Roderick Lewis v. Dushan Zatecky

993 F.3d 994

• Court: Court of Appeals for the Seventh Circuit
• Decided: April 13, 2021
• Docket: 20-1642
• Status: Published

Opinion

In the

United States Court of Appeals For the Seventh Circuit ____________________ No. 20‐1642 RODERICK V. LEWIS, Petitioner‐Appellant, v.

DUSHAN ZATECKY, Respondent‐Appellee. ____________________

Appeal from the United States District Court for the Southern District of Indiana, Indianapolis Division. No. 1:19‐cv‐01515‐RLY‐MPB — Richard L. Young, Judge. ____________________

ARGUED SEPTEMBER 30, 2020 — DECIDED APRIL 13, 2021 ____________________

Before SYKES, Chief Judge, and WOOD and BRENNAN, Circuit Judges. WOOD, Circuit Judge. When has a client charged with a se‐ rious crime received not merely inadequate assistance of counsel, but a failure of representation so serious that “coun‐ sel has entirely failed to function as the client’s advocate”? Florida v. Nixon, 543 U.S. 175, 189 (2004). This is the situation the Supreme Court first addressed in United States v. Cronic, 466 U.S. 648 (1984). Although such a total breakdown is rare, 2 No. 20‐1642

the Court has never wavered from the recognition that it can occur. In such cases, unlike those presenting more conven‐ tional ineffective‐assistance claims, the defendant does not need to make an independent showing of prejudice. See Strickland v. Washington, 466 U.S. 668 (1984). The failing is so profound that prejudice is inherent in the situation. In the case before us, Roderick Lewis argues that his is one of the extraordinary cases to which the Cronic rule applies. Standing convicted of felony murder, he received literally no assistance from his lawyer during the sentencing stage of the trial. After proceedings in the state courts, which we detail below, he turned to federal court and filed a petition for a writ of habeas corpus under 28 U.S.C. § 2254. The district court de‐ nied relief, but it issued a certificate of appealability to Lewis. We conclude that the decision of the last responsible state court was contrary to Supreme Court precedent, insofar as it held that Strickland, not Cronic, furnished the applicable rule, and it was an unreasonable application of Cronic, insofar as it focused on that case.1 We thus reverse and remand for issu‐ ance of the writ, limited to sentencing.

1 In order to be fair to the state court, we consider its decision under each of the two distinct branches of section 2254(d)(1), as courts commonly do. See, e.g., Aki‐Khuam v. Davis, 339 F.3d 521, 529 (7th Cir. 2003) (“… it is clear that the state trial court proceedings, and the state supreme court review thereof, resulted in a decision contrary to, and involving an unrea‐ sonable application of, federal law as determined by the United States Su‐ preme Court[.]”); Bailey v. Rae, 339 F.3d 1107, 1118–19 (9th Cir. 2003) (state court’s application of a standard “is ‘contrary to’ clear Supreme Court precedent[] [and] [t]he state court’s denial of the Brady claim was also ob‐ jectively ‘unreasonable’[.]”); Pazden v. Maurer, 424 F.3d 303, 306 (3d Cir. 2005) (the state court’s determination “was both contrary to, and an un‐ reasonable application of, clearly established law as proclaimed by the No. 20‐1642 3

I A We take our account of the underlying facts from the sec‐ ond opinion of the Court of Appeals of Indiana, the last state court to consider this case. See Lewis v. State (Lewis II), 116 N.E.3d 1144 (Ind. Ct. App. 2018). That court in turn relied on the facts it had reported on direct appeal, see Lewis v. State(Lewis I), 973 N.E.2d 110 (Table), (Ind. Ct. App. 2012), but we can largely disregard that detail. The case involved a toxic mixture: drugs, robbery plans, guns, and immaturity. Richard Rogers, then 16 years old, ran a drug house in Fort Wayne, Indiana, with Sidney Wilson, 14 years old. On June 29, 1999, Rogers invited Christopher Hale to visit the drug house, but Hale declined because of tensions with Wilson. Later that evening, Hale, petitioner Roderick Lewis, and Kajuanta Mays came up with a plan to rob Rogers and Wilson of both drugs and money. They first confirmed that Rogers and Wilson were alone by sending Angela

Supreme Court.”); Fratta v. Quarterman, 536 F.3d 485, 502–03 (5th Cir. 2008) (“The district court was thus correct in determining that the CCA’s deci‐ sion was contrary to, and involved an unreasonable application of, clearly established federal law.”); Breakiron v. Horn, 642 F.3d 126, 139 (3d Cir. 2011) (the state court’s ruling “is both contrary to and an unreasonable application of Strickland.”); Dennis v. Sec’y, Pennsylvania Dep’t of Corr., 834 F.3d 263, 285 (3d Cir. 2016) (“We conclude…that the [state court]’s deci‐ sion … rested on … unreasonable applications of clearly established law, or were contrary to United States Supreme Court precedent.”); Rivera v. Thompson, 879 F.3d 7, 16–17 (1st Cir. 2018) (“[W]e conclude that … the [state court]’s holding was contrary to governing Supreme Court law[.] … Thus, the [state court]’s conclusion ‘involved an unreasonable application of[] clearly established Federal law, as determined by the Supreme Court of the United States.’”). 4 No. 20‐1642

Lawson to the house to buy drugs. Hale then showed up, fol‐ lowed by Lewis and Mays. The group smoked and drank to‐ gether. Two of them were armed: Lewis had a .38 special re‐ volver, and Hale had a 9 mm firearm. At one point Hale went upstairs. When he returned, he said “die bitch” and shot Wilson five times, killing him. Rog‐ ers and Lewis then each reached for a shotgun. Hale told Lewis to kill Rogers, but Lewis refused, instead handing his revolver to Mays and saying, “if you want it … you do it.” Mays did not hesitate: he shot Rogers multiple times, fatally. Lewis, Hale, and Mays then collected the drugs and money and fled. They wound up in a hotel where they laughed and partied through the night. Later, Lewis had his uncle bury the murder weapon. For the next few years, the crime remained unsolved and Lewis traveled around the country, living in Arizona and In‐ diana. Ultimately, however, investigators in Fort Wayne iden‐ tified him as a suspect in the 1999 murders. They found him in a prison in May 2009 and interviewed him; on February 25, 2011, the State of Indiana charged him with two counts of fel‐ ony murder and two counts of robbery. He was arrested on June 27, 2011. B At trial, Lewis was represented by Attorney Jeffrey Raff. Raff tried to get Lewis seriously to consider some plea offers, but Lewis was uninterested, perhaps because he did not un‐ derstand the concept of felony murder and thought that, be‐ cause he did not shoot either Wilson or Rogers, he was not guilty. If that was his impression, he was mistaken. The jury found Lewis guilty as charged. No. 20‐1642 5

The problems that bring Lewis before us today arose at the sentencing phase. Here is how the Indiana Court of Appeals described Raff’s assistance to Lewis at that critical point: “Judge I’m going to defer to Mr. Lewis if he has any comments. I don’t have anything to add.” Sentencing Transcript at 23–24. This is the sum total of trial coun‐ sel’s participation at Lewis’s sentencing hearing, at which Lewis was being sentenced for two counts of fel‐ ony murder and faced a maximum sentence of 130 years in prison. The trial court found no mitigating cir‐ cumstances—none being asserted by the defense—and sentenced Lewis to the maximum aggregate sentence of 130 years in prison. Lewis II, ¶ 1. Represented by new counsel, Lewis took a direct appeal, but it was unsuccessful. See Lewis I. Acting pro se, Lewis then filed a post‐conviction petition in the state court in 2013. Post‐conviction counsel amended that petition in Octo‐ ber 2016, and the court held an evidentiary hearing on July 7, 2017. Lewis called Attorney Raff, among others, to testify at that hearing. The state conceded that Raff “basically did not do any advocacy at the sentencing hearing” but argued that he could not have made a difference anyway. Raff himself testi‐ fied about his normal procedures for preparing for a sentenc‐ ing hearing. He also described quite a few things that he did not do: He made no inquiries about Petitioner’s mental health history, and was not aware that Petitioner had attempted suicide at the Allen County Jail. … He did not ask Petitioner about his upbringing or his family 6 No. 20‐1642

members, did not speak to his relatives or friends, and did not have him examined by a mental health profes‐ sional. He did not prepare Petitioner to make a state‐ ment at sentencing, and explained that Petitioner did not take his advice well. Lewis II, ¶ 16, quoting from the post‐conviction court’s find‐ ings of fact. Essentially Raff thought that Lewis was a hope‐ less cause, and so there was nothing useful Raff could do. Other witnesses at the post‐conviction hearing spoke about evidence that might have had an impact at sentencing, includ‐ ing a psychologist who diagnosed Lewis with bipolar II dis‐ order and discussed his associated substance‐abuse problem, physical abuse by his mother’s boyfriends, mental disorders in other family members, and his attempted suicide. None of this, it bears repeating, was brought out during the sentencing stage of the trial. On state post‐conviction review, the state appellate court “agree[d] with Lewis that trial counsel’s performance at sen‐ tencing was clearly deficient.” Lewis II, ¶¶ 4, 20. Nevertheless, the court held that “our review leaves us with the firm con‐ viction that Lewis was not prejudiced by counsel’s deficient performance.” Id. ¶ 20. It reviewed the following potential mitigating circumstances: Lewis’s role as an accomplice; his age; his difficult childhood; and his mental health. None of these could have supported a finding of prejudice, in the court’s view, nor was it troubled by his consecutive sentences. Finally, the court turned to the issue that has survived to reach us: whether the proper standard for assessing Lewis’s case comes from Cronic, as Lewis argues, or Strickland. If it is Strickland, then Lewis’s case is over: we cannot say that the Indiana Court of Appeals was unreasonable when it found No. 20‐1642 7

that Lewis had not been prejudiced by his attorney’s substand‐ ard performance. (We add that we are not necessarily saying that we would have resolved the prejudice issue the same way. We mean only that we are satisfied that the state court acted within the generous boundaries delineated for it by 28 U.S.C. § 2254(d)(1).) If Cronic applies, however, then matters are quite different, because prejudice need not be shown. But the state court found that Lewis’s case did not fit within the Cronic framework. Lewis II, ¶ 39. Its finding of no prejudice for Strickland purposes required it to affirm the trial court’s denial of post‐conviction relief. After exhausting his state‐court remedies, Lewis filed a pe‐ tition under 28 U.S.C. § 2254 in the federal court. The district court acknowledged that in Miller v. Martin, 481 F.3d 468 (7th Cir. 2007), this court had found that the Indiana court had un‐ reasonably failed to apply Cronic to an attorney’s performance at sentencing, and thus that petitioner Miller was entitled to the issuance of a writ of habeas corpus. Id. at 470. But, the court thought, Miller had been undermined by two decisions of the U.S. Supreme Court: Wright v. Van Patten, 552 U.S. 120 (2008), and Woods v. Donald, 575 U.S. 312 (2015). It character‐ ized what happened in Lewis’s case as counsel’s complete failure to subject the prosecution’s case to adversarial testing, and it then concluded that no Supreme Court decision squarely addressed that situation. It thus concluded that the criteria for the issuance of a writ were not met. But the court also recognized that “[r]easonable jurists could disagree about whether Cronic clearly establishes an exception to Strickland’s prejudice requirement” on these facts, and so it is‐ sued a certificate of appealability to Lewis limited to this is‐ sue. 8 No. 20‐1642

II We begin with a review of the Supreme Court’s Cronic de‐ cision and how it fits within the broader context of ineffective‐ assistance‐of‐counsel cases. For purposes of section 2254(d), the only relevant law is that which is “clearly established Fed‐ eral law, as determined by the Supreme Court of the United States[.]” 28 U.S.C. § 2254(d)(1). Our own decisions, as well as those of other circuits or state courts, are informative only in‐ sofar as they may shed light on our understanding of the au‐ thoritative Supreme Court precedents. 1. Cronic The underlying facts of Cronic involved a common check‐ kiting scheme, in which defendant Cronic and his co‐defend‐ ants relayed checks back and forth between two bank ac‐ counts (one in Tampa, Florida, and the other in Norman, Ok‐ lahoma) in order to create falsely inflated balances in each one. They ran almost $4.8 million through the Tampa bank and $4.6 million through the Norman bank. They ultimately were caught and indicted on federal mail‐fraud charges. Cronic was initially represented by one lawyer, but shortly before trial, that lawyer withdrew from the case. In his place, “[t]he court appointed a young lawyer with a real estate prac‐ tice” to represent Cronic. It allowed the substitute lawyer only 25 days to prepare for trial, even though the government had been working for over four and a half years on the case and “had reviewed thousands of documents.” 466 U.S. at 649. Cronic’s co‐defendants agreed to testify for the government. Despite these disadvantages, Cronic’s lawyer managed to do a few things. He was able to establish some points favora‐ ble to Cronic on cross‐examination. He did not, however, put No. 20‐1642 9

on any defense. In the end, Cronic was convicted on most counts of the indictment and sentenced to 25 years’ imprison‐ ment. His trial lawyer filed a timely appeal for him, although two months later Cronic filed a motion asking that a new law‐ yer be appointed for the appeal. The court of appeals obliged him, though it declined to ap‐ point the lawyer Cronic had requested. It reversed Cronic’s conviction on the basis of several factors: the limited time for investigation and preparation, counsel’s lack of experience, the gravity of the charge, the complexity of possible defenses, and counsel’s access to witnesses. The Supreme Court re‐ versed. Id. at 667. The Court began its analysis by reiterating the critical role that counsel plays in the criminal justice system and the con‐ sequent need to assure counsel’s competence: The right to the effective assistance of counsel is thus the right of the accused to require the prosecution’s case to survive the crucible of meaningful adversarial testing. Id. at 656. It then reaffirmed the general rule, under which “[a]bsent some effect of challenged conduct on the reliability of the trial process, the Sixth Amendment guarantee is gener‐ ally not implicated.” Id. at 658. That said, the Court then spelled out some “circumstances that are so likely to prejudice the accused that the cost of liti‐ gating their effect in a particular case is unjustified.” Id. It sin‐ gled out the following examples:  The complete absence of counsel at a critical stage of the trial. Id. at 659. 10 No. 20‐1642

 Counsel’s total failure to subject the prosecution’s case to meaningful adversarial testing. Id.  Other circumstances under which, “although coun‐ sel is available to assist the accused during trial, the likelihood that any lawyer, even a fully competent one, could provide effective assistance is so small that a presumption of prejudice is appropriate without inquiry into the actual conduct of the trial.” Id. at 659–60. The Court concluded that Cronic had not suffered from an ex‐ treme deprivation along the lines of its examples, and so he could prevail only if he could identify specific errors that trial counsel made. Id. at 666. That issue, the Court held, could be explored on remand. In the years since it was decided, Cronic has made few ap‐ pearances in Supreme Court opinions. The district court iden‐ tified two of those in its effort to distinguish our decision in Miller, 481 F.3d 468: Wright v. Van Patten, 552 U.S. 120 (2008), and Woods v. Donald, 575 U.S. 312 (2015). But neither case of‐ fers help in answering the question we face. Wright v. Van Pat‐ ten concerned whether it was clearly established that a de‐ fense counsel’s appearance at a plea hearing by speakerphone was the equivalent of a complete denial of counsel. The Court ruled that Cronic did not go that far. 552 U.S. at 125. Indeed, recent experience has shown that remote presence is a rela‐ tively good substitute for a great many things. The problem for Lewis was not, as in Van Patten, that his lawyer’s assistance was occurring via telephone or Zoom; it is that his attorney did absolutely nothing for him regardless of format. The Court similarly declined to apply Cronic in Woods v. Donald, a situation in which defense counsel was briefly absent from his No. 20‐1642 11

client’s joint criminal trial and missed the beginning of a gov‐ ernment witness’s testimony about codefendants’ actions. 575 U.S. at 317–19. Lewis does not allege that his attorney was physically absent at any relevant time. Neither Van Patten nor Woods thus advances the analysis here. Cronic is far from a dead letter in the Supreme Court. To the contrary, as recently as October Term 2018, in Garza v. Idaho, 139 S. Ct. 738 (2019), the Court reaffirmed Cronic’s hold‐ ing. Garza was a case in which an attorney in a criminal case failed to file a notice of appeal for a defendant, despite the fact that the defendant asked the attorney to do so. Normally, un‐ der those circumstances prejudice is presumed, regardless of how likely an appeal would be to change the result. See Roe v. Flores‐Ortega, 528 U.S. 470, 484 (2000). But Garza had signed two plea agreements in which he waived his right to appeal. 139 S. Ct. at 742. Based on those waivers, counsel informed Garza that he was not going to initiate an appeal. Garza sought post‐convic‐ tion relief in Idaho’s state courts, but they ruled that he needed to show both deficient performance and prejudice, and that he had not done so. The Supreme Court reversed. It began by reiterating that “in certain Sixth Amendment contexts, … prejudice is presumed.” Id. at 744 (quotations omitted). It elaborated as follows: For example, no showing of prejudice is necessary if the accused is denied counsel at a critical stage of his trial, United States v. Cronic, 466 U.S. 648, 659 (1984), or left entirely without the assistance of counsel on ap‐ peal, Penson v. Ohio, 488 U.S. 75, 88 (1988). Similarly, prejudice is presumed if counsel entirely fails to subject 12 No. 20‐1642

the prosecution’s case to meaningful adversarial test‐ ing. Cronic, 466 U.S. at 659. And, most relevant here, prejudice is presumed when counsel’s constitutionally deficient performance deprives a defendant of an ap‐ peal that he otherwise would have taken. Flores‐Ortega, 528 U.S. at 484. We hold today that this final presump‐ tion applies even when the defendant has signed an appeal waiver. 139 S. Ct. at 744 (cleaned up). With respect to the case before it, the Court underscored that appeal waivers do not inevita‐ bly block all further recourse: some matters might fall outside the scope of a waiver, the prosecution might forfeit the benefit of a waiver, or the waiver might be unenforceable on the ground that it was unknowing or involuntary. Id. at 744–45. It therefore found that the presumption of prejudice recognized in Flores‐Ortega applied to Garza’s case and remanded for fur‐ ther proceedings. Years earlier, the Court had also discussed the Cronic rule, albeit in the context of a proceeding in which it found that the rule did not apply. The case was Nixon, supra, 543 U.S. 175 (2004), a capital case that came to the Supreme Court after the Florida Supreme Court resolved the defendant’s state post‐ conviction petition. Before trial, counsel informed his client, Nixon, that he intended to concede guilt at the outset of the trial, in the hope that this would persuade the jury not to rec‐ ommend death during the penalty phase. Id. at 178. Nixon did not respond one way or the other to this statement, and so counsel followed his planned strategy. The Florida Supreme Court found that counsel’s decision effectively to concede guilt without having Nixon’s express consent to do so amounted to constitutionally ineffective No. 20‐1642 13

performance, and that such a concession of guilt triggered the Cronic presumption of prejudice. Id. at 188–89. The Supreme Court held that the state court erred in both respects. At least in a case such as Nixon’s, where counsel fully informed the client of his proposed trial strategy and the client raised no objection, counsel was entitled to proceed as planned. More‐ over, in a passage that is relevant to our case, it found that this was not an occasion for presumed prejudice: Cronic recognized a narrow exception to Strick‐ land’s holding that a defendant who asserts ineffective assistance of counsel must demonstrate not only that his attorney’s performance was deficient, but also that the deficiency prejudiced the defense. Cronic in‐ structed that a presumption of prejudice would be in order in circumstances that are so likely to prejudice the accused that the cost of litigating their effect in a particular case is unjustified. 466 U.S., at 658. The Court elaborated: “[I]f counsel entirely fails to subject the prosecution’s case to meaningful adversarial test‐ ing, then there has been a denial of Sixth Amendment rights that makes the adversary process itself pre‐ sumptively unreliable.” Id., at 659; see Bell v. Cone, 535 U.S. 685, 696–697 (2002) (for Cronic’s presumed preju‐ dice standard to apply, counsel’s failure must be com‐ plete). We illustrated just how infrequently the sur‐ rounding circumstances [will] justify a presumption of ineffectiveness in Cronic itself. In that case, we reversed a Court of Appeals ruling that ranked as prejudicially inadequate the performance of an inexperienced, un‐ derprepared attorney in a complex mail fraud trial. 466 U.S., at 662, 666. 14 No. 20‐1642

543 U.S. at 190 (cleaned up). See also Kansas v. Ventris, 556 U.S. 586, 591 (2009) (citing Cronic for the proposition that the right to counsel “ensur[es] that the prosecution’s case is subjected to ‘the crucible of meaningful adversary testing’”). Although it is possible, as the Supreme Court itself did in Cronic and as the district court here did, to identify particular circumstances in which the Cronic rule will apply, we must take the Court at its word when it says that it is simply offer‐ ing illustrations of the rule announced by the Court. We have cautioned before that “[j]udicial opinions must not be con‐ fused with statutes, and general expressions must be read in light of the subject under consideration.” United States v. Skoien, 614 F.3d 638, 640 (7th Cir. 2010) (en banc) (citing Zenith Radio Corp. v. United States, 437 U.S. 443, 462 (1978)). More to the point, the Supreme Court itself has expressly stated that section 2254(d)(1) does not demand a clone in prior law. Wil‐ liams v. Taylor, 529 U.S. 362, 407–09 (2000). As the Court has put this point, section 2254(d)(1) does not “require state and federal courts to wait for some nearly identical factual pattern before a legal rule must be applied” nor does it “prohibit a federal court from finding an application of a principle unrea‐ sonable when it involves a set of facts different from those of the case in which the principle was announced.” Panetti v. Quarterman, 551 U.S. 930, 953 (2007) (cleaned up). Instead, “state courts must reasonably apply the rules squarely estab‐ lished by [the Supreme Court]’s holdings to the facts of each case.” White v. Woodall, 572 U.S. 415, 427 (2014). As applied here, that means that we must pay heed to Cronic’s core hold‐ ing: that a showing of prejudice is not necessary in “situations in which counsel has entirely failed to function as the client’s advocate.” Nixon, 543 U.S. at 189. Implicit in this formulation No. 20‐1642 15

is the need to show that this extreme failure occurred during a critical stage of the proceedings. Lewis has done just that. 2. Critical Stage Before proceeding to apply these principles to Lewis’s case, we confirm that the Supreme Court has emphasized for years the “critical nature of sentencing in a criminal case[.]” See Mempa v. Rhay, 389 U.S. 128, 134 (1967). The Court recon‐ firmed that the Sixth Amendment right to counsel exists dur‐ ing the sentencing phase in Lafler v. Cooper, 566 U.S. 156 (2012): The precedents also establish that there exists a right to counsel during sentencing in both noncapital, see Glover v. United States, 531 U.S. 198, 203–204 (2001); Mempa v. Rhay, 389 U.S. 128 (1967), and capital cases, see Wiggins v. Smith, 539 U.S. 510, 538 (2003). Even though sentencing does not concern the defend‐ ant’s guilt or innocence, ineffective assistance of coun‐ sel during a sentencing hearing can result in Strick‐ land prejudice because “any amount of [additional] jail time has Sixth Amendment significance.” Glover, su‐ pra, at 203. 566 U.S. at 165. This passage relates only to the question whether the Sixth Amendment applies at all to proceedings before and after trial; it does not address the distinction be‐ tween a Strickland claim and a Cronic claim; hence, the refer‐ ence to prejudice is of no moment. To the extent a court of appeals decision is relevant, we think it helpful to explain why our en banc decision in Schmidt v. Foster, 911 F.3d 469 (7th Cir. 2018), has no bearing on whether the sentencing phase in Lewis’s case was a “critical phase” for Sixth Amendment purposes. In Schmidt, before 16 No. 20‐1642

defendant Schmidt’s trial on first‐degree intentional homicide charges began, the trial court had to make a decision about whether to admit evidence relating to a defense of adequate provocation. It decided to hold an ex parte, in camera examina‐ tion of Schmidt to help it assess whether he could pursue this defense. Schmidt’s lawyer was present for that examination, but the court admonished counsel not to say anything, and the lawyer complied. The trial court ultimately decided to dis‐ allow the defense; Schmidt was convicted; and after proceed‐ ings that need not detain us, he argued in a petition under section 2254 that he had suffered a complete deprivation of counsel at a critical phase and thus was entitled to relief under Cronic. Bearing in mind the highly deferential approach to state‐ court rulings that section 2254(d) requires, the en banc court rejected Schmidt’s argument. After canvassing the relevant cases, the court first held that the Supreme Court had never had the occasion to consider the unusual circumstances that Schmidt’s case presented: a deprivation of counsel during a pre‐trial, in camera, examination that related to the admissibility of evidence. 911 F.3d at 480. Moreover, the en banc court viewed the course of events through a broader lens than Schmidt had urged. That broader perspective showed that counsel was able to help Schmidt in several ways with respect to the proposed evidence: he filed a notice of the provocation defense, he argued for its application in court hearings, he briefed the law, he submitted a detailed offer of proof, and he gave the court a witness list. The fact that counsel was barred from offering assistance during the in camera hearing, the court said, did not render the rest of counsel’s assistance meaningless. Or at least, the court held, the state courts were No. 20‐1642 17

entitled to view the case this way, and that was enough to re‐ quire denial of the writ. In our case, unlike Schmidt, the Supreme Court has spoken specifically to the question whether the phase in question— sentencing—is a “critical” one. As we noted earlier, the an‐ swer is an unambiguous yes. And as we will see below, the other distinction between Lewis’s case and Schmidt’s is the degree of help that counsel offered—significant for Schmidt, nonexistent for Lewis. III We do not need to decide for ourselves whether trial coun‐ sel’s performance at sentencing was deficient. We have only to defer to the finding of the Indiana Court of Appeals in Lewis II, to which we referred at the outset of this opinion. That court described counsel’s performance as “clearly defi‐ cient,” and we agree with that assessment. Where we part company is with Lewis II’s approach to Cronic. The court be‐ gan by reviewing the three situations that Cronic itself had mentioned and that we noted earlier: ( 1) a complete denial of counsel at a critical stage of trial; (2) the entire failure to sub‐ ject the prosecution’s case to meaningful adversarial testing; and (3) a situation in which counsel is called upon to render assistance under circumstances where even competent coun‐ sel could not do so. Lewis II, ¶ 38. In the state court, Lewis stressed the second of those three considerations, but the court did not confine its analysis to that situation. Rightly so, we think—as we noted earlier, judicial opinions are not stat‐ utes and should not be treated in such a rigid way. But nothing turns on whether we see Cronic as establishing three exclusive categories, or as stating a principle and 18 No. 20‐1642

offering three illustrations. In the end, the state court simply noted (accurately) that the Cronic exception is a narrow one, rarely applied by the Supreme Court. Without another word, it then turned to Strickland, which it read as confining pre‐ sumed prejudice in various ways. 466 U.S. at 692–93. It did not explain why, in the case before it, Lewis had not suffered ex‐ actly the fate the Strickland Court had mentioned: the actual or constructive absolute denial of the assistance of counsel (Cronic category one). Naturally, someone whose lawyer has left him in the lurch that way will also fail to subject the pros‐ ecutor’s case to meaningful adversarial testing (Cronic cate‐ gory two). In cases such as Lewis’s, there is thus no operative difference between the first and the second of Cronic’s exam‐ ples. The closest the state court came to supporting its conclu‐ sion that Cronic does not apply to Lewis’s case is in the fol‐ lowing passage: Moreover, since Cronic was decided in 1984, the U.S. Supreme Court has never applied the second ex‐ ception [i.e. lack of adversarial testing] to relieve a con‐ victed defendant of the need to prove prejudice, nor has the Indiana Supreme Court. In Bell [v. Cone, 535 U.S. 685 (2002)], the Court simply spoke of “the possi‐ bility of presuming prejudice based on an attorney’s failure to test the prosecutor’s case” where the attor‐ ney’s failure is complete. Bell, 535 U.S. at 696–97 (em‐ phasis supplied). Ultimately, the Court concluded in Bell: “The aspects of counsel’s performance challenged by respondent—the failure to adduce mitigating evi‐ dence and the waiver of closing argument—are plainly of the same ilk as other specific attorney errors we have No. 20‐1642 19

held subject to Strickland’s performance and prejudice components.” Id. at 697–98. We are not persuaded that Lewis’s claim falls within one of the limited circumstances of extreme magnitude that justify a presumption of ineffective‐ ness under Cronic. The post‐conviction court, there‐ fore, correctly determined that Lewis was required to establish prejudice under Strickland. Lewis II at ¶¶ 42, 43. Entirely missing from the state court’s brief discussion is an acknowledgment of the day‐and‐night difference between the assistance that Cone received during the sentencing phase of his case and that which Lewis got. Cone, in a word, had plenty of help. The sentencing hearing was a separate part of Cone’s trial. The state opened by telling the jury that it planned to prove four aggravating factors that would justify the death penalty. Defense counsel responded in his own opening statement by calling to the jury’s attention “the miti‐ gating evidence already before them”, and by suggesting that Cone “was under the influence of extreme mental disturbance or duress, that he was an addict whose drug and other prob‐ lems stemmed from the stress of his military service, and that he felt remorse.” 535 U.S. at 691. Counsel also urged the jury to be merciful. Already, we note, Cone received far more than Lewis did. But that was just the start for Cone—there was much more. His lawyer cross‐examined the state’s witnesses and objected to the introduction of gory photographs. He chose to waive final argument because this prevented the state from arguing in rebuttal. It is hardly a surprise that the Supreme Court did not regard Cone’s lawyer’s performance as either the equivalent of a total lack of counsel, or the 20 No. 20‐1642

“entire” failure to subject the prosecutions’ case to meaning‐ ful adversarial testing. It was neither. Let’s take another look at Attorney Raff’s “assistance” dur‐ ing the entire sentencing phase. In essence, it was nothing but a statement that he was bowing out. He uttered two short sen‐ tences: “Judge, I’m going to defer to Mr. Lewis if he has any comments. I don’t have anything to add.” This went beyond a failure to conduct adversarial testing; it was an announce‐ ment of abandonment. The state suggests that Raff did have a strategy, and that was to allow Lewis to speak for himself in the hope that he might express remorse. This has the flaw of having no support in the record. Raff never communicated any such strategy to Lewis, and so Lewis had no guidance from counsel about what he might do with his allocution when he had the chance to speak. This theory also conflicts with Raff’s testimony at the post‐conviction hearing. He never said that he was trying to guide Lewis in this way. Instead, he said that he thought that there were no mitigating factors in Lewis’s case. Actually, he had no idea one way or the other, because he never asked Lewis about his mental‐health history and he never requested Lewis’s medical records. He did not try to prepare Lewis for the hearing because he found Lewis “difficult” and “angry.” It is of no moment that four jurists (whom we presume were acting in good faith) disagreed with Cronic’s application. The question is an objective one and does not rest “on the simple fact that at least one of the Nation’s jurists has applied the relevant federal law in the same man‐ ner the state court did … .” Williams, 529 U.S. at 409–10. Contrary to the dissent’s assertions, our opinion in no way conflicts with the holdings in Woods v. Donald, Cone, or Nixon. None of these cases involved the total absence of counsel (or No. 20‐1642 21

its functional equivalent) at a critical stage. That is what we have here. By contrast, in Woods counsel was briefly absent during the testimony of a co‐defendant. As discussed earlier, counsel in Nixon fully informed his client of his proposed strategy, and counsel in Cone subjected the prosecution’s case to adversarial testing. Today’s outcome faithfully follows Cronic, because we are faced with the extraordinary situation of a lawyer’s total abandonment of his client at the critical sen‐ tencing state. IV If Raff was going to fall back to a plea for mercy, or an ef‐ fort to convince Lewis to demonstrate remorse, he had to take some step in that direction. He did not. Instead, he gave up on Lewis and left him entirely without the assistance of counsel at the sentencing stage of a felony murder case. Rare though Cronic cases may be, we think that this one qualifies. We therefore REVERSE the judgment of the district court and REMAND this case for the issuance of a writ of habeas cor‐ pus, limited to the sentencing phase of petitioner Roderick Lewis’s case. 22 No. 20‐1642

BRENNAN, Circuit Judge, dissenting. If this were a direct ap‐ peal, I might join the majority opinion. All can agree Roderick Lewis’s counsel should have done more on his behalf at sen‐ tencing. Such minimal involvement occurred at a critical stage in a criminal case. But Lewis’s appeal comes to us as a collat‐ eral attack on a state court judgment under the Antiterrorism and Effective Death Penalty Act of 1996 (“AEDPA”). In his petition for a writ of habeas corpus, Lewis contends his coun‐ sel’s silence at sentencing requires us to apply the presump‐ tion of prejudice described in United States v. Cronic, 466 U.S. 648 (1984), to an ineffective assistance of counsel claim other‐ wise governed by Strickland v. Washington, 466 U.S. 668 (1984). AEDPA requires that we grant habeas relief to Lewis only when the Supreme Court has answered the specific question of whether Cronic—and not Strickland—applies, and the state court has issued a decision contravening that answer. 28 U.S.C. § 2254(d)(1). No Supreme Court decision holds that silence at sentenc‐ ing by defense counsel triggers Cronic’s presumption of prej‐ udice. Three courts have declined Lewis’s invitation to apply Cronic in this novel circumstance. Despite the stringent stand‐ ards of AEDPA, our court accepts the invitation. This decision avoids AEDPA’s confines and expands Cronic’s scope, read‐ ing it too generally and combining its exceptions. Review of Lewis’s habeas petition should end with AEDPA, so I respect‐ fully dissent. I. AEDPA Review AEDPA’s strict standard of review results in great defer‐ ence to state courts. The grant of Lewis’s habeas petition lacks the requisite precision under 28 U.S.C. § 2254(d)(1), neglecting No. 20‐1642 23

the critical importance of comity to our federal habeas system. A. AEDPA’s Strictures AEDPA deference is more than a judicial guidepost; it is a Congressional mandate. See Woodford v. Garceau, 538 U.S. 202, 206 (2003). “Section 2254(d) reflects the view that habeas cor‐ pus is a ‘guard against extreme malfunctions in the state crim‐ inal justice systems,’ not a substitute for ordinary error correction through appeal.” Harrington v. Richter, 562 U.S. 86, 102–03 (2011) (quoting Jackson v. Virginia, 443 U.S. 307, 332 n.5 (1979) (Stevens, J., concurring in judgment)). By its plain text, AEDPA precludes a federal court from granting a state pris‐ oner’s habeas petition unless the state court’s merits adjudi‐ cation “resulted in a decision that was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States[.]” 28 U.S.C. § 2254(d)(1) (emphases added). “If this standard is difficult to meet, that is because it was meant to be.” Harring‐ ton, 562 U.S. at 102. Section 2254(d)(1)’s “contrary to” and “unreasonable ap‐ plication of” clauses have independent meaning. Bell v. Cone, 535 U.S. 685, 694 (2002); see also Williams v. Taylor, 529 U.S. 362, 404–05 (2000). A federal court may grant habeas relief under the “contrary to” clause “if the state court applies a rule dif‐ ferent from the governing law set forth in our cases, or if it decides a case differently than we have done on a set of mate‐ rially indistinguishable facts.” Cone, 535 U.S. at 694. For exam‐ ple, if a state court applies Strickland when it should apply Cronic, we may issue the writ as the state court judgment is “contrary to” Supreme Court precedent. Cf. id. at 698 (reject‐ ing a petitioner’s claim that Cronic, not Strickland, should ap‐ ply and noting “we find no merit in respondent’s contention 24 No. 20‐1642

that the state court’s adjudication was contrary to our clearly established law” (emphasis added)). Under the “unreasona‐ ble application of” clause, a federal court may grant habeas relief “if the state court correctly identifies the governing legal principle from our decisions but unreasonably applies it to the facts of the particular case.” Id. at 694. So if a state court applies Strickland to the facts of a case “in an objectively un‐ reasonable manner[,]” we may issue the writ. Id. at 699. Under either clause of § 2254(d)(1), a petitioner’s habeas claim is measured against the last reasoned state‐court deci‐ sion on the merits. Wilson v. Sellers, 138 S. Ct. 1188, 1192 (2018). To grant relief, that state court’s decision must be “contrary to” or an “unreasonable application of” Supreme Court prec‐ edent, not our own. Glebe v. Frost, 574 U.S. 21, 24 (2014) (per curiam) (“[C]ircuit precedent does not constitute ‘clearly es‐ tablished Federal law, as determined by the Supreme Court.’” (quoting 28 U.S.C. § 2254(d)(1)). Standing alone, AEDPA exudes deference. But for ineffec‐ tive assistance of counsel claims, “[t]he federal courts as a whole engage in ‘doubly deferential’ review” under AEDPA. Wilborn v. Jones, 964 F.3d 618, 620 (7th Cir. 2020), cert. denied, No. 20‐913, 2021 WL 666799 (U.S. Feb. 22, 2021) (quoting Knowles v. Mirzayance, 556 U.S. 111, 123 (2009)). We layer def‐ erence upon deference in these cases because federal courts must give “both the state court and the defense attorney the benefit of the doubt.” Burt v. Titlow, 571 U.S. 12, 15 (2013). Even without AEDPA, ineffective assistance of counsel claims remain difficult to prove as “counsel is strongly presumed to have rendered adequate assistance and made all significant decisions in the exercise of reasonable professional judg‐ ment.” Strickland, 466 U.S. at 690. “That hill is even steeper” No. 20‐1642 25

for claims governed by AEDPA. Myers v. Neal, 975 F.3d 611, 620 (7th Cir. 2020). As the Supreme Court said recently, AEDPA takes on a “special importance” when a state prisoner asserts the ineffectiveness of his counsel. Shinn v. Kayer, 141 S. Ct. 517, 523 (2020). B. The Majority Opinion’s Application of AEDPA The majority opinion holds that the Court of Appeals of Indiana’s decision in Lewis v. State, 116 N.E.3d 1144 (Ind. Ct. App. 2018) (Lewis II), is both “contrary to” and an “unreason‐ able application of” Supreme Court precedent under § 2254(d)(1). In its only specific reference to the text of that standard, it states: “We conclude that the decision of the last responsible state court was contrary to Supreme Court prece‐ dent, insofar as it held that Strickland, not Cronic, furnished the applicable rule, and it was an unreasonable application of Cronic, insofar as it focused on that case.” Majority Op. at p. 2 (footnote omitted). But these clauses are distinct, with each having independent meaning. Cone, 535 U.S. at 695 (“[Section] 2254(d)(1)’s ‘contrary to’ and ‘unreasonable application’ clauses have independent meaning.”). Implicit within an “un‐ reasonable application of” Supreme Court precedent is that the state court applied the correct legal rule but did so unrea‐ sonably to the facts at hand. Conversely, if a state court ap‐ plies the incorrect rule, its decision is “contrary to” Supreme Court precedent from the start and we need not reach the rea‐ sonability of its application. For the majority opinion, Lewis II is worthy of correction under both clauses. AEDPA requires more precision. Cone teaches that if a state court applies Strickland when it should apply Cronic (or vice versa), that error implicates the “contrary to” clause of § 2254(d)(1). See Cone, 535 U.S at 698. This is because “[f]or 26 No. 20‐1642

purposes of distinguishing between the rule of Strickland and that of Cronic, this difference is not of degree but of kind.” Id. at 697 (footnote omitted). Admittedly, the Supreme Court has not always followed this procedure when engaging with Cronic. Compare Wright v. Van Patten, 552 U.S. 120, 126 (2008) (per curiam) (rejecting Cronic’s application under the “unrea‐ sonable application of” clause), with Woods v. Donald, 575 U.S. 312, 317–18 (2015) (per curiam) (rejecting Cronic’s application under both the “contrary to” and “unreasonable application of” clauses). That may be because, in these cases, the Supreme Court has denied relief, not granted it. In other words, whether a state court’s decision was “contrary to” or an “un‐ reasonable application of” Supreme Court precedent did not matter. The petitioner’s claim failed either way. When a peti‐ tioner’s claim succeeds, however, and a federal court on ha‐ beas review overrules a state court’s decision, precision is a must. Otherwise, we disregard AEDPA’s text and do not re‐ spect the independent meaning of each clause.1 To be sure, the “contrary to” clause might be an easier path for a habeas petitioner than the “unreasonable applica‐ tion of” clause. But see Williams, 529 U.S. at 405 (“The word ‘contrary’ is commonly understood to mean ‘diametrically

1 The majority opinion collects several cases to support consideration of the state court’s decision under both clauses of § 2254(d)(1). Maj. Op. at p. 2 n.1. But among the cases cited are rulings that recognized the inde‐ pendent meaning of each clause, even if the state court decision at issue in the end violated both. In Bailey v. Rae, the Ninth Circuit explained the dis‐ tinct clauses, and gave independent reasoning under each. 339 F.3d 1107, 1111–12, 1118–19 (9th Cir. 2003). Pazden v. Maurer is much the same. 424 F.3d 303, 311–12, 319 (3d Cir. 2005), as is Breakiron v. Horn, 642 F.3d 126, 131, 139 (3d Cir. 2011). To the extent these cases correctly state the law, they recognize that the two clauses differ. No. 20‐1642 27

different,’ ‘opposite in character or nature,’ or ‘mutually op‐ posed.’” (quoting WEBSTER’S THIRD NEW INTERNATIONAL DICTIONARY 495 (1976)). An “unreasonable application” means more than just error: “The question under AEDPA is not whether a federal court believes the state court’s determi‐ nation was incorrect but whether that determination was un‐ reasonable—a substantially higher threshold.” Schriro v. Landrigan, 550 U.S. 465, 473 (2007). Or as this court has said, “[w]e must deny the writ if we can posit arguments or theo‐ ries that could have supported the state court’s decision, and if fairminded jurists could disagree about whether those ar‐ guments or theories are inconsistent with Supreme Court holdings.” Kidd v. Lemke, 734 F.3d 696, 703 (7th Cir. 2013). To grant Lewis habeas relief under the “unreasonable applica‐ tion of” clause, the state court’s application of Strickland, not Cronic, must be unreasonable; that is how the Supreme Court approached Cone, and that is how to ensure the independent meaning of the two clauses. See Cone, 535 U.S. at 698. Yet even if the state court unreasonably applied Cronic, the implication of the majority opinion is that no “fairminded ju‐ rist could disagree” that Cronic should supplant Strickland here. On direct review, four jurists disagreed with Cronic’s ap‐ plication; two more did so under AEDPA.2 Although this nu‐ merical disparity does not alone doom Lewis’s appeal, it shows that under the majority opinion, this would be the only court that has embraced Cronic, doing so under both clauses of § 2254(d)(1). If AEDPA is satisfied here, the clause under

2One judge from the Allen County Superior Court, three judges from the Court of Appeals of Indiana, one judge from the U.S. District Court for the Northern District of Indiana, and myself. 28 No. 20‐1642

which the state court’s decision is purportedly incorrect should be specified as one or the other. Animating AEDPA’s strictness is a faith in comity. “AEDPA recognizes a foundational principle of our federal system: State courts are adequate forums for the vindication of federal rights.” Titlow, 571 U.S. at 19. Indeed, “AEDPA’s re‐ quirements reflect a ‘presumption that state courts know and follow the law.’” Woods, 575 U.S. at 316 (quoting Woodford v. Visciotti, 537 U.S. 19, 24 (2002) (per curiam)). That means “[w]hen reviewing state criminal convictions on collateral re‐ view, federal judges are required to afford state courts due respect by overturning their decisions only when there could be no reasonable dispute that they were wrong.” Woods, 575 U.S. at 316. This is particularly so when state courts adjudicate ineffective assistance of counsel claims: Especially where a case involves such a common claim as ineffective assistance of counsel under Strickland—a claim state courts have now adjudicated in countless criminal cases for nearly 30 years—“there is no intrin‐ sic reason why the fact that a man is a federal judge should make him more competent, or conscientious, or learned … than his neighbor in the state courthouse.” Titlow, 571 U.S. at 19 (quoting Stone v. Powell, 428 U.S. 465, 494 n.35 (1976)). AEDPA review is so rigorous for Strickland claims because comity demands it. See Calderon v. Thompson, 523 U.S. 538, 555–56 (1998) (“Federal habeas review of state convic‐ tions frustrates both the States’ sovereign power to punish of‐ fenders and their good‐faith attempts to honor constitutional rights.” (internal quotation marks omitted)). No. 20‐1642 29

Comity’s force here is not just in principle, but in practice. The Indiana state courts have worked on Lewis’s case for some time and have a significant interest in this litigation. In 2012, Lewis went to trial in the Allen County Superior Court for his role in a crime committed 13 years earlier in 1999. Lewis II, 116 N.E.3d at 1148–50, ¶¶ 6–12. After his conviction and sentencing, Lewis directly appealed, challenging the suffi‐ ciency of the evidence against him. Lewis v. State, 973 N.E.2d 110, No. 02A03‐1201‐CR‐18, 2012 WL 3777134 (Ind. Ct. App. 2012) (unpublished table decision) (Lewis I). In 2013, Lewis filed a pro se post‐conviction petition, and in 2016, post‐con‐ viction counsel amended that petition. Lewis II, 116 N.E.3d at 1150, ¶ 14. In 2017, the Allen County Superior Court held an evidentiary hearing. Id. In 2018, that court, in a “lengthy or‐ der[,]” denied Lewis relief and, as relevant in this appeal, also rejected Cronic. Id at 1150–53, ¶¶ 16–17. In the decision at issue here, Lewis II, the Court of Appeals of Indiana affirmed that denial of relief later in 2018, id. at 1160, ¶¶ 45–46, with the Indiana Supreme Court ultimately denying leave to transfer in 2019. Lewis v. State, 124 N.E.3d 41 (Ind. 2019) (unpublished table decision). This procedural history shows that Lewis’s case has received thorough consideration by various Indiana courts, not to mention the district court here. AEDPA makes clear “that state courts are the principal forum for asserting constitutional challenges to state convictions.” Harrington, 562 U.S. at 103. All of this is true even before accounting for the “special importance” of AEDPA to ineffective assistance of counsel claims adjudicated by state courts. Shinn, 141 S. Ct. at 523. Ab‐ sent from the majority opinion’s treatment of Lewis’s claim is recognition of the “doubly deferential” nature of the review this court must conduct. Mirzayance, 556 U.S. at 123. Because 30 No. 20‐1642

“state courts know and follow the law[,]” Woods, 575 U.S. at 316 (internal quotation marks omitted), they understand that the scope of Cronic has been significantly curtailed; indeed, that is why they declined to apply it here. Lewis II, 116 N.E.3d at 1159, ¶ 43 (“We are not persuaded that Lewis’s claim falls within one of the limited circumstances of extreme magnitude that justify a presumption of ineffectiveness under Cronic.” (footnote omitted)). All told, the type of claim the petitioner makes requires him to overcome AEDPA, surpass Strickland, and trigger Cronic. That is quite the gauntlet. In fact, it is one of the most doctrinally difficult challenges a state prisoner can make in this area of law. AEDPA’s text, along with its di‐ rective of deference, instructs that we must give independent meaning to each clause of § 2254(d)(1). II. The Narrow Scope of Cronic The Supreme Court and this court narrowly construe and rarely apply Cronic. The majority opinion expands Cronic’s scope and unsoundly combines its first and second excep‐ tions. A. Cronic Defined Cronic is a hard‐to‐meet exception to the already hard‐to‐ meet standard of Strickland. Under Strickland, an ineffective assistance of counsel claim requires a showing that the attor‐ ney’s performance was not only deficient, but also prejudicial. 466 U.S. at 687. Deficiency occurs when “counsel made errors so serious that counsel was not functioning as the ‘counsel’ guaranteed the defendant by the Sixth Amendment.” Id. Proving prejudice requires that “there is a reasonable proba‐ bility that, but for counsel’s unprofessional errors, the result of the proceeding would have been different.” Id. at 694. No. 20‐1642 31

Strickland sets a high bar, which makes Cronic’s presumption of prejudice an appealing option for litigants. Cf. Padilla v. Kentucky, 559 U.S. 356, 371 (2010) (“Surmounting Strickland’s high bar is never an easy task.”). With Cronic’s strength, though, comes its rarity. See, e.g., Florida v. Nixon, 543 U.S. 175, 190 (2004). Under Cronic, courts may presume prejudice only when there are “circumstances that are so likely to prejudice the accused that the cost of liti‐ gating their effect in a particular case is unjustified.” 466 U.S. at 658 (footnote omitted). But even Cronic itself did not result in this presumption of prejudice. Id. at 666. Instead, the Su‐ preme Court in Cronic pronounced three exceptions to Strick‐ land that permit the presumption of prejudice: 1. When there has been a “complete denial of coun‐ sel[,]” because “a trial is unfair if the accused is denied counsel at a critical stage of his trial.”; 2. “[I]f counsel entirely fails to subject the prosecution’s case to meaningful adversarial testing, then there has been a denial of Sixth Amendment rights that makes the adversary process itself presumptively unrelia‐ ble.”; and 3. “[W]hen although counsel is available to assist the accused during trial, the likelihood that any lawyer, even a fully competent one, could provide effective as‐ sistance is so small that a presumption of prejudice is appropriate without inquiry into the actual conduct of the trial.” 466 U.S. at 659–60 (footnote omitted). When a defendant’s case presents one of these three circumstances, Cronic is 32 No. 20‐1642

triggered, prejudice is presumed, and Strickland’s second prong is satisfied. Cronic’s narrowness derives not just from its result, but from its reasoning. Whether called “illustrations,” “exam‐ ples,” “circumstances,” “scenarios,” or “situations,” what matters is that each operates as an exception to the onerous Strickland standard, and that there are three—and only three—of them. See, e.g., Cone, 535 U.S. at 695 (identifying “three situations implicating the right to counsel that in‐ volved circumstances ‘so likely to prejudice the accused that the cost of litigating their effect in a particular case is unjusti‐ fied’” (quoting Cronic, 466 U.S. at 658)); Reynolds v. Hepp, 902 F.3d 699, 705 (7th Cir. 2018) (same). That Cronic is triggered in only three ways reflects the narrowness with which its pre‐ sumption should be applied. Recently, this court endorsed this understanding of Cronic in Schmidt v. Foster, an en banc decision declining under AEDPA to presume prejudice when a trial judge conducted an ex parte, in camera examination without defense counsel’s active participation. 911 F.3d 469, 478 (7th Cir. 2018) (en banc). This court began its analysis in Schmidt by noting that Cronic and its progeny “come with two caveats.” Id. at 479. First, Cronic’s presumption of prejudice is “narrow” and “arises only when the denial of counsel is extreme enough to render the prosecution presumptively unreliable.” Schmidt, 911 F.3d at 479. And second, because the Supreme Court has spoken only generally about Cronic, “the ‘precise contours’ of these rights ‘remain unclear.’” Schmidt, 911 F.3d at 479 (quoting Woods, 575 U.S. at 318). That means “[s]tate courts therefore enjoy broad discretion in their adjudication of them.” Schmidt, 911 F.3d at 479 (footnote and internal quotation marks No. 20‐1642 33

omitted). Schmidt recognized the three exceptions of Cronic and rejected applying the first—“the complete denial of coun‐ sel during a critical stage.” 911 F.3d at 478 & n.2, 480. Because Schmidt’s counsel had assisted him before, during a recess, and after the in camera examination, he did not suffer the “complete” deprivation of counsel necessary to presume prej‐ udice under Cronic’s first exception, even though the trial court prevented his counsel from speaking during the in chambers hearing. Id. at 480–85.3 Although Schmidt’s outcome may not control, and these facts differ, this court’s reasoning is instructive. Bound by AEDPA, this court engaged in a comprehensive examination of Cronic and cases interpreting it to decide that “[n]o clearly established holding of the Supreme Court mandate[d]” the presumption of prejudice for an ex parte, in camera examina‐ tion without defense counsel’s active participation. 911 F.3d at 481. In doing so, this court rejected an attempt to generalize what, under AEDPA, must be specific: “If we must take sev‐ eral dissimilar decisions and reduce them to blanket princi‐ ples in order to arrive at a general proposition applicable here, the proposition is ‘far too abstract to establish clearly the spe‐ cific rule’ [petitioner] needs.” Id. at 483–84 (quoting Lopez v. Smith, 574 U.S. 1, 6 (2014) (per curiam)). Relevant here, this court also observed that “the Supreme Court has never ad‐ dressed a case like [Schmidt’s].” Schmidt, 911 F.3d at 485. So “[w]ithout clearly established law mandating relief, we [could not] grant it under AEDPA.” Id.

3 Schmidt assumed that the ex parte, in camera examination was a “crit‐ ical stage.” 911 F.3d at 480. 34 No. 20‐1642

B. Cronic Applied The Supreme Court rarely applies Cronic, and when it does, it reads the decision narrowly. Since Cronic’s advent nearly 40 years ago, the Supreme Court has applied it only once to presume prejudice. See Penson v. Ohio, 488 U.S. 75, 88 (1988) (holding that “the presumption of prejudice must ex‐ tend as well to the denial of counsel on appeal” when the granting of an attorney’s motion to withdraw had left the pe‐ titioner “entirely without the assistance of counsel on ap‐ peal”). Although the majority opinion is correct that the Supreme Court recently cited Cronic in Garza v. Idaho, 139 S. Ct. 738, 744 (2019), the Supreme Court has not applied Cronic to grant relief since Penson. This court has followed the Supreme Court’s lead, reading Cronic narrowly and applying it rarely. See, e.g., Schmidt, 911 F.3d at 479. Lewis musters only two cases where this court applied Cronic’s presumption on habeas review. Miller v. Mar‐ tin, 481 F.3d 468, 473 (7th Cir. 2007) (post‐AEDPA); Patrasso v. Nelson, 121 F.3d 297, 305 (7th Cir. 1997) (pre‐AEDPA). Other circuit courts have taken this same guarded approach. See, e.g., United States v. Roy, 855 F.3d 1133, 1144 (11th Cir. 2017) (en banc) (“The difficulty of carrying that ‘very heavy’ burden and the ‘very narrow’ scope of the Cronic exception are evi‐ dent from the fact that the Supreme Court has repeatedly re‐ fused to find it applicable.”). Exceptions by their nature are narrow, so it is no surprise that the Supreme Court has limited Cronic to the three de‐ scribed above. See, e.g., Cone, 535 U.S. at 695. When the Court has not specifically listed the three exceptions, it has still de‐ nied relief. See, e.g., Woods, 575 U.S. at 317–18, 319. This court has read Cronic the same way. See Schmidt, 911 F.3d at 478 & No. 20‐1642 35

n.2. Even in Miller, when this court applied Cronic’s presump‐ tion, it did so within the three‐exception framework. Miller, 481 F.3d at 472–73. Patrasso applied Cronic’s second exception, too. 121 F.3d at 303–05. Neither the Supreme Court nor this court has adopted the broad reading of Cronic in the majority opinion. For the majority opinion, “[a]lthough it is possible, as the Supreme Court itself did in Cronic and as the district court here did, to identify particular circumstances in which the Cronic rule will apply, we must take the Court at its word when it says that it is simply offering illustrations of the rule announced by the Court.” Maj. Op. at p. 14. But the Court in Cone and this court in Schmidt did not read Cronic so expan‐ sively, as we recently acknowledged. Cf. Fayemi v. Ruskin, 966 F.3d 591, 594 (7th Cir. 2020) (“We have been told not to extend Cronic on collateral review.”). If adopted in future cases, this broad conception of Cronic could swallow Strickland’s prejudice prong. For example, if Cronic is read this broadly, several Supreme Court cases should have come out differently. Counsel would have trig‐ gered Cronic by conceding guilt in a capital case (Nixon), fail‐ ing to affirmatively mount some case for life imprisonment in a capital case’s penalty phase (Cone), and being absent during certain trial testimony concerning a codefendant (Woods). That each request to presume prejudice in these cases failed should give pause before applying such an expansive reading of Cronic. Even where the majority opinion does engage with the three exceptions, it makes general under Cronic what must be specific under AEDPA. The majority opinion combines Cronic’s first and second exceptions. See Maj. Op. at p. 18 (“In cases such as Lewis’s, there is thus no operative difference 36 No. 20‐1642

between the first and the second of Cronic’s examples.”). Ra‐ ther than “nothing” turning on whether Cronic established three exclusive categories, id. at 17, its three exceptions have independent meaning, like the two clauses of § 2254(d)(1). Cf. Cone, 535 U.S. at 685. Schmidt recognized this three‐exception framework and rejected applying Cronic’s first exception for complete denial of counsel at a critical stage. Schmidt, 911 F.3d at 478 & n.2, 480, 485. As we stated there, “only once in the thirty‐plus years since Cronic has the Court applied the pre‐ sumption of prejudice it described in a critical‐stage case.” Id. at 479 (citing Penson, 488 U.S. at 88). And in no case since Cronic has the Supreme Court applied the presumption of prejudice described in the second exception.4 Yet despite the strictures of AEDPA, the rarity of Cronic, and the narrowness with which that case has been applied, the majority opinion finds Lewis’s claim strong enough to fit both exceptions. This court should not backtrack from the understanding of Cronic endorsed in Schmidt. According to the majority opin‐ ion, Schmidt differs in its “critical stage” and “degree of help that counsel offered.” Maj. Op. at pp. 15, 17. But Schmidt’s rel‐ evance here is in its mode of analysis. That en banc decision teaches three lessons about Cronic: it is narrow in its rule, it gives state courts “broad discretion” in adjudicating the ap‐ plication of its exceptions, and it has three—and only three— exceptions. Schmidt, 911 F.3d at 478 & n.2. These lessons led this court in Schmidt to address that Cronic‐based habeas petition with the requisite particularity under AEDPA. Generalizing Cronic did not win the day there

4 No Supreme Court case since Cronic appears to have applied its third

exception, either. No. 20‐1642 37

and should not do so here. Id. at 483–84 (“If we must take sev‐ eral dissimilar decisions and reduce them to blanket princi‐ ples in order to arrive at a general proposition applicable here, the proposition is far too abstract to establish clearly the spe‐ cific rule [petitioner] needs.” (internal quotation marks omit‐ ted)). As AEDPA required, Schmidt considered whether the Supreme Court had ever addressed a claim like that raised by the petitioner. It had not, so this court denied relief. Id. at 485. Because the same is true here, Schmidt should guide us to re‐ ject Lewis’s habeas petition. III. Supreme Court Treatment of Cronic and AEDPA The Supreme Court has never confronted the novel cir‐ cumstances presented by Lewis’s claim. That should be enough to preclude habeas relief under AEDPA. The majority opinion emphasizes one case—Cone, where the Supreme Court declined to presume prejudice—at the expense of the rest of Cronic’s progeny. A. What AEDPA Means for Cases Invoking Cronic Looking through AEDPA’s lens, we may grant habeas re‐ lief only when the Supreme Court has answered the “specific question” of whether Cronic—and not Strickland—applies and the state court has issued a decision “contrary to” this answer. Woods, 575 U.S. at 317 (quoting Lopez, 574 U.S. at 6); 28 U.S.C. § 2254(d)(1). Lewis cannot meet this heavy burden imposed by AEDPA. A direct appeal would be a lighter lift, as it would turn on whether his lawyer’s silence at sentencing fell within Cronic. AEDPA, however, constricts our review and requires that we ask whether the Supreme Court has held that silence at sentencing triggers the presumption of prejudice—the 38 No. 20‐1642

“specific question.” Woods, 575 U.S. at 317 (internal quotation marks omitted). This difference is dispositive: A tour through four Supreme Court cases addressing Cronic—Cone, Nixon, Van Patten, and Woods—demonstrates that the Supreme Court has never presumed prejudice based on the type of claim Lewis brings. That is “[a]ll that matters here[.]” Id. at 319. Time and again, the Supreme Court has declined to apply Cronic. In Cone, the Court considered and rejected an argu‐ ment that Cronic’s second exception for lack of meaningful ad‐ versarial testing applied when counsel failed to “mount some case for life [imprisonment]” in a capital case’s penalty phase. Cone, 535 U.S. at 696 (internal quotation marks omitted). In Nixon, the Court, outside the strictures of AEDPA, again re‐ jected application of Cronic’s second exception by holding that a concession of guilt in a capital case “does not rank as a failure to function in any meaningful sense as the Govern‐ ment’s adversary.” Nixon, 534 U.S. at 190 (quoting Cronic, 466 U.S. at 666 (footnote omitted)).5 In Van Patten, the Court de‐ nied Cronic’s first exception because its own precedents “do not clearly hold that counsel’s participation by speakerphone should be treated as a ‘complete denial of counsel,’ on par with total absence.” Van Patten, 552 U.S. at 125. And in Woods, the Court’s most recent engagement with Cronic, it avoided presuming prejudice, without mentioning a specific excep‐ tion, because it had “never addressed whether the rule announced in Cronic applies to testimony regarding code‐ fendants’ actions”—as in whether counsel’s absence during

5The standard the majority opinion quotes when invoking Cronic to grant relief is from Nixon, 543 U.S. at 189. Maj. Op. at pp. 1, 14. The Court in Nixon declined to presume prejudice under Cronic. Nixon, 543 U.S. at 190. No. 20‐1642 39

that testimony triggers the presumption of prejudice. Woods, 575 U.S. at 317. The Supreme Court has been reluctant to presume preju‐ dice under Cronic. We should be as well. Of course, rare does not mean never. What matters under AEDPA is that the Court has never answered, let alone affirmatively, the question of whether silence at sentencing by defense counsel triggers Cronic. See Van Patten, 552 U.S. at 126 (“Because our cases give no clear answer to the question presented, let alone one in Van Patten’s favor, it cannot be said that the state court unreason‐ ably applied clearly established Federal law.” (alterations and internal quotation marks omitted)). This is so regardless of the clause under which the majority opinion proceeds. See Woods, 575 U.S. at 317 (“Because none of our cases confront the spe‐ cific question presented by this case, the state court’s decision could not be contrary to any holding from this Court.” (inter‐ nal quotation marks omitted)). As the Court said in its last en‐ gagement with Cronic under AEDPA, “[a]ll that matters here, and all that should have mattered to the Sixth Circuit, is that we have not held that Cronic applies to the circumstances pre‐ sented in this case. For that reason, federal habeas relief based upon Cronic is unavailable.” Id. at 319. So it is here for this court. B. Correctly Applying AEDPA to Cronic in this Case Under either clause of § 2254(d)(1), the majority opinion’s broad reading of Cronic cuts against applying it on these facts. Such expansive treatment of Cronic comes closer to de novo review, which AEDPA does not permit. Lewis II is neither “contrary to” nor an “unreasonable ap‐ plication of” Supreme Court precedent concerning Cronic. To 40 No. 20‐1642

implicate the “contrary to” clause, a state court decision must apply the wrong legal rule or deviate from a factually indis‐ tinguishable case. See Cone, 535 U.S. at 694. If the majority opinion proceeds under that clause, it commits the same er‐ rors the Supreme Court corrected in Woods, 575 U.S. at 317 (noting that the Sixth Circuit’s application of Cronic was “dou‐ bly wrong”). First, as in Woods, the majority opinion relies upon cases that are only “similar to” Supreme Court prece‐ dent, which means “the state’s court’s decision is not ‘con‐ trary to’ the holdings in those cases.” Id. But the Sixth Circuit in Woods at least had affirmative case law to rely on. None of the decisions cited in the majority opinion granted relief un‐ der Cronic. Second, the majority opinion “frame[s] the issue at too high a level of generality.” Woods, 575 U.S. at 318. As dis‐ cussed, no Supreme Court case holds that silence at sentenc‐ ing triggers Cronic’s presumption of prejudice. Cf., e.g., id. at 317–19; Van Patten, 552 U.S. at 124–26; Nixon, 543 U.S. at 190‐93; Cone, 535 U.S. at 693–98; but see Penson, 488 U.S. at 88‐89. In sum, like the petitioner in Woods, Lewis cannot show that Lewis II was “contrary to” Supreme Court precedent. The fate of Lewis’s claim is the same under the “unreason‐ able application of” clause of § 2254(d)(1). Under that clause, the broader the rule, the more room state courts have to apply it. Yarborough v. Alvarado, 541 U.S. 652, 664 (2004) (“The more general the rule, the more leeway courts have in reaching out‐ comes in case‐by‐case determinations.”). For our purposes, the broader the majority opinion reads Cronic, the more room Lewis II has to apply it. Cf. Maj. Op. at p. 14. This means that Lewis II must be “not merely wrong” or “even clear error.” Woods, 575 U.S. at 316 (quoting Woodall, 572 U.S. at 419). Ra‐ ther, it must be so objectively unreasonable that no No. 20‐1642 41

“fairminded jurist” could reach its conclusion. Kidd, 734 F.3d at 703. The question then remains: Is Lewis II that wrong? It is not. Given that the Supreme Court almost never ap‐ plies Cronic, the Court of Appeals of Indiana’s similar reti‐ cence is reasonable. The majority opinion faults Lewis II for its scant reasoning, despite the “broad discretion” conferred to state courts interpreting Cronic. Woods, 575 U.S. at 318 (inter‐ nal quotation marks omitted). True, the state appeals court could have said more.6 But perhaps Lewis II’s cursory treat‐ ment of Cronic shows how obviously it does not apply. Cf. White v. Woodall, 572 U.S. 415, 427 (2014) (“[R]elief is available under § 2254(d)(1)’s unreasonable‐application clause if, and only if, it is so obvious that a clearly established rule applies to a given set of facts that there could be no ‘fairminded disa‐ greement’ on the question[.]” (quoting Harrington, 562 U.S. at 103) (emphasis added)). The opposite does not hold true: It is not “so obvious” that Cronic, and not Strickland, should apply here. White, 572 U.S. at 427. Under either clause, the majority opinion’s analysis of Lewis II comes closer to de novo review than the “doubly def‐ erential” standard mandated under AEDPA for ineffective as‐ sistance of counsel claims. Mirzayance, 556 U.S. at 123; cf. Harrington, 562 U.S. at 101 (“Here it is not apparent how the Court of Appeals’ analysis would have been any different without AEDPA.”). In Lewis II, the Court of Appeals of Indi‐ ana considered, and rejected, Lewis’s assertion of Cronic’s

6 Like the Indiana Court of Appeals, I also “note that Attorney Raff’s lack of advocacy at the sentencing hearing appears to have been, at least in part, invited by Lewis, who expressed clear disdain for counsel.” Lewis II, 116 N.E.3d at 1160 n.10, ¶ 43 n.10. 42 No. 20‐1642

second exception for lack for meaningful adversarial testing. 116 N.E.3d at 1159, ¶¶ 42–43 Ostensibly under AEDPA re‐ view, the majority opinion nevertheless crafts a hybrid rule— combining Cronic’s first and second exceptions—to cover Lewis’s claim. See Maj. Op. at p. 18. For the majority opinion, Cone supports this proposition: Lewis’s lawyer did less than the lawyer in Cone, which means Cronic should apply. See id. at pp. 18–21. But Cronic’s trigger is not so general, and Cone’s lesson is not so simple. If Lewis is to secure relief, he must fit within one of Cronic’s three exceptions, which Cone itself rec‐ ognized. Cone, 535 U.S. at 696. What is more, Cone is not the only benchmark by which to measure the merits of Lewis’s claim under Cronic. The major‐ ity opinion rejects Van Patten and Woods because “Lewis does not allege that his attorney was physically absent at any rele‐ vant time. Neither Van Patten nor Woods thus advances the analysis here.” Maj. Op. at p. 11. But if only constructive ab‐ sence cases were relevant to our analysis, then the majority opinion would have little, let alone recent, affirmative support. Herring v. New York, 422 U.S. 853, 864–65 (1975) (pre‐ suming prejudice when a state law barred summation of the evidence); Ferguson v. Georgia, 365 U.S. 570, 571, 596 (1961) (presuming prejudice when a state law barred elicitation of client’s trial testimony); see also Schmidt, 911 F.3d at 481. In‐ stead, Van Patten and Woods serve as further examples of the only thing that matters in this appeal: No Supreme Court case has held that silence at sentencing by defense counsel triggers Cronic. On that ground, Lewis’s petition should fail. IV. Conclusion Cronic’s scope is narrow, AEDPA review is narrow, and AEDPA review of a Cronic case is especially narrow. Bound No. 20‐1642 43

by AEDPA, I would reject Lewis’s habeas petition because no Supreme Court case has applied Cronic to the novel circum‐ stances presented by his claim. I respectfully dissent.