Oral argument — Brown v. Davenport

We'll hear argument first this morning in case 2826, Brown v. Davenport. Ms. Hamoud? Mr. Chief Justice, and may it please the Court, Davenport's concession that Brecht doesn't always subsume ACPA narrows the dispute here. But the modified Brecht-only approach he suggests gives no deference to state courts' merits adjudications and absolves habeas petitioners of their burden under 2254 D.I. Even if federal judges relied only on material permissible under ACPA within its Brecht analysis, the inquiry is not over. It is not enough for federal judges to believe in their own minds that an error substantially influenced the verdict. Before granting relief, they must look through ACPA's highly deferential lens and ask whether all other fair-minded jurists would disagree with the state court's conclusion. When Congress enacted ACPA, it did not give federal judges the option of ignoring this crucial deference. That is why, as a precondition to habeas relief, they must apply both Brecht and ACPA. Failing to do so contravenes this Court's modern habeas jurisprudence, including IALA, which reaffirmed that ACPA's limitations are distinct from Brecht. The Sixth Circuit's Brecht-only approach failed to defer to the Michigan courts. It also extended this Court's holdings, relied on circuit precedent, conducted an independent review of the record, and used extrajudicial social science studies, all of which are prohibited under ACPA. As Judge Thapar said in his en banc dissent, federal judges can't simply ignore ACPA's guardrails whenever they find actual prejudice under Brecht. We ask this Court to articulate the correct standard and to reverse the Sixth Circuit. If you were writing on a clean slate, how would you coordinate Brecht and ACPA? Would you say that, for example, that one subsumes the other? In a case of denial, in a case of denial of relief, applying the other would be a mere formality. As far as this Court found at the state court's conclusion So it really wouldn't matter if you denied? If you denied applying the other, formally applying it would not matter because it would be a mere formality. However, if a court were to grant relief under either, it must go to the next test. So if they were to grant relief under Brecht, as the Sixth Circuit did, it must apply ACPA as a precondition to the grant of relief. And if a state court used the wrong standard or it was contrary to this Court's precedent, then if a Petitioner prevails under ACPA, Brecht must be applied as well prior to relief, Your Honor. I hope that answers your question. Thank you. Why — that's how you think it would be applied. But why would a rational legislature set the system up this way? In other words, okay, let's have this inquiry under Brecht. Then let's have this separate inquiry under — under ACPA. Would somebody just sitting down on a clean slate put that system together? Well, one, we know that they are different tests. They're distinct tests. They ask different questions. And when Congress enacted 2254-D1, that was three years after Brecht. So Brecht could never consider the limitations that ACPA set in place. And Brecht applies whether or not there's a state court determination. So they're not two of the same. Each hold different burdens as well. Well, if that's really — if they sat down and decided that's what we're going to do, don't you think they would have made it a little clearer than to have us sitting here now and saying, well, how do we reconcile these two things? Because, you know, they're addressed to the same question, I guess, at a broad level. In other words, it would seem to me odd that they would leave it implicit that AEDPA and Brecht would coexist. They have to coexist because when 2254 specifically applies to state court's merits determination, Brecht doesn't need a state court's merits adjudication for it to apply. And we know that this Court said in Frye that on collateral review, whether there's a state court's merits adjudication or not, Brecht applies on collateral review. Now, once there is a state court's merits adjudication, that was the heart of AEDPA, was to protect that. And that's the basic structure. Now that there is a state court merits adjudication, then that needs to be protected. And it can't be ignored. And it doesn't offer a menu of options. We must give it deference. And so they do ask different questions. And as Justice Tappar said, different questions often lead to different answers. Ms. Samu, do we have to overrule IALA to side with you? And if not, how do we handle that subsumes language in IALA? The Court does not have to overrule IALA because the question that's presented here was never asked in IALA. And, in fact, in IALA, the Court applied both. And the Court made clear that AEDPA is a precondition to relief. And in IALA, the Court did not grant relief. So in terms of the subsumes language, I think that the Court can clarify that if a Federal court were to grant relief under Brecht, we ask this Court to do exactly and say what it said, reiterate AEDPA's limitations, that AEDPA remains a precondition to relief. What is your understanding of the meaning of the term subsume? Your Honor, we know that it can't mean ignore or make dull. However, our reading of it is it could subsume, which means a court does not have to formally apply AEDPA if the Petitioner was not entitled to relief under Brecht, doesn't have to go through a separate application because there is no grant of relief in that case. It would kind of subsume that conclusion, that decision. Well, if I look up the definition in the dictionary, well, I find something like this, include as a component. Is that a meaning of the term subsume? And I know that the definition, Your Honor, has been debated with what does it mean, right? Judge Radler said it can't mean consume. Certainly, the Sixth Circuit thinks that it means you can completely leapfrog. That AEDPA would all – that AEDPA would be – all of AEDPA's limitations would be included in Brecht and that a Federal court could leapfrog and ignore AEDPA. We know that at least it can't mean that, which is why I think this case – Well, it means include as a component. And so if 2254d is included as a component of Brecht, then doesn't that mean that a court purporting to apply Brecht still has to satisfy 2254d? They're not two of the same. And, you know, when we look at the two tests differently, we know that Brecht doesn't ask the questions that AEDPA asks. One, they're distinct. One is an independent review as to what, as this Court said in O'Neill, me as a Federal court judge believe in my own mind. And as opposed to AEDPA, they have to look and ask the question, is there a fair-minded disagreement on this? Well, sometimes judicial opinions can confuse things. So maybe it's helpful to go back to first principles. Isn't Federal habeas relief entirely statutory except in those circumstances And yes, but this Court's – Yes. Okay. The answer to that is yes. 2254d is a statute. On what basis could a Federal court say, we're not going to follow 2254d, we're going to follow a judicially created standard in Brecht? What do you understand to have been the basis for Brecht? It wasn't in the Federal habeas statute at that time, was it? No, Your Honor. In fact, this Court, one, they must – they must both apply the Federally mandated congressional statute. It was an – it was our understanding, it was our application of the equity that a Federal court exercises when it provides Federal habeas relief. It was an equitable rule that was read into the previous statute, the previous version of the statute. And so if there were a conflict between that and a subsequently enacted statute, which would prevail? They must both prevail because Brecht does not need a State court's merits adjudication in order to apply. And two, both of them can exist at the same time. Especially on collateral review. But I do think, Ms. Hammoud, that the language in Davis v. Ayala, and also in Frey, which Davis v. Ayala quotes and refers to as a holding, that that language goes, you know, something like this. It says, we've looked at these two tests, and what we think is that the stricter standard is the – the – Brecht. Brecht standard. And so if a court does Brecht, that's good enough for us. If a court does Brecht only, that's good enough for us. Now, you might contest that. You might say, well, that was too hasty to just say that Brecht is stricter in all circumstances. But, in fact, that's what the Court twice said, said in no uncertain terms, that the one subsumes the other because the Brecht test is stricter than the Chapman test. So if a court does the Brecht test, it's sufficient. That's the way I view, I read, and I think it's the only way to read both of these decisions. Now, I understand the point that they were wrong in saying that. I mean, I understand the argument you're making, but they say what they say, don't they? Justice Kagan, that is correct. That language was included in Frey. And, as Your Honor stated, Brecht and Chapman were compared in Frey. And, in fact, in Frey, there was no harmless error determination by a State subject to deference in Frey with no AEDPA overlay. What our position is, is that when there is an AEDPA overlay that's distinct from Frey, sure, when one is comparing Brecht and Chapman, you can compare into which one is friendlier to a criminal defendant. However, the AEDPA overlay asks different questions. And that is not what that specific judge thinks, but whether there is fair-minded disagreement. Kagan. I hear you on that. I hear you. But what I'm suggesting is that that's an argument that could have been made to the Davis v. Ayala court. It's an argument that could have been made to the Frey court. But that the language in both of those cases essentially rejects that argument. It basically says, look, we think that the Brecht standard is, you know, it's just going to do all the work here, so we think that the Brecht standard is enough. This Court in Frey did not consider this question, and this Court in Ayala did not consider this question. In fact, this Court in Ayala specifically stated that Frey did not abrogate the AEDPA. And this Court has repeatedly stated that the two tests are distinct not only with different burdens as well in terms of who carries the burden under each test. We know from this Court that the State court's ruling has to have been so lacking in justification that there was an error well understood and comprehended in existing law beyond any possibility for fair-minded disagreement. Counsel. Those are different tests. Yes, Your Honor. All of those things you're saying, those language, fair-minded disagreement, nobody else can think of this this way. Let's have the language of the statute. The statute just says, and I'm reading 2254d, CABIA shall not be granted with respect to any claim that was adjudicated on the merits in a State, I'm sorry, resulted in a decision that was contrary to, and this is the operative language, or involved in unreasonable application of clearly established Federal law. Yes. Explain to me in layman's terms when a court under Brey, under Chapman, under any test that you want to set forth, basically says the constitutional violation here had to have substantially injured, caused substantial and injurious effect on a verdict, aren't they saying by definition that whatever interpretation you give, it can't be reasonable? Isn't that what Davis meant by subsumes? Isn't that what Frey meant by the same concept? How can it ever be reasonable to conclude that there was no injury to a verdict when a judge finds there was? Thank you, Justice Sotomayor. I'm going to address both of your questions. First, I want to go back to the statute. And the reading of the statute is, it shall not be granted, that's a command, unless the State court's adjudication in layman's terms. That's exactly what we said in Davis. In Davis, when we talked about the Breck standard, that Aiella had to meet the Breck standard, and that while a Federal habeas need not formally apply both Breck and Havis, EPA nevertheless sets forth a precondition to the grant of habeas. And that's exactly what we want this Court to say. Well, that's exactly what the Court said. You don't have to apply both. Yes. The Court, in its analysis in Aiella, want to say that there's no basis for finding that Aiella suffered actual prejudice, and there was no causal statement between the two. The decision of the California Supreme Court represented an entirely reasonable application of controlling precedent. What the Sixth Circuit didn't do, according to the statute that Your Honor just cited, 2254d1, at no point did they consider whether or not it was an unreasonable application that the State court did. All you're asking us to do in this case today is to tell the courts below, apply both Breck and Chapman EPA? Is that all you're asking us to do today? We're asking that the Court articulate that prior to the grant of relief, they must apply both Breck and EPA. And we believe that the Court — All right. So you'd be happy if that's all we said here? I'm sorry? That's all you're asking us to do, to remand it and say, apply both, don't rely on circuit precedent, and don't use social science data? Is that what you're asking us to do? By applying both, that's already included, because they're different questions and they consider different actions. What's happening here is a statement in IL — Just answer my question. What do you want our judgment line to say? Exactly what Your Honor had stated, is that prior to the grant of relief, they must apply both. And we believe it would be prudent and it would offer the State — it would offer the bar and bench guidance if this Court were to go and articulate the difference between the two standards, and exactly why the Sixth Circuit failed to give deference to State courts and to abide by a congressionally mandated statute. And we think that the best way to answer that is for the Court to even go as far as applying it. But at the end, in our question, we do ask that this Court articulate the correct standard. Well, you've told us that the Sixth Circuit didn't do that, right? It did not do that. And is it our common practice? Isn't it against our common practice to do something in the first instance? Don't we lay out standards and let the Court below apply them? That's correct, Your Honor. And Your Honor, we're asking this Court to, one, articulate exactly why the Sixth Circuit did not do that, and two, the district court in this case, prior to it going to the Sixth Circuit, actually, their decision was that the State courts, they applied AEDPA, and that the State courts' merits adjudication was not objectively unreasonable. They asked those questions prior to it going to the Sixth Circuit. I thought your brief ended by asking that, say, the Court should reverse the Sixth Circuit's judgment, not remand it. Yes. Reverse it on that jurisprudentially significant question, Your Honor. And, Ms. Simut, can I ask you, what was the last adjudication on the merits? Why shouldn't we? I mean, it seemed to me that below, in saying that the Michigan Supreme Court's probably was, you basically conceded that it was, but now you're saying that after your brief says that after considered reflection, you think it was the Court of Appeals. Why shouldn't we hold you to your earlier concession? Your Honor, when we filed our briefing in district court, and the district courts adjudication and review of our case, their decision was that the Michigan Court of Appeals was the last reason decision based on their analysis. And since then, we have carried that position, but we've always said, whichever one, they still deserve deference. And in our briefing, we did state that after reconsideration of that legal question, it's the Michigan Court of Appeals decision that's the last reason decision, because that's what deference is. You have to take them at their word. They made a decision not to decide the case, and they denied leave to appeal. But they offered, I mean, it's unusual in that they didn't take the case, but they also had a bit of an opinion. I mean, they offered some views about the merits. Yes, and they offered that at the end of their denial. And our courts often sometimes, and sometimes they don't, and sometimes they do offer guidance to lower courts. And in Michigan, that's consistent not just with our state laws, but with this court's jurisprudence as to the fact that even if they offer guidance at the end, that's not considered to be the last merits adjudication on the case. All right. My difficulty with this case is I believe that you understand it, and I believe that the lawyers in front of me understand it. And my colleagues spent time on it. So did I. And I have a terrible time understanding where all these different standards are and how they fit together. And I doubt that a lot of habeas judges will understand it either. Maybe they will, but many will not, no matter what we say. So I began to think of how could we deal with this. The problem comes up because Breck says if you're the habeas person, you want habeas, I have to find me the habeas judge. That's correct. You have proved that it was harmful. Right? That's correct, Your Honor. No problem. But the State court, the DA there, what she did was prove beyond a reasonable doubt that it wasn't harmless. Oh, wait a minute. Prove beyond a reasonable doubt. Prove that it was harmless beyond a reasonable doubt. And so then you get where Justice Kagan was and you say, but that's just contradictory. Ah, not quite. Because we've said that when you look at that State court decision, as long as a reasonable jurist could have found that it was harmless beyond a reasonable doubt, it has to stand up. What's my problem? I think it was harmful. But I can't bring myself to say that no reasonable jurist could have agreed with that person over in the State court. And so you say that's what we should apply. Now, have an idea. The purpose of this whole thing is to get the habeas judge to pay some attention to what they did on this in the District court, in the Federal — in the State court. That's the purpose, isn't it? Pay attention, Federal habeas judge, to the fact that those are good judges over there, too, and they came out the opposite. You say it was harmful, but they said no reasonable — beyond a reasonable doubt it was harmless. Pay attention to it. So why don't we just say that? Why don't we just say this is one of the questions, or if you ever have such a situation, Federal habeas judge, please pay some attention. And instead of writing it in a legal standard that no one can understand, just tell them what to do. Pay some attention. Now, I grant you, that would leave it all up to them. It would be very hard to review. But we leave lots of things to district judges. And there we get our objective. Pay attention to the fact that the State court did come out the opposite on this than you did. What about that? This Court has already done that. This Court in Richter has specifically stated, Federal judges can't use this test as a test of its own confidence in the result they would reach in a de novo review. That they cannot grant petitions because merely they disagree with the State court's harmlessness determination. In this case, not only did you have 11 Michigan judges that believed it was harmless. All right. All right. Stop right there. Just say, okay, we said it already. We'll just repeat that. And we'll say, whatever the technicalities here of this language, Chapman, Brecht, which nobody really is going to understand with these two hours of study, just do what we said there. Pay some attention to the State court, the fact that they found the opposite. As Judge Sutton had put it. Would you be happy with that? No. And I wish it worked, Your Honor. But this Court has said that repeatedly. Two and a half decades after Congress enacted ACFA, this Court have said that. And as Judge Sutton put it, there is vexing language. And he stated in his concurring opinion in the en banc denial, I suspect every Federal judge in the nation would benefit from articulating the standard and clarifying this language. And that is why we believe that it's important that this Court articulate, because this question has not been asked, and it's not been in front of this Court before, that if a petitioner were to prevail under Brecht, can a State court, can a Federal court leapfrog ACFA and consider it? No. I mean, I've been trying to figure out how this question matters. And I'm going to have some questions for Ms. Pahal on this point, too. Because, frankly, I'm not sure that it matters all that much. But let me put this to you, which is, has there ever been a case where a court granted relief under Brecht and then said, sorry, we can't grant relief because of EDPA Chapman? Your Honor, this case is the perfect case to do that. And this case — No, but has there ever been a case where any judge ever said that? I mean, I think the reason Ayala and Frey look the way they do is essentially that the Court made a judgment that they could not imagine a court saying that. And, in fact, as far as I can see, no court has ever said that. This is the first time that a court actually grants relief without applying EDPA Chapman. And the reason why we don't have more of those decisions is because circuit courts have been applying EDPA Chapman, and the Sixth Circuit did not conform to that. I see that my time is up, and I would like to reserve the rest of my time for rebuttal. You'll have rebuttal. Justice Thomas? Justice Breyer? Justice Kavanaugh? No further questions, Chief. Thank you. Thank you very much, Counsel. Thank you, Mr. Chief Justice. We'll hear now from you, Ms. Bahal. Mr. Chief Justice, and may it please the Court, Brecht and EDPA Chapman are both preconditions to habeas relief, and both standards have been met here. Mr. Davenport was actually prejudiced by the unconstitutional shackling, as the Court of Appeals found under Brecht. The State has not sought review of that Brecht determination before this Court. The finding of actual prejudice necessarily means the State court adjudication on the merits was an unreasonable application of the Chapman standard. There is a clear and logical relationship between Brecht and EDPA Chapman, with Brecht setting the higher hurdle. Chapman requires the State to prove on direct review the error was harmless beyond a reasonable doubt. That means the State must show there was no reasonable possibility the error contributed to the verdict. EDPA then asks whether a fair-minded jurist could agree with that Chapman determination. Brecht in turn asks whether there is more than a reasonable possibility the error contributed to the verdict. Comparing the standards, where there is more than a reasonable possibility the error contributed to the verdict, as is the case here, no fair-minded jurist could agree there is no reasonable possibility the error contributed to the verdict. Put differently, a fair-minded jurist confronted with more than a reasonable possibility of harm could not find the error harmless beyond a reasonable doubt. That relationship between the standards was recognized by this Court in Frey and again in Ayala, and it's also been recognized in the practical experience of Federal courts applying these standards for more than 20 years. Through multiple rounds of briefing, the State has never identified a single case in which Brecht was satisfied but EDPA Chapman was not satisfied. Therefore, where a finding of actual prejudice under Brecht has been made, and that finding does not rest on sources of law that would not be permissible to consider under 2254d, the Brecht inquiry answers the EDPA questions. I'd now be happy to take the Court's questions. Counsel, would you comment or respond to Justice Alito's point as to the stature or status of Brecht as an equitable doctrine in comparison with EDPA, which is statutory? Does one have preference over the other, the statutory over the equitable, or are they both to be treated given the same weight? Our position in this case, Your Honor, is that both Brecht and EDPA Chapman are both preconditions to habeas relief and that both have been satisfied in this case. Well, I understand that, but if you had to choose between one or the other, which has the higher status? The Brecht question asks a question that requires a more difficult hurdle for a defendant to satisfy, but I believe that they are both equally important in the granting of habeas relief. Well, Brecht is an opinion, a decision from this Court, and as I said, it's equitable. EDPA is statutory, and you don't think there's any difference as far as which has the higher stature and which one should command more of our attention? I think they both must be satisfied before habeas relief should be granted, as they were in this case. Well, if you don't think that they can be, if you don't think they are compatible, let's assume, just for the sake of discussion, that someone thinks they're incompatible. Which takes precedent? I'm not sure I know how to answer the question, as we're not conceding that one test is more important or less important than the other. We think they both must be satisfied, as they were here. An act of Congress is important, as is this Court's precedent. I mean, if that's true, Ms. Bahal, that both have to be satisfied, then why not just tell courts that both have to be satisfied? You know, it seems like kind of a waste of pages. But, you know, just go through the motions, do it twice. And I understand why you don't want that, because that's not the way the Sixth Circuit decision reads. So it's unfair, perhaps, to ask you to answer this question, because, you know, your client has a real interest in keeping this judgment. But, I mean, I guess my question here is, if one, you know, generally subsumes the other, but maybe contra Ayala and contra Frye, we could imagine a case in which that wasn't true, just have the courts go through both, and we'll be sure. Courts can do formal application of both. That's up to the courts. The question here is whether or not to. Yeah, I mean, the question is whether to require it, right? And so why not just say, you know, you have to do it just so we're sure that no errors are taking place, and that AEDPA is being considered in the right way? To require parties and courts to go through the time, effort, energy of briefing, arguing two separate questions, the 2254D question first, when it's answered, and then require that whole round of time, energy, effort to then answer the Brecht question will answer the AEDPA-Chapman inquiry seems unnecessary. Courts can do it. That's fine. But here the question is whether it's error not to do it. And this — Breyer. You can make up cases where it could really lead to a different result. The habeas judge sits there and says, Smith, the juror saw the shackle. I'm sure he saw the shackle. And so it's not harmless. It's harmful. And then he says, of course, the court of appeals over there in the State, what they said is that Smith didn't see the shackle because he was looking out the window. And I don't believe that, but I think a reasonable juror could have believed it. See? Now we've got different results from the two tests. And so they're saying, well, that could have happened. And then you say, well, it never happened. That's hardly surprising because nobody could understand the test. But regardless, the good happened. So what are we supposed to do? And that's sort of where I'm stuck. I can imagine cases where it happens. And they seem far and few and far between, but I can imagine it. And so what are we supposed to do? Well, in the hypothetical you just gave, Justice Breyer, that question goes to whether there was an underlying constitutional violation in the first place, if someone sees the shackles or not. Here the record is undisputed. There was. There was. There was. Well, let's make it just make a different thing. I mean, see, make a different thing. Was this witness believable? Maybe a judge says, yeah, I think he's absolutely believable, and therefore this omission here of the witness was really harmful. You know, the other one says, no, it wasn't. It wasn't believable at all. First judge. I think I agree with that second judge. No, no, I don't. But I could see a reasonable juror might. Now, that's being a little too honest. But you see the problem. The context that we're advocating here and our approach here is limited to the context where there is an underlying constitutional violation. The weighing the credibility might not fall into that category. And so the relationship between the standards, as I described them, is limited to where Chapman is the underlying clearly established law. I'm sorry. No. Please finish. Because there is an underlying constitutional violation where Chapman applies, Brecht would subsume the Edpa-Chapman inquiry. Well, Brecht calls on the federal habeas judge to make a personal judgment. The federal habeas judge could say, I personally have a grave doubt. I personally think that this had a substantial effect on the outcome. But Edpa looks at something different. Couldn't a judge say, I personally think this had a substantial effect, but a fair-minded jurist could reasonably reach the opposite conclusion? They're looking at two different things, aren't they? The standards are an apples-to-apples comparison because they're all looking at whether the constitutional trial error affected the verdict, and they're setting different hurdles for that, with Brecht being the higher hurdle. You can't surpass the Brecht hurdle without also satisfying the Edpa-Chapman hurdle. Well, why is that so? Isn't what I just said possible? A judge could say, I personally think that it had a substantial effect, but — and I have no grave doubt about that. On the other hand, a reasonable jurist could reach the opposite conclusion. Is that irrational? Is that inconsistent? It would be like a prosecutor standing up at closing argument and saying, there is more than a reasonable possibility that this defendant is innocent, but I, the State, still prove Tim guilty beyond a reasonable doubt. Well, no, it's not at all the same. The Brecht standard, because it subsumes the Edpa inquiry, you cannot have a finding of grave doubt on one hand with a fair-minded jurist concluding on the other that the harm was harmless beyond a reasonable doubt. Well, maybe our opinions have confused things by introducing this concept of one subsuming the other. Why shouldn't we just get rid of that? Edpa is a statute. It says in unequivocal terms, you can't grant Federal habeas relief unless the decision is based on an unreasonable application of Federal law defined in a certain way, period. There's no way that Federal relief, Federal habeas relief can be granted unless that is satisfied. So forget about what subsumes something subsuming the other. Brecht was an equitable decision. It continues to have force in a situation in which there isn't an applicable Edpa provision, which is what Frey addressed when there wasn't a harmlessness determination by the State court. Doesn't that simplify things? And is there anything wrong with it? If you disagree with the logical relationship as I laid out and require formal application of both tests, application, formal application of Edpa here confirms the result. The Michigan Supreme Court opinion was contrary to clearly established law. The law that was to be applied was Chapman, which requires the State to prove the error harmless beyond a reasonable doubt. This is not — Edpa was a sea change in habeas law. And this is why it's — and this is the argument your friend makes, different from Brecht. It said, you've made your determination under Brecht, and that's fine. We don't care whether there's one judge who disagrees with the State court. We want to make sure that that determination is unreasonable, that there's no reasonable jurist out there. That's a totally different inquiry. And the same with respect to the materials that are before it. Yes, State — you know, you may have looked at a wide range of materials, you, the Federal habeas judge, and made your determination. Edpa says, for review, we want to look at only the Supreme Court cases. We don't care about the lower courts. It elevated the importance of the State court determination. So the idea that's informal or, you know, you could — they ask the same question. I think — and it's — maybe I'm just repeating Justice Alito's point, but they don't ask the same question. They ask a question as to whether what the State court did was not a reasonable application of clearly established Federal law. In this case, where the underlying constitutional violation requires Chapman, that is a very different review than when the underlying determination is, for instance, sufficiency under Jackson or inefficiency of counsel under Strickland. Both Jackson and Strickland require deference to the State. And then when you add Edpa on top of that, this Court has called that dual deference. When you're applying Chapman as the underlying standard, that requires the State to prove beyond a reasonable doubt that the error was harmless. It's a question, as this Court called it in the Kneader case, whether the evidence could rationally lead to a different verdict. If it could, then reversal is required. The Edpa lens in this case needs to be viewed in the context of Chapman, unlike the other standards. But, Ms. Bahal, I don't understand you to be arguing for straight-up Brecht. Don't you kind of argue for Brecht, but as limited with the guardrails of Edpa and that why the Sixth Circuit's decision was okay here is that it was Brecht, but they only considered clearly established Supreme Court law, just putting aside, just assuming that they did. And all of the differences that Judge Tappar points out in his dissent from the book review, you say, well, they did all that. It was just the substantive standards. You're advocating this hybrid thing, which seems to me kind of confusing. That's not really what Brecht said, so why not just for the sake of clarity to make it, you know, as Justice Breyer's pointing out, it's hard to unpack all this. For the sake of clarity, why not just tell courts, apply both. Kind of explain it like Judge Easterbrook did. Apply Edpa, and even if Edpa's relitigation bar would permit it, you know, apply Brecht too, and they have to pass both in order to get relief. So I agree. The approach we advocate here is applying Brecht, and if the Brecht inquiry finds actual prejudice without relying on sources of law that would be impermissible under 2254D, we know the answer to the Edpa inquiry. If there are sources that are relied upon, then formal application of Edpa might make sense. But Brecht didn't require that because Brecht preceded 2254D-1. So you're not really asking just for the application of Brecht. You're trying to meld the two together in a new test, right? I don't think of it as a new test. I think of it as assurance or a check that the Brecht test will actually subsume the Edpa inquiry. But again, here, formal application of Edpa confirms the result. The State court adjudication on the merits was contrary to clearly established law. They found the error harmless because there was an unacceptable risk of impermissible factors coming into play. That is not the Chapman test. That was a standard from this court's opinion in Holbrook that applied to determine if there was a constitutional violation by having four uniformed officers sitting behind the bar in the courtroom. That is not what should have been applied here. My friend on the other side agrees that Chapman is the underlying law. So formal application of here confirms the result. What about the fact that all the jurors testified that the shackles did not influence the verdict? Thank you, Justice Kavanaugh. This court has made clear first in Holbrook and again in DEC that relying on juror testimony as to whether the effect of shackles affected their verdict is unreliable because a juror will not always be aware of the effect of seeing a defendant in shackles. It has sort of a subconscious effect on the jurors. And so it is not at all surprising that a juror was not able to testify on the remand proceedings that yes, I saw the shackles and yes, they affected the verdict. This court recognized in Holbrook and DEC that the effect of shackling is implicit in the juror and they will not be able to articulate the reasons why the shackling is prejudicial. And second question, Chief Judge Sutton in his opinion joined by Judge Edpa, if the state courts issued a ruling on the harmlessness question under Chapman, then you apply Edpa. If the state court did not issue a ruling on the merits of the harmlessness question, then you apply Bracht. So not really applying both in every case, but first making that determination, did the state court actually conduct a harmlessness analysis? If so, Edpa. If not, Bracht. Anything to say for that approach? The statute itself requires where there is an adjudication on the merits that Edpa will apply. Here both sides agree there was an adjudication on the merits and so the 2254D question does apply as does Bracht. We think the Bracht question answers the 2254 inquiry, but that is one way that Edpa can be informally applied through Bracht. Both tests apply and both tests have been satisfied here. You mentioned that the state Supreme Court referred to an unacceptable risk. Is there any reason why that phrase in a very short opinion should not be understood to mean a risk that cannot be ruled out beyond a reasonable doubt? They cited the test from Holbrook. We know the context in which the Holbrook court used that test. It was a test to determine whether there was a constitutional violation in the first place. The state concedes in their briefing that it was not a harmless error test. The test to be applied, as we all agree, was the Chapman test. There is no indication from the Supreme Court opinion that they applied Chapman. They certainly didn't cite it and there is no indication that they applied it at all. They did not hold the state to that burden of proving the error harmless beyond a reasonable doubt. If there are no further questions, I would ask this court to please affirm. I guess I have one further one. We talk about informally applying AEDPA and formally applying it. What do you understand that difference to be? I use that terminology in light of this Court's opinions in Frye. No, no. I know we've used the terminology. I just want you to explain to me what we meant. The formal application of AEDPA, as I understand it, requires making a determination as to what is the last reasoned opinion. Informal application through an understanding that the Brecht test will subsume the AEDPA inquiry means no matter what the last reasoned opinion from the state court is, it was unreasonable because there has been a finding of actual prejudice. And so the informal application doesn't require specifically making the determination as to the last reasoned opinion. Justice Thomas? Alito? Justice Kavanaugh, do you have anything further? No further questions. Thank you. Thank you. Rebuttal, Ms. Hamoud. Thank you, Mr. Chief Justice. If I may, I'd like to address Justice Thomas' and Justice Alito's question in terms of which takes precedent when there's a congressional mandate. And when there is a state court merits adjudication, this is the way we believe the test should work because that is the basic structure of AEDPA. If there is a state court merits adjudication, then they must start with AEDPA first. The point that we were trying to make is let's say a petitioner prevails because a court, for example, stated that Chapman is not beyond a reasonable doubt, but by probable cause standard. The petitioner then wouldn't go to a direct Chapman, pure Chapman application, then as the court stated in Frye, Brecht would apply. So in terms of what the court should articulate, if there is a state court merits adjudication, then AEDPA's highly deferential standards kicks in. And it makes sense that they should start there. And if a petitioner prevails under AEDPA, then we move over to the next test. I know that my friend had stated that the Sixth Circuit did just that when they asked the question in Brecht. And again, that's an independent question. We know me as a judge. We have to take the Sixth Circuit at their word when they specifically stated that the test, and there is no reason to ask whether the state court unreasonably applied Chapman. So to say that Brecht encompasses AEDPA is simply not true because, again, they ask different questions. And in this specific case and in cases to follow, it is important that this court, like the tests that have been suggested, when there is a state court merits adjudication, we start with AEDPA Chapman. If a petitioner prevails, then you move over to Brecht. But what happened here, if a state court finds that there is substantial or an injurious effect on the verdict, it thinks that that's already been articulated, the question is different. Just because I, a federal judge, disagree, or even as this court's jurisprudence had articulated, if I, a federal judge, believe they are wrong, they must still ask the question, is it beyond all fair-minded disagreement? Or could fair-minded, not biased, fair-minded jurists agree with the state court's conclusion? And that's at the heart of it. This is not really a case where somebody's saying, look, I believe one thing, let's call it X, but at the same time, I think a fair-minded person could disagree with me. Because what you're looking to the fair-minded person to decide is something completely different. The standard in Brecht is so much higher than the standard in Chapman that even when you import that level of deference, what the court said in Ayala, what the court said in Chapman, even when you import some deference, there's such a gap between the Chapman and the Brecht standard that the Brecht standard is necessarily going to be the greater one. Thank you, Justice Kagan. I agree with you when you're comparing Brecht and Chapman. Those are two harmless error tests. Aetba is completely different. Aetba is an overlay. But Aetba, you're exactly right. Aetba is an overlay on Chapman. And essentially what we decided in Ayala and in Frey is that even with that Aetba overlay, the Brecht standard doesn't get close to the Chapman standard, doesn't get close to the Brecht standard. Thank you, Your Honor. The question, Aetba was not a play in Frey. And never once in the court in Ayala reaffirmed that, that it displaced Aetba. And the court in Ayala went through an extensive analysis. And I understand the two sentences that talk about subsumes. And this is why this case is a perfect vehicle for this court to apply both of the tests and distinguish between the different standards, the different limitations, the different burdens. But this court in Ayala went through and did an extensive analysis showing how Ayala did not meet the test under Brecht. And separately, under Aetba Chapman. And that's what we're asking the court to do here today because those differences matter. And because there is confusion and tension. And this court should clarify, we believe, through its application and articulate the test, that you have articulated but not answered this question. And the Sixth Circuit certainly did not give the State's reference. Thank you, Your Honors. Thank you, Counsel. The case is submitted.