. City of Indianap‐
olis, 566 U.S. 673, 681 (2012), a standard that is very difficult to
satisfy. This is not the sort of argument that moves courts to
overlook  procedural  failures  to  free  wrongly  convicted  de‐
fendants.  The  state  court  did  not  unreasonably  apply  Su‐
preme Court precedent in holding that a missed opportunity
to raise such an argument in an already untimely fashion was
not an “adverse effect” within the meaning of Sullivan. Rey‐
nolds’s  claim  cannot  meet  the  stringent  standard  of
§ 2254(d)(1).5
5 Our dissenting colleague proposes as a remedy that Reynolds’s con‐
victions be set aside. That remedy would not fit the argued violation of his
right to counsel on appeal. If the financial conflict produced an “adverse
effect” within the meaning of Sullivan, that effect was counsel’s refusal to
investigate an already‐forfeited equal protection challenge to Wisconsin’s
carjacking  statute.  That  potential  challenge  still  has  not  been  shown  to
have any arguable merit. If there were a constitutional violation here in
the appellate representation, the remedy would seem to be a fresh appeal,
not a new trial based on the assumption that the equal protection theory
not only was not forfeited but would have produced a victory on appeal.
No. 16‐3430  21
III. Ineffective Assistance of Appellate Counsel
Finally, Reynolds claims that he received ineffective assis‐
tance  of  appellate  and  trial  counsel  because  both  attorneys
failed  to  challenge  the  jury  instructions  on  his  carjacking
counts. This issue is outside the scope of the certificate of ap‐
pealability. We certified Reynolds’s claim that the State vio‐
lated his right to the effective assistance of counsel “when the
state public defender’s office allegedly informed Reynolds’s
appointed counsel that he would not be paid for future work
on Reynolds’s case and when counsel, citing this reason, de‐
clined  to  conduct  additional  investigation  that  Reynolds  re‐
quested.” We did not certify any questions related to the jury
instructions, and Reynolds did not request permission to ar‐
gue any non‐certified claims. See Welch v. Hepp, 793 F.3d 734,
737 (7th Cir. 2015). Also, since we know (a) that the state stat‐
ute authorizes conviction for either use or threatened use of a
weapon,  see  Wis.  Stat.  § 943.23(1g),  and  (b)  that  a  deadly
weapon was certainly used in this case, we see no reason to
consider expanding the scope of our certificate of appealabil‐
ity. We therefore decline to review this issue. Welch, 793 F.3d
at 737–38.
See, e.g., Shaw v. Wilson, 721 F.3d 908, 919 (7th Cir. 2013) (ordering new
appeal where appellate counsel had been ineffective); see generally Roe v.
Flores‐Ortega, 528 U.S. 470, 484 (2000) (where appellate counsel abandoned
defendant  by  failing  to  file  notice  of  appeal,  and  defendant  otherwise
would have taken appeal, defendant is entitled to new appeal); Penson v.
Ohio, 488 U.S. 75, 87 (1988) (where state appellate court found several ar‐
guable issues in record, remedy for denial of counsel on appeal was new
appeal); Dodd v. Knight, 533 F. Supp. 2d 844, 854 (N.D. Ind. 2008) (ordering
new appeal where appellate counsel was ineffective).
22  No. 16‐3430
The  district  court’s  judgment  denying  a  writ  of  habeas
corpus is
AFFIRMED.
No. 16-3430 23
WOOD, Chief Judge, dissenting. Few would doubt that fi-
nancial conflicts of interest can destroy the integrity of a rela-
tionship. That is why, in Tumey v. Ohio, 273 U.S. 510 (1927),
the Supreme Court found that due process was violated when
a local judge’s salary depended on the fines that resulted from
convictions in cases before him. See also Connally v. Georgia,
429 U.S. 245 (1977) (due process violated by local law provid-
ing that judge was compensated only when he issued search
warrants, not when he denied them). It is why judges must
give extra scrutiny to the performance of counsel in class ac-
tions, lest the lawyer sell out the class and walk away with an
enormous fee. See RICHARD A. POSNER, ECONOMIC ANALYSIS
OF LAW 803–04 (9th ed. 2014); Jack B. Weinstein, Ethical Dilem-
mas in Mass Tort Litigation, 88 NW. U. L. REV. 469, 502–03
(1994). And it is why the Rules of Professional Conduct for
Attorneys in Wisconsin, which are based on the American Bar
Association’s Model Rules with some additions not material
here, state that “a lawyer shall not represent a client if the rep-
resentation involves a concurrent conflict of interest.”
SCR 20:1.7(a). One way in which a forbidden conflict can arise
is if “there is a significant risk that the representation of one
or more clients will be materially limited by … a personal in-
terest of the lawyer.” Id. at 20:1.7(a)(2). The lawyer is relieved
of this responsibility only if, among other things, the “affected
client gives informed consent, confirmed in a writing signed
by the client.” Id. at 20:1.7(b)(4). In addition, lawyers may not
limit the scope of their representation of a client unless “the
client gives informed consent,” again in writing. Id. at
20:1.2(c).
The obvious concern is that, faced with a conflict between
one’s own financial interest and the interests of another per-
son, the former will win out. If the lawyer puts his or her own
24 No. 16-3430
interest first, the client is the loser. This is precisely what Cor-
nell Reynolds says happened to him in the course of the crim-
inal prosecution the State of Wisconsin brought against him
in the early 2000s.
The key events in the case can be summarized quickly. In
May 2001, Reynolds was charged with carjacking resulting in
death. The primary issue at the trial was identification. A jury
convicted him, and he was sentenced in early 2002. As Wis-
consin law permits, Reynolds then filed a direct appeal and a
post-conviction motion simultaneously. At that point, the
Wisconsin State Public Defender’s Office appointed attorney
Terry Williams to serve as Reynolds’s post-conviction and ap-
pellate counsel. Williams filed a motion for a new trial on the
ground that Reynolds’s trial counsel had been ineffective in
failing to raise an alibi defense. The state trial court denied the
post-conviction motion without holding an evidentiary hear-
ing.
Still with Williams’s help, Reynolds appealed to the Wis-
consin Court of Appeals, which reversed and remanded the
ruling on the post-conviction motion so that the trial court
could hold an evidentiary hearing on the ineffectiveness
point. The trial court obliged, but it again concluded that
Reynolds was not entitled to a new trial.
So far, so good. The problem now before us arose when
Williams filed a new appeal on Reynolds’s behalf. While that
appeal was pending, someone on the staff of the Wisconsin
State Public Defender’s Office told Williams that he had spent
too much time on the case. The consequences were serious for
Williams: (1) he would not be paid for some of the work he
already had done; (2) he would not be paid for any more work
on Reynolds’s case; and (3) he would no longer receive case
No. 16-3430 25
assignments from the Public Defender. At that point, Williams
halted all further work on the case. Reynolds asked him to
pursue an equal protection challenge to the car-jacking stat-
ute, but Williams refused to do so. Importantly, his stated rea-
son had nothing to do with the potential merit of such an ar-
gument. Williams frankly said that he was shutting down his
work because the state was not paying him. He told Reynolds
that he would do additional work if Reynolds paid him for it,
but Reynolds was unable to do so. Williams simply filed the
brief he already had prepared, and after the court of appeals
affirmed the trial court’s order, he later filed a petition for re-
view on Reynolds’s behalf in the Wisconsin Supreme Court.
After some additional state proceedings, Reynolds filed a
petition for habeas corpus under 28 U.S.C. § 2254 in the fed-
eral court. He presented several arguments: first, that he was
constructively denied counsel after the Public Defender’s Of-
fice cut off Williams’s pay, in violation of United States v.
Cronic, 466 U.S. 648 (1984); second, that Williams had an ac-
tual financial conflict of interest, which resulted in the denial
of effective assistance of counsel under the standards of Cuyler
v. Sullivan, 446 U.S. 335 (1980); and third, that his appellate
counsel was ineffective pursuant to Strickland v. Washington,
466 U.S. 668 (1984), for failing to object to trial counsel’s own
failure to object to certain jury instructions. While I am per-
suaded that the state courts’ decisions fell within the generous
boundaries provided by section 2254 with respect to the first
and third of these arguments, I conclude that the state courts,
as well as the majority here, have failed to apply Cuyler
properly and that Reynolds is entitled to issuance of the writ
on this ground.
26 No. 16-3430
In Cuyler, the Supreme Court addressed the question
“whether a state prisoner may obtain a federal writ of habeas
corpus by showing that his retained defense counsel repre-
sented potentially conflicting interests.” 446 U.S. at 337. In the
case before the Court, two lawyers had represented three co-
defendants in a murder prosecution. The Court accepted that
this amounted to multiple representation. Id. at 342. But it
clarified that “multiple representation does not violate the
Sixth Amendment unless it gives rise to a conflict of interest.”
Id. at 348.
The burden on a defendant who seeks to establish a viola-
tion of the Sixth Amendment in such cases, the Court held, is
to “demonstrate that an actual conflict of interest adversely af-
fected his lawyer’s performance.” Id. (emphasis added). Such a
showing might rest on a failure to cross-examine a witness, or
a failure to object to the introduction of certain evidence. Id.
at 348–49. Critically, it is the lawyer’s performance that is the
focus of the inquiry, not the ultimate outcome of the trial. The
Court expressly declared that once unconstitutional multiple
representation has been demonstrated—that is, once the de-
fendant has shown an actual conflict and an adverse effect on
the lawyer’s performance—there “is never harmless error.” Id.
at 349 (emphasis added). Underscoring that fact, the Court
continued as follows:
Thus, a defendant who shows that a conflict of
interest actually affected the adequacy of his
representation need not demonstrate prejudice
in order to obtain relief.
Id. at 349–50. Because the court of appeals had used a different
standard, the Court remanded the case for further proceed-
ings.
No. 16-3430 27
As my colleagues note, some conflicts and ethical viola-
tions do not present difficulties akin to multiple representa-
tion. Ante at 17. But we must not lose sight of the fact that
Cuyler created a standard to be applied not just when one at-
torney represents clients with competing interests, but “in sit-
uations where Strickland itself is evidently inadequate to as-
sure vindication of the defendant’s Sixth Amendment right to
counsel.” Mickens v. Taylor, 535 U.S. 162, 176 (2002). The exist-
ence of an actual conflict of interest between attorney and cli-
ent is one of those instances. Cf. United States v. Gonzalez-Lopez,
548 U.S. 140, 146 (2006) (finding structural error in the depri-
vation of the right to counsel of choice, rejecting harmless er-
ror, and holding that “no additional showing of prejudice is
required to make the violation ‘complete’”).
Reynolds has established the two elements required by
Cuyler: he has shown that there was an actual conflict of inter-
est between him and his lawyer, and he has shown that the
conflict had an adverse effect on his lawyer’s performance, in-
sofar as the lawyer limited the scope of his representation to
the matters he already had researched and flatly refused to
look into anything else. The Supreme Court has not distin-
guished between an actual conflict of interest arising out of
multiple representation, as was present in Cuyler, and a more
direct conflict of interest between attorney and client based on
financial considerations. (Indeed, the latter situation involves
a form of multiple representation: Williams was trying to rep-
resent both himself and Reynolds.) Nor can there be any
doubt that the conflict between Reynolds and Williams af-
fected Williams’s performance. Without obtaining Reynolds’s
permission, as contemplated by professional ethics norms
and over Reynolds’s direct objection, Williams froze his work
for Reynolds and did nothing but file what he already had
28 No. 16-3430
completed, following up on the same limited points with the
state supreme court.
At the state post-conviction stage, Reynolds presented this
conflict-of-interest claim to the state courts. The Wisconsin
court of appeals agreed that “a conflict of interest may also
arise when an attorney and his client have divergent inter-
ests.” Sep. App. at 70 (quotation marks omitted). The court
then correctly noted that the claimant had to show an actual
conflict and an adverse effect on the lawyer’s performance. Id.
It then discussed Williams’s performance for a bit, but it piv-
oted over to the ultimate result in the case, when it said that
had Williams not had the conflict, he would have determined
that the issue Reynolds wanted to raise was not properly be-
fore the court in any event. That is the language of harmless
error, not a reference to Williams’s own performance. It says,
in essence, that the conflict affected Williams’s performance
insofar as he failed to raise the issues that Reynolds wanted to
see explored, but that Williams’s failure could not have af-
fected the outcome.
That reasoning is contrary to the holding of Cuyler, which
cannot have been more clear in rejecting a harmless-error ap-
proach. At the very least, it is an unreasonable application of
Cuyler, since it shifts from a focus on Williams’s performance
to one on the ultimate impact of Williams’s decisions in a
manner that Cuyler forbids. See also Mickens, 535 U.S. at 170
(2002) (stressing that the conflict must affect counsel’s perfor-
mance, rather than be merely a theoretical division of loyal-
ties). Tempting as it may be to engage in harmless-error anal-
ysis in every case, the Supreme Court has decided that there
are some situations in which it is off the table.
No. 16-3430 29
The only remaining question is what the appropriate rem-
edy is in this case. A common remedy for a failure of counsel
at the appellate stage would be, as my colleagues have noted,
a fresh appeal with unconflicted counsel. See Shaw v. Wilson,
721 F.3d 908, 919 (7th Cir. 2013). But nothing in the statute re-
stricts the district court’s options to only that remedy. To the
contrary, the law provides that “[t]he court shall summarily
hear and determine the facts, and dispose of the matter as law
and justice require.” 28 U.S.C. § 2243, ¶ 8. The Supreme Court
has recognized that “a court has broad discretion in condi-
tioning a judgment granting habeas relief.” Hilton v. Braun-
skill, 481 U.S. 770, 775 (1987). The question here is thus what
remedy would suffice to redress the serious miscarriage of
justice suffered by Reynolds.
In my view, this is a situation that demands more than a
fresh appeal. The state courts have already decided—in the
absence of a proper adversary presentation on Reynolds’s be-
half—that any additional points that may have been raised on
appeal were forfeited or meritless. I have no reason to believe
that they would not simply reiterate that point, if we were to
limit the remedy to a new appeal. (Principles of law-of-the-
case might even require them to do so.) Because a new appeal
would not redress the constitutional violation in Reynolds’s
case, I would issue a conditional writ granting him a new trial
(if the state chooses to take that step). This is well within the
scope of the statute, as our sister circuits have recognized. See,
e.g., Ramchair v. Conway, 601 F.3d 66, 78 (2d Cir. 2010) (peti-
tioner deprived of effective appellate counsel entitled to new
trial as remedy); Turner v. Bagley, 401 F.3d 718, 727 (6th Cir.
2005) (granting unconditional writ requiring new trial for
deprivation of access to appellate process because of undue
delay); Eagle v. Linahan, 279 F.3d 926, 943–44 (11th Cir. 2001)
30 No. 16-3430
(appellate counsel’s failure to raise Batson claim remedied by
grant of new trial).
In this respect, I agree with the observations of the Tenth
Circuit in Clayton v. Jones, 700 F.3d 435 (10th Cir. 2012), a case
in which the defendant was denied his right to appeal because
of ineffective assistance of counsel. As that opinion put it,
“[c]ertainly a district court is not confined to only granting an
appeal out of time when a petitioner has been denied his right
to a direct appeal. It is not unprecedented for a district court
to fashion a remedy for ineffective assistance of counsel on
appeal that goes beyond an appeal out of time.” Id. at 443–44.
The only problem in Clayton was that the district court had
not sufficiently explained why a new appeal would not suf-
fice. I have indicated why I believe this case is a proper candi-
date for a broader remedy. Reynolds has tried hard to navi-
gate the habeas corpus process—no mean feat for even an ex-
perienced lawyer—and in my view he has shown that he
should receive a new trial.
I therefore respectfully dissent.