Dennis v. Secretary, Pennsylvania Department of Corrections

HARDIMAN, Circuit Judge,

dissenting,

joined by SMITH and FISHER, Circuit Judges.

At the outset of its analysis of James Dennis’s Brady claims, the Majority notes that the Antiterrorism and Effective Death Penalty Act (AEDPA) “dictates” our review. Majority Op. 280. The opinion describes with precision AEDPA’s strictures. Federal courts are prohibited from granting habeas corpus relief unless the state-court adjudication (1) “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,” or (2) “resulted in a decision that was based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding.” Id. (quoting 28 U.S.C. § 2254(d)). These fetters on our review, the Majority notes, come close to “imposing a complete bar on federal-court relitigation of claims already rejected in state-court proceedings.” Id. at 281 (quoting Harrington v. Richter, 562 U.S. 86, 102, 131 S.Ct. 770, 178 L.Ed.2d 624 (2011)).

It is one thing to recite these demanding limits; it is quite another to abide by them.1 And as Judge Fisher’s dissenting opinion cogently explains, they quickly fall *370by the wayside once the Majority turns to actually reviewing Dennis’s claims. I join Judge Fisher’s opinion in full, but write separately to note that I would reverse the District Court’s judgment even if there were no “analytical gap[s]” in the Pennsylvania Supreme Court’s decision rejecting Dennis’s Brady claims. Fisher Dissent 360-61. Consistent with the text of AED-PA and the precedents of the United States Supreme Court, I would hold that regardless of the thoroughness — or even the correctness — of the Pennsylvania Supreme Court’s stated reasoning, its judgment may not be upset so long as its decision did not contravene or unreasonably apply clearly established federal law and did not rest on an unreasonable determination of the facts. Whatever its flaws, the state court’s decision passes this test.

I

It is a virtue of our judicial system that courts explain their decisions in writing. When an explanation is not good enough— whether due to a legal, logical, factual, or other defect — the decision it supports is often reversed. AEDPA displaces this traditional approach to error review by imposing strict constraints on the writ of habeas corpus designed to stay the hand of federal courts over all but the most glaring of state-court errors. We may issue the writ only “where there is no possibility fairminded jurists could disagree that the state court’s decision conflicts with” the precedents of the Supreme Court. Harrington v. Richter, 562 U.S. 86, 102, 131 S.Ct. 770, 178 L.Ed.2d 624 (2011). For a prisoner in state custody to obtain habeas relief from a federal court, he must demonstrate that the state court’s decision on the claim presented before the federal court “was so lacking in justification that there was an error well understood and comprehended in existing law beyond any possibility for fairminded disagreement.” Id. at 103, 131 S.Ct. 770. “If this standard is *371difficult to meet,” the Supreme Court has explained, “that is because it was meant to be.” Id. at 102,131 S.Ct. 770.

A

By its terms, AEDPA applies to federal review of state-court decisions — not to the specific explanations that support them. See 28 U.S.C. § 2254(d). This distinction might seem technical, but the Supreme Court’s decision in Harrington v. Richter rendered it critical. There,' the Court was faced with the question of AEDPA’s application to a state-court decision that dismissed in a one-sentence summary order a habeas petitioner’s ineffective assistance of counsel claim. 562 U.S. at 96-97, 131 S.Ct. 770. The Court was presented with two issues: whether the state-court decision constituted an “adjudication] on the merits” under AEDPA, and if so, how the Court should go about determining whether the decision was unreasonable under AEDPA given that the opinion provided no reasoning. Id. at 97-102, 131 S.Ct. 770.

The Court’s answer to the first question rested on a straightforward application of AEDPA. Since the text of AEDPA “refers only to a ‘decision’” resulting from an “adjudication” — making no mention of the need for a “statement of reasons” — the Court held that summary decisions unaccompanied by an explanation usually qualify as merits adjudications under AEDPA. Id. at 98,131 S.Ct. 770. Hence, even where the state-court decision under federal review is devoid of reasoning, AEDPA’s deference requirements apply. It followed that “the habeas petitioner’s burden still must be met by showing there was no reasonable basis for the state court to deny relief.” Id. (emphasis added). This rule obtains regardless of “whether or not the state court reveals which of the elements in a multipart claim it found insufficient, for § 2254(d) applies when a ‘claim,’ not a component of one, has been adjudicated.” Id.

The Court’s answer to the second question in Richter — how to assess the reasonableness of a summary state-court decision under AEDPA — is particularly instructive here. The Court held that AEDPA requires federal courts to consider what explanations would nevertheless support the decision under federal law. As the Court explained, “a habeas court must determine what arguments or theories supported or, as here, could have supported, the state court’s decision; and then it must ask whether it is possible fairminded jurists could disagree that those arguments or theories are inconsistent with the holding in a prior decision of this Court.” Id. at 102, 131 S.Ct. 770 (emphasis added). At a minimum, then, when a state-court decision is unaccompanied by an explanation, Richter requires us to ascertain whether it was reasonable.

Circuit co'urts of appeals have divided over whether Richter extends beyond the precise circumstances of that case.2 Those courts that have chosen to cabin Richter can readily point to a limiting principle: single-sentence decisions versus multiple-sentence decisions. That distinction strikes me as unprincipled, however, because neither Richter’s logic nor AEDPA’s text limits the reason-supplying rule to cases in which the state-court “decision” is expressed in just one sentence. A decision is a decision, after all, and AEDPA does not distinguish among one-sentence decisions, one-paragraph decisions, or ten-page decisions; all of them are subject to the same deferential standard. Although the first *372portion of Richter focused on the fact that the state-court decision provided no explanation for the outcome, the reasonableness standard articulated in the rest of the opinion is tied to AEDPA’s general standard itself. “Where a state court’s decision is unaccompanied by an explanation, the habeas petitioner’s burden still must be met by showing there was no reasonable basis for the state court to deny relief.” Id. (emphasis added). In other words, regardless of how extensive or sparse the reasoning of a state-court opinion, the same AEDPA reasonableness test applies to all decisions on the merits.

This approach to AEDPA’s reasonableness standard finds support in Premo v. Moore. There, the petitioner claimed his counsel was ineffective for failing to move for suppression of the petitioner’s confession before advising him regarding a guilty plea. 562 U.S. 115, 119, 131 S.Ct. 733, 178 L.Ed.2d 649 (2011). The state court concluded that the petitioner had not established ineffective assistance of counsel under Strickland v. Washington, reasoning that a “motion to suppress would have been fruitless in light of the other admissible confession by [the petitioner], to which two witnesses could testify.” Id. at 119, 131 S.Ct. 733 (internal quotation marks omitted). Even though the state court “did not specify whether this was because there was no deficient performance under Strickland or because [the petitioner] suffered no Strickland prejudice, or both,” the Supreme Court stated that for a federal habeas court to properly eschew AED-PA deference, it “had to conclude that both findings would have involved an unreasonable application of clearly established federal law.” Id. at 131, 131 S.Ct. 733 (emphasis added).

Athough the state court’s reasoning was quite bare and did not explicitly engage the Strickland prongs, the Court held that its decision was entitled to AEDPA deference because reasons existed that would have supported the decision. Specifically, it highlighted that counsel had explained in state court that his decision to discuss plea bargaining before challenging the petitioner’s confession was based on his rationale that “suppression would serve little purpose in light of [the petitioner’s] other full and admissible confession.” Id. at 123-24, 131 S.Ct. 733. “The state court,” the Supreme Court explained, “would not have been unreasonable to accept this explanation.” Id. at 124, 131 S.Ct.'733 (emphasis added). Indeed, the Court found it unnecessary to consider a second justification that counsel had offered in the underlying proceedings because the first “confirms that his representation was adequate under Strickland, or at least that it would have been reasonable for the state court to reach that conclusion.” Id. (emphasis added). In short, presented with a state-court decision that was not a summary disposition but that provided only some vague reasoning for its decision, the Premo Court looked to the record to posit a rationale that would have supported that decision, finding it not to be an unreasonable application of federal law.3 We should approach Dennis’s case the same way.4

*373B

My understanding of Richter is supported by notions of consistency and coherence as well. If we were to limit Richter to cases involving one-sentence decisions, the outcome of federal review would turn on the state court’s opinion-writing technique. Consider a federal court faced with a state-court decision that rejected a petitioner’s claim that his conviction was invalid because it stemmed from an illegal arrest. Assume the record was unclear with respect to whether the arresting officer had probable cause, but that fairminded jurists could disagree as to whether a Supreme Court precedent demanded the conclusion that there was no probable cause. If the state court re*374jected the petitioner’s claim via summary disposition, Richter requires the reviewing federal court to infer the supportive rationale. Because the record would arguably support probable cause for the arrest, the conviction would be affirmed. But what if the very same claim had been rejected in a partially reasoned state-court opinion with problematic gaps in the logic from which adverse inferences could be drawn or in an opinion that gave incorrect reasons to justify the decision (say, by stating that the arrest was valid because there was “reasonable suspicion”)? Absurdly, appellate courts that circumscribe Richter in the way the Majority has here would require the reviewing federal court to ignore the supportive rationale on de novo review (where a weak case for probable cause wouldn’t be enough) and grant relief.

The asymmetry illustrated by my hypothetical makes a mess of the scheme established by AEDPA. How could a state-court decision be “reasonable” under AEDPA where the state court gives no reasons to explain itself but where we can think of one, yet be “unreasonable” under AEDPA where — although the very same good reason to support the decision exists — the decision is supported by undeveloped or incorrect reasons?5 See Mann, 774 F.3d at 1224-25 (Kozinski, J., concurring and dissenting) (“A habeas petitioner is not entitled to any reasoning at all, so reversing a state court’s reasonable decision on the grounds of incorrect reasoning risks treating defendants inconsistently: Those who are given incorrect reasoning get relief while those who aren’t given any reasoning do not.”). To make AEDPA reasonableness turn on a state court’s drafting decision is inconsistent with AEDPA’s directive that federal courts review the reasonableness of decisions, not opinions. And because it makes AEDPA deference inversely proportional to the amount of information the state court provides, it creates a perverse incentive for state courts to earn the deference of federal courts by saying less.6

II

To sum up, I would hold that when gaps or errors afflict a state court’s habeas adjudication, federal courts may not reverse unless the decision itself is unreasonable. In Dennis’s case, this principle is most pertinent to the Cason receipt. As Judge Fisher explains, the reasons proffered by the Pennsylvania Supreme Court for rejecting Dennis’s Brady claims regarding the Howard police activity report and the Frazier documents are themselves sufficient to pass AEDPA review without any inference from us. The Pennsylvania Supreme Court’s analysis of, the Cason receipt, on the other hand, is incomplete and might ungenerously be read as incorrect. For the reasons explained by Judge Fisher, however, a rationale consistent with Supreme Court precedent supports the decision, and so it must stand. I would simply add that AEDPA would require us to supply this rationale even if the state court’s treatment of the Cason receipt were in fact wrong. After all, “[a] state court could *375write that it rejected a defendant’s claim because Tarot cards dictated that result, but its decision might nonetheless be a sound one.” Brady, 711 F.3d at 827 (Wood, J.).

In my view, AEDPA requires federal courts to take the following approach to habeas review. Where the state court denies relief summarily, Richter requires federal courts to consider what arguments or theories could have supported the state court’s decision such that fairminded jurists could disagree whether those arguments or theories are inconsistent with the holding in a prior decision of the Supreme Court. Where the state court denies relief but addresses only certain prongs of a test or components of a claim, the reviewing federal court should likewise consider what reasons regarding an unaddressed prong or component could have supported the decision. And where, as here, the state court denies relief through vague, ambiguous, incomplete, or even incorrect reasoning, AEDPA still requires the reviewing federal court to consider what theories could have supported the decision, under AEDPA.7

By ignoring these principles, the Majority empowers itself to reweigh evidence that is decades old. Like the District Court, the Majority takes a fresh look at the evidence and concludes, contrary to the consistent testimony of three eyewitnesses, that the alleged Brady violations “effectively gutted the Commonwealth’s case against Dennis.” Majority Op. 269 (emphasis added). AEDPA proscribes such searching review. Because fairminded jurists could disagree as to whether the *376Pennsylvania Supreme Court’s decision was inconsistent with federal law, we owe it our deference. I respectfully dissent from the Majority’s decision to do otherwise.

. The inability of federal courts to follow AEDPA has reached epidemic proportions. As I pointed out in 2012, since 2000

the Supreme Court has granted certiorari in ninety-four cases arising under AEDPA, forty-six of which involved questions of federal court deference to decisions of state courts. Thirty-four of those cases (approximately seventy-four percent) have been reversed because the court of appeals failed to afford sufficient deference to the state court. Remarkably, twenty-two of those cases — almost fifty percent — were reversed without dissent.

Garrus v. Sec’y of Pennsylvania Dep't of Corr., 694 F.3d 394, 412-14 (3d Cir. 2012) (en banc) (Hardiman, J., dissenting) (collecting cases).

In the four short years since we decided Garrus, the errors have continued apace. By my count, of the nineteen cases arising under AEDPA in which the Supreme Court has granted certiorari, fourteen involved questions of federal court deference to state-court decisions. Thirteen of those cases were reversed — ten without dissent. See Kernan v. Hinojosa, -U.S. -, 136 S.Ct. 1603, 1606, 194 L.Ed.2d 701 (2016) (per curiam) (reversing the Ninth Circuit's treatment of a summary decision as a non-merits adjudication and noting that "the Ninth Circuit has already held that state-court denials of claims identical to [the petitioner's] are not contrary to clearly established federal law”); Woods v. Etherton,-U.S. -, 136 S.Ct. 1149, 1153, 194 L.Ed.2d 333 (2016) (unanimously reversing the Sixth Circuit because "a fairminded jurist — applying the deference due the state court under AEDPA — could certainly conclude that the court was not objectively unreasonable in deciding that appellate counsel was not incompetent under Strickland, when she determined that trial counsel was not incompetent under Strickland”); White v. Wheeler,-U.S.-, 136 S.Ct. 456, 461-62, 193 L.Ed.2d 384 (2015) (unanimously reversing the Sixth Circuit’s grant of habeas relief because it "did not properly apply the deference it was required to accord the state-court ruling”); Davis v. Ayala, - U.S. -, 135 S.Ct. 2187, 2208, 192 L.Ed.2d 323 (2015) (reversing the Ninth Circuit’s grant of the writ on the ground that fairminded jurists could disagree as to whether a state court’s exclusion of a defendant's attorney from part of a Batson hearing was harmless error); Woods v. Donald,-U.S.-, 135 S.Ct. 1372, 1377, 191 L.Ed.2d 464 (2015) (unanimously reversing the Sixth Circuit’s grant of habeas relief because the state court's conclusion that the petitioner's counsel was not per se ineffective "was not contrary to any clearly established holding” of the Court); Glebe v. Frost, - U.S. -, 135 S.Ct. 429, 430, 190 L.Ed.2d 317 (2014) (unanimously reversing the Ninth Circuit's conclusion that the state court "unreasonably applied clearly established federal law by failing to classify the trial court’s re*370striction of closing argument as structural error" because no Supreme Court precedent clearly established that such mistakes rank as structural error); Lopez v. Smith, - U.S. -, 135 S.Ct. 1, 5, 190 L.Ed.2d 1 (2014) (unanimously reversing the Ninth Circuit where it “had no basis to reject the state court’s assessment that [the petitioner] was adequately apprised of the possibility of conviction on an aiding-and-abetting theory”); White v. Woodall, - U.S. -, 134 S.Ct. 1697, 1702-04, 188 L.Ed.2d 698 (2014) (reversing the Sixth Circuit’s grant of habeas relief because the state court’s determination that the trial court’s jury instructions did not violate clearly established federal law was not "objectively unreasonable”); Burt v. Titlow, - U.S. -, 134 S.Ct. 10, 17-18, 187 L.Ed.2d 348 (2013) (reversing without dissent the Sixth Circuit’s judgment that the state court's conclusion that counsel’s performance was ineffective was unreasonable); Nevada v. Jackson,-U.S.-, 133 S.Ct. 1990, 1994, 186 L.Ed.2d 62 (2013) (per curiam) (unanimously reversing the Ninth Circuit's grant of habeas relief where the state court reasonably applied federal law in determining that the petitioner had not been denied the right to present a complete defense when he was not allowed to present certain, extrinsic evidence); Metrish v. Lancaster, — U.S.-, 133 S.Ct. 1781, 1792, 185 L.Ed.2d 988 (2013) (unanimously reversing the Sixth Circuit’s grant of the writ where Supreme Court had "never found a due process violation in circumstances remotely resembling [the petitioner’s] case”); Johnson v. Williams, - U.S. -, 133 S.Ct. 1088, 1097, 185 L.Ed.2d 105 (2013) (reversing without dissent the Ninth Circuit’s grant of relief based on the faulty conclusion that the state court had overlooked a meritorious Sixth Amendment claim); Ryan v. Gonzales, -U.S. -, 133 S.Ct. 696, 700, 184 L.Ed.2d 528 (2013) (killing two birds with one stone in unanimously reversing both the Sixth Circuit’s and Ninth Circuit's grants of relief where the courts wrongly concluded that federal law provides a right to incompetent prisoners to suspend their federal habeas proceedings); but see Brumfield v. Cain, — U.S.-, 135 S.Ct. 2269, 2281, 192 L.Ed.2d 356 (2015) (finding the state court's determination of the facts regarding a defendant with an IQ of 75 unreasonable).

. See Noam Biale, Beyond A Reasonable Disagreement: Judging Habeas Corpüs, 83 U. Cin. L. Rev. 1337, 1391 (2015) ("Since Richter ... the circuits have split on whether the opinion’s ‘could have supported’ language for decisions unaccompanied by a reasoned opinion applies to decisions that do include a reasoned opinion.”).

. The Majority and Judge Jordan conclude that the Supreme Court's decision in Lafler v. Cooper “implies a limit” to the reason-supplying rule announced in Richter. Jordan Concurrence 351. I do not read Lafler that way. Significantly, habeas relief in that case rested on the Supreme Court's holding that Michigan Court of Appeals’ application of Strickland was "contrary to" — not an “unreasonable application of” — clearly established federal law. - U.S. --, 132 S.Ct 1376, 1390, 182 L.Ed.2d 398 (2012). Specifically, rather than applying the Strickland ineffective-assistance-of-counsel standard, the state court applied a (completely wrong) "knowing and voluntary” plea rejection rule. Id. Because a decision is categorically “contrary to” clearly established federal law if the state court “applies a rule that contradicts the governing law,” Williams v. Taylor, 529 U.S. 362, 405, 120 S.Ct. 1495, 146 L.Ed.2d 389 *373(2000), AEDPA deference was inappropriate, so de novo review applied. Lafler, 132 S.Ct. at 1390. Consequently, the case was not amenable to Richter's "could have supported” analysis to determine whether the state court decision was an unreasonable application of federal law. For these reasons, I disagree with the Majority and Judge Jordan that La-fler instructs federal courts to "take the state court's decision as written” and apply Richter only in the small subset of cases in which the state court left so-called "gaps” to be filled. Jordan Concurrence 351.

Nor do I read Wetzel v. Lambert to imply any limit on Richter. Although the opinion in that case did not include Richter’s "theories [that] ... could have supported” language in its recitation of AEDPA's general standard, see-U.S. -, 132 S.Ct. 1195, 1198, 182 L.Ed.2d 35 (2012), the Court did not reject that approach by implication. Rather, in Wet-zel the reasons for upholding the state court’s decision under AEDPA were expressed by the state court. The petitioner claimed the prosecution violated Brady by suppressing a police activity sheet consisting of a photo display marked with written notations suggesting that "someone other than or in addition to” the petitioner had committed the crime. Id. at 1196-97. We granted habeas relief, but the Supreme Court vacated and remanded, explaining that we had "overlooked the determination of the state courts that the notations were ... 'not exculpatory or impeaching' but instead 'entirely ambiguous.”' Id. at 1198. The Court criticized us for "focus[ing] solely on the [state court’s] alternative ground that any impeachment value that might have been obtained from the notations would have been cumulative.” Id. The problem was that "[i]f the conclusion in the state courts about the content of the document was reasonable — not necessarily correct, but reasonable — whatever those courts had to say about cumulative impeachment evidence would be beside the point.” Id. Hence, by failing to recognize — as the state courts did' — the " 'ambiguous' nature of the notations” and the “ speculative]’ nature of [the petitioner's] reading of them,” we ran afoul of AEDPA. Id. Far from implying a limitation on Richter, Wetzel merely requires federal habeas courts to review state court opinions in search of a reasonable reading that would support the decision under federal law.

. Some courts have begun to recognize Richter s true reach. See, e.g, Holland v. Rivard, 800 F.3d 224, 235_(6th Cir. 2015) (concluding that although "a state court decision unaccompanied by any explanation differs from a state court decision based on erroneous reasoning ... Richter suggests that this is not a meaningful distinction” and that AEDPA requires a habeas petitioner to show that there was "no reasonable basis for the state court to deny relief ... whether or not the state court reveals [its reasoning]”); Trottie v. Stephens, 720 F.3d 231, 240-41 (5th Cir. 2013) (“We review only the ultimate legal determination by the state court — not every link in its reasoning.”); Brady v. Pfister, 711 F.3d 818, 827 (7th Cir. 2013) (Wood, J.) ("[I]t is clear that a bad reason does not necessarily mean that the ultimate result was an unreasonable application of established doctrine.... If a state court's rationale does not pass muster ... for Section 2254(d)(1) cases, the only consequence is that further inquiry is necessary.”); Mann v. Ryan, 774 F.3d 1203, 1224-25 (9th Cir. 2014) (Kozinski, J., concurring and dissenting) ("I have misgivings about whether, in light .of the Supreme Court's decision in Richter, we are still entitled to reverse a state court’s reasonable decision based on what we consider to be its incorrect reasoning.... After Richter, it seems clear that we should *374assess the reasonableness of a state court’s decision, not its reasoning.”).

. Such arbitrariness is all the more perplexing in light of the fact that AEDPA "does not require citation of [Supreme Court] cases— indeed, it does not even require awareness of [Supreme Court] cases.” Early v. Packer, 537 U.S. 3, 8, 123 S.Ct. 362, 154 L.Ed.2d 263 (2002).

. See Hodges v. Colson, 727 F.3d 517, 537 n.5 (6th Cir. 2013) ("[If Richter is limited to summary dispositions], the more information the state court provides, the less deference we grant it. This is contrary not only to the language of the statute, which speaks of 'claims' not components of claims, but also contrary to the spirit of § 2254(d), which is designed to give more deference to a state court judgment on the merits.”).

. I disagree with Judge Jordan that my understanding of Richter conflicts with Ylst v. Nunnemaker and Wiggins v. Smith. Both of those cases involved the threshold question of whether the petitioners' claim had been decided on the merits. The Ylst Court was faced with an "unexplained” State supreme court order denying the petitioner's habeas petition, wherein it was unclear whether the court rested its denial on a procedural default (the basis of the lower court's holding) or on the merits of his Miranda claim. 501 U.S. 797, 801, 111 S.Ct. 2590, 115 L.Ed.2d 706 (1991). The Supreme Court reversed the Ninth Circuit's conclusion that the decision was on the merits, explaining that, "where, as here, the last reasoned opinion on the claim explicitly imposes a procedural default, we will presume that a later decision rejecting the claim did not silently disregard that bar and consider the merits.” Id. at 803, 111 S.Ct. 2590. To the extent that Ylst requires us to "look through” unreasoned state court opinions to the last reasoned opinion, I have no quarrel with Judge Jordan that we ought to first consider whether the state court's stated explanation is reasonable before deigning to supply reasons of our own under Richter. As for Wiggins, we have explained that the reason the Court declined to apply deference with respect to the prejudice prong of the petitioner's Strickland claim was that the state courts had not decided the Strickland prejudice issue "on the merits.” Palmer v. Hendricks, 592 F.3d 386, 400 (3d Cir. 2010); see also Wiggins, 539 U.S. 510, 534, 123 S.Ct. 2527, 156 L.Ed.2d 471 (2003) (“[Ojur review is not circumscribed by a state court conclusion with respect to prejudice, as neither of the state courts below reached this prong of the Strickland analysis.”). Because AEDPA deference only extends to "any claim that was adjudicated on the merits in State court proceedings,” 28 U.S.C.A. § 2254(d), the determination whether the state-court decision under federal review was made on the merits is prior to the consideration, vel non, of whether adequate reasons exist in support of that decision. I do nevertheless agree with Judge Jordan that Wiggins is in some tension with my approach because it engaged in de novo review of the second prong of Strickland even though the state court denied relief but addressed only the first prong. However, Richter — decided after Wiggins — speaks clearly on this point. "[Aj^habeas petitioner's burden still must be met by showing there was no reasonable basis for the state court to deny relief.... whether or not the state court reveals which of the elements in a multipart claim it found insufficient, for § 2254(d) applies when a 'claim,' not a component of one, has been adjudicated.” Richter, 562 U.S. at 98, 131 S.Ct. 770.