MCGRADY_FINAL.DOCX (DO NOT DELETE) 11/22/2011 9:07 AM WHOSE LINE IS IT ANYWAY?: A RETROSPECTIVE STUDY OF THE SUPREME COURT’S SPLIT ANALYSIS OF § 2254(D)(1) SINCE 2000 Daniel J. McGrady∗ I. INTRODUCTION In 1996, Congress passed the Antiterrorism and Effective Death Penalty Act (AEDPA or “the Act”).1 As the Supreme Court noted, AEDPA’s purpose was “to reduce delays in the execution of state and federal criminal sentences, particularly in capital cases.”2 Thus, the Act sought “to prevent ‘retrials’ on federal habeas and to give effect to state convictions to the extent possible under law” in an effort to advance the “principles of comity, finality, and federalism.”3 In sum, Congress attempted to curb what it deemed as an abuse by convicted prisoners of the extensive appeals process in the United States court system at that time.4 From its inception, scholars have recognized that the Act has indeed greatly limited the access of convicted state prisoners to the federal habeas system.5 ∗ J.D., May 2011, Seton Hall University School of Law; B.A., 2007, The College of William & Mary. Thanks to Professor John Cornwell for his guidance and to Sarah Geers for her comments and assistance. 1 Pub. L. No. 104-132, 110 Stat. 1214 (codified at 28 U.S.C. § 2254 (2006)). 2 Woodford v. Visciotti, 538 U.S. 202, 206 (2003). 3 Id. at 206 (quoting Williams v. Taylor, 529 U.S. 362, 386 (2000)). 4 See H.R. REP. NO. 104-518, at 111 (1996) (Conf. Rep.). Some scholars, howev- er, have suggested that Congress adopted AEDPA more for political reasons than be- cause of any concern for federalism or the efficiency of the judicial process. See, e.g., Bryan A. Stevenson, The Politics of Fear and Death: Successive Problems in Capital Federal Habeas Cases, 77 N.Y.U. L. REV. 699, 701 (2002) (“AEDPA was drafted, enacted, and signed in an atmosphere of fear. The legislation, which includes substantial cutbacks in the federal habeas corpus remedy, was Congress’s response to the tragedy of the Oklahoma City bombing.”). 5 See, e.g., James O. Nygard, Current Developments in the Law: A Survey of Cases Af- fecting the Anti-terrorism and Effective Death Penalty Act of 1996, 6 B.U. PUB. INT. L.J. 333, 338 (1996) (“The Antiterrorism and Effective Death Penalty Act of 1996 limits the Court’s power to review petitions for writs of habeas corpus.”); Tung Yin, A Better Mousetrap: Procedural Default as a Retroactivity Alternative to Teague v. Lane and the Anti- terrorism and Effective Death Penalty Act of 1996, 25 AM. J. CRIM. L. 203, 206 (1998) (“The Act limits the scope of federal habeas review, wherein the state has rendered a deci- sion on the merits on the claim, to determining whether that decision ‘was contrary 1599 MCGRADY_FINAL.DOCX (DO NOT DELETE) 11/22/2011 9:07 AM 1600 SETON HALL LAW REVIEW [Vol. 41:1599 Prior to AEDPA, federal habeas corpus courts reviewed “the peti- tioner’s legal claims de novo in the strictest sense of that term.”6 With respect to “questions of law or mixed questions of law and fact, the federal courts treated the petition as a wholly new complaint, which originated ‘an independent civil suit’ and deserved to be adjudicated ‘from scratch.’”7 AEDPA changed these norms in a number of ways. Perhaps the most significant alteration was § 2254(d)(1). The sta- tute, in pertinent part, states the following: (d) An application for a writ of habeas corpus on behalf of a per- son in custody pursuant to the judgment of a State court shall not be granted with respect to any claim that was adjudicated on the merits in State court proceedings unless the adjudication of the claim— (1) resulted in a decision that was contrary to, or involved an un- reasonable application of, clearly established Federal law, as deter- mined by the Supreme Court of the United States.8 First, note that the statute, by its terms, applies only to state-prisoner habeas corpus petitions.9 As the provision suggests, the threshold question is whether a claim has been “adjudicated on the merits” at the state court level.10 Yet even if this requirement is satisfied, the pe- titioner may not simply assert any and all issues that he or she has with the state court decision. To the contrary, for a federal court to grant the writ of habeas corpus under § 2254(d)(1), a petitioner must demonstrate that the state court decision was inconsistent with feder- al law.11 In the determination of this latter point, AEDPA drastically changed the preexisting habeas corpus process. The statute provides two requirements for establishing a reversible federal-law error. First, a federal court may only review federal-law claims, which the provi- sion limits to “clearly established” precedents of the “Supreme Court of the United States.”12 Because federal courts formerly had a much broader body of law from which to draw when making decisions,13 to, or involved an unreasonable application of, clearly established Federal law, as de- termined by the Supreme Court of the United States.’”). 6 2 RANDY HERTZ & JAMES S. LIEBMAN, FEDERAL HABEAS CORPUS PRACTICE AND PROCEDURE § 32.1 (5th ed. 2005). 7 Id. 8 28 U.S.C. § 2254(d)(1) (2006) (emphasis added). 9 HERTZ & LIEBMAN, supra note 6, § 32.1. 10 § 2254(d). 11 § 2254(d)(1). 12 Id. 13 HERTZ & LIEBMAN, supra note 6, § 32.1 MCGRADY_FINAL.DOCX (DO NOT DELETE) 11/22/2011 9:07 AM 2011] COMMENT 1601 Congress effectually reduced the potential grounds for habeas review. Second, the provision requires that a state decision be “contrary to, or involve[] an unreasonable application of” that clearly established law before a court may grant a habeas petition.14 Thus, in addition to limiting the potential grounds for habeas review, Congress also al- tered the de novo standard of review of federal legal claims that ex- isted prior to AEDPA. By restricting the reviewable “federal law” and adopting the more stringent “contrary to” and “unreasonable appli- cation” language, AEDPA radically diminished the scope of federal review. After years of difficulty for the circuit courts in interpreting § 2254(d)(1),15 in 2000 the Supreme Court issued a definitive state- ment in Williams v. Taylor,16 which continues to be the governing law today. Essentially, Williams established that the “contrary to” and “unreasonable application” language in the statute formulated two different tests within the standard even though they are contained in the same sentence.17 A decision is “contrary to” federal law when a state court applies the wrong federal law entirely.18 By contrast, a state court unreasonably applies federal law when its application of the correct precedent is “objectively unreasonable.”19 The “contrary to” test is typically, but not always, uncontroversial, because state courts usually will at least identify the correct federal law to apply.20 Thus, the focus of this Comment is the “unreasonable application” test. Here, the Williams Court established its lasting legacy by insist- ing that an incorrect or erroneous application of federal law is different Until 1989 . . . federal courts could adjudicate the prisoner’s “com- plaint” on the basis of legal principles that were not even in existence at the time the state ruled. And until AEDPA was passed, federal courts could adjudicate the complaint on the basis of legal principles that had never become binding on state courts, i.e., on the basis of federal cir- cuit law developed in the absence of any controlling Supreme Court law on point. Id. 14 Id. 15 See infra note 64. 16 529 U.S. 362 (2000). 17 Id. at 404–05. 18 See id. at 405–06. 19 Id. at 409. 20 An example of how the “contrary to” test operates can be found in Williams it- self. Justice O’Connor held that the Virginia Supreme Court’s decision was “contrary to” clearly established federal law because the state court applied Lockhart v. Fretwell, 506 U.S. 364 (1993), instead of Strickland v. Washington, 466 U.S. 668 (1984). Wil- liams, 529 U.S. at 413–14 (O’Connor, J., concurring). MCGRADY_FINAL.DOCX (DO NOT DELETE) 11/22/2011 9:07 AM 1602 SETON HALL LAW REVIEW [Vol. 41:1599 from an objectively unreasonable application of that law.21 Accordingly, Williams cemented a notion that had existed, albeit tenuously, since the adoption of AEDPA: even if a federal court believes in its inde- pendent judgment that a state-court decision applied federal law in- correctly, it must deny a petition for habeas corpus so long as the de- cision was “objectively reasonable.”22 The interpretative gap between incorrect and unreasonable has been the subject of many Supreme Court decisions since 2000. Be- cause this Comment focuses exclusively on the application of § 2254(d)(1) to capital punishment cases, its analysis is necessarily li- mited to decisions involving capital sentences.23 In this context, the results are staggering. The so-called conservative bloc of the Court, which at various times has consisted of Justices Scalia, Rehnquist, Thomas, Roberts, and Alito, has found a state-court decision to be an “unreasonable application” of federal law in only 4–14% of cases.24 By contrast, Justices Ginsburg, Breyer, and Stevens—the so-called lib- eral bloc of the Court—have found the states’ decisions to be an “un- reasonable application” of the law 53–64% of the time.25 A review of the Court’s cases interpreting § 2254(d)(1) in the capital context also reveals that the decisions in this area are often split five-to-four, with the holdings hinging solely on the analyses of Justices Kennedy and O’Connor.26 The latter two justices have applied a more genuine— yet entirely undefined—standard in reaching their conclusions.27 A retrospective analysis of the Court’s decisions since Williams v. Taylor demonstrates that the two sides of the Court are not merely disagreeing over the application of the present standard; they are ac- tually applying two completely different standards. The conservative bloc, to whom I refer throughout this Comment as the “blind defe- rence camp,” has blindly deferred to the state court’s interpretation of federal law.28 On the other hand, the more liberal justices, to whom I refer throughout this Comment as the “de novo camp,” have 21 Id. at 412. 22 See infra Part II.B. 23 Contrary to what its name suggests, the AEDPA is not exclusively limited to capital cases. For purposes of this Comment—in the interest of clarity and brevity, and also because I view the results in capital cases to be particularly troublesome—I have elected to restrict my focus accordingly. 24 See infra Part IV.A. 25 See infra Part IV.A. 26 See infra Part V.A. 27 See infra Part V.A. 28 See infra Part IV.A. MCGRADY_FINAL.DOCX (DO NOT DELETE) 11/22/2011 9:07 AM 2011] COMMENT 1603 applied de novo review akin to what was in place prior to AEDPA.29 This separation between the two blocs is stark and evidences three critical problems stemming from the provision as it has been applied since Williams v. Taylor. First, several cases with similar facts have led to inconsistent re- sults. Often split five-to-four, such decisions demonstrate that the standard has become arbitrary with results hinging on which camp Justices Kennedy and O’Connor aligned with in a particular case.30 The second issue flows from the first: namely, the provision fails to provide petitioners with the proper notice of what is required to chal- lenge a state decision.31 Finally, and most importantly, the inconsis- tent standard is dangerous because the differences in opinion regard- ing the provision in capital cases result in life-or-death consequences.32 Given these considerations and the fact that federal courts are generally in a better position to evaluate federal law than the state courts where these claims originate, the Supreme Court should at the very least define the standard by which “reasonableness” is to be judged. Yet it seems that the optimal solution to the current dilemma is a congressional revision of the statue creating a new ex- ception: in capital cases, federal law claims should be reviewed as they were prior to AEDPA—de novo.33 Part II of this comment will outline habeas corpus procedure as it existed prior to AEDPA, as well as the legislative concerns that led to its adoption. Part III will discuss the Williams v. Taylor standard and how it changed the existing interpretations of § 2254(d)(1). Part IV will provide an overview of how the Supreme Court has inter- preted the provision since 2000 and identify the wide and problemat- ic split between liberal and conservative justices over the meaning of § 2254(d)(1). Part V will analyze the three main problems with the Court’s split analysis: namely, that the standard has become arbitrary, that it fails to provide proper notice, and that it is dangerous in the context of life-or-death consequences. Part VI will propose potential solutions and explain why a balancing of the relevant harms weighs in favor of altering the statute to provide a de novo standard of review. 29 See infra Part IV.A. 30 See infra Part V.A. 31 See infra Part V.B. 32 See infra Part V.C. 33 Recognizing the unique concerns involved with capital punishment is not un- precedented. See, e.g., Death Penalty Is Repealed in New Mexico, N.Y. TIMES, Mar. 18, 2009, at A16 (explaining that fifteen states have now barred the death penalty). MCGRADY_FINAL.DOCX (DO NOT DELETE) 11/22/2011 9:07 AM 1604 SETON HALL LAW REVIEW [Vol. 41:1599 II. HABEAS CORPUS PROCEDURE PRIOR TO AEDPA AND LEGISLATIVE HISTORY A. Habeas Prior to AEDPA The writ of habeas corpus derives from English common law. In England, the writ established a system enabling convicted prisoners to contest illegal imprisonments.34 The United States acknowledged the writ in its Constitution, which states that “the writ of habeas cor- pus shall not be suspended, unless when in cases of rebellion or inva- sion the public safety may require it.”35 In 1867, Congress extended the writ of habeas corpus to state prisoners via the Habeas Corpus Act; however, the act disallowed claims alleging violations of state law and restricted federal habeas to claims stating a federal-law viola- tion.36 Thus, grant of the writ in federal court has always required a petitioner to demonstrate that he was being held in “custody in viola- tion of the Constitution or laws or treaties of the United States.”37 Given this, a federal court will not grant the writ of habeas corpus where there has been an error in the application of state law.38 Ac- cordingly, the “vast majority of habeas corpus petitions by state pris- oners allege a violation of the constitution.”39 As discussed earlier, prior to AEDPA, the courts were permitted to find that convicted state prisoners were being held “in violation of the Constitution or laws or treaties of the United States” where a low- er court applied federal law incorrectly.40 An example of the Supreme Court’s application of de novo review to federal law claims prior to AEDPA is Burger v. Kemp, where the Court affirmed the denial of a writ of habeas corpus to a petitioner claiming ineffective counsel.41 In that case, a Georgia jury sentenced petitioner Burger to death for murder.42 Burger claimed that his counsel was ineffective because (1) a conflict of interest existed when his counsel’s partners represented his co-indictees, and (2) his counsel failed to “develop and present 34 MARK E. CAMMACK & NORMAN M. GARLAND, ADVANCED CRIMINAL PROCEDURE IN A NUTSHELL 483 (2d ed. 2006). 35 U.S. CONST. art. I, § 9, cl. 2. 36 The Habeas Corpus Act of 1867, ch. 28, 14 Stat. 385 (current version at 28 U.S.C. § 2241); see also CAMMACK & GARLAND, supra note 34, at 483. 37 CAMMACK & GARLAND, supra note 34, at 483. 38 Id. at 487. 39 Id. 40 See supra notes 6–7 and accompanying text. 41 483 U.S. 776 (1987). 42 Id. at 777. MCGRADY_FINAL.DOCX (DO NOT DELETE) 11/22/2011 9:07 AM 2011] COMMENT 1605 mitigating evidence” at the sentencing hearings.43 While the Court rightly deferred to the lower courts’ factual findings, it applied those facts de novo with respect to the federal legal claims. The Court rea- soned that the Strickland standard,44 as it is laid out, requires defe- rence to counsel’s decisions.45 Then, the majority directly evaluated the evidence presented of Burger’s troublesome circumstances and came to an independent legal judgment that Burger had not shown that, “in light of all the circumstances, the identified acts or omissions [of counsel] were outside the wide range of professionally competent assistance.”46 The Court did not base the holding on the “objective reasonability” of the state court’s application of federal law regarding ineffective assistance of counsel; rather, the Court simply believed that the state court did not apply the Strickland standard incorrectly. As Burger v. Kemp illustrates, federal courts possessed considerable flexibility in evaluating federal law claims prior to AEDPA. B. Legislative History In the wake of the Oklahoma City bombing, some lawmakers began to view the flexibility of the habeas corpus law as an enabling device for convicted prisoners to delay their sentences.47 If courts re- viewed every claim de novo, they feared, then the extensive appeals 43 Id. at 788. 44 The Strickland standard states the following: A convicted defendant’s claim that counsel’s assistance was so defective as to require reversal of a conviction or death sentence has two com- ponents. First, the defendant must show that counsel’s performance was deficient. This requires showing that counsel made errors so se- rious that counsel was not functioning as the “counsel” guaranteed the defendant by the Sixth Amendment. Second, the defendant must show that the deficient performance prejudiced the defense. This re- quires showing that counsel’s errors were so serious as to deprive the defendant of a fair trial, a trial whose result is reliable. Unless a defen- dant makes both showings, it cannot be said that the conviction or death sentence resulted from a breakdown in the adversary process that renders the result unreliable. Strickland v. Washington, 466 U.S. 668, 687 (1984). 45 This use of deference should not be confused with the level of review federal courts apply to this standard. In other words, at that time federal courts were re- quired to review de novo this standard of deference to counsel. 46 Burger, 483 U.S. at 795. 47 See H.R. REP. NO. 104-518, at 111 (1996) (Conf. Rep.); 142 CONG. REC. S3454, 3463 (daily ed. April 17, 1996) (statement of Sen. Bradley); see also Stevenson, supra note 4, at 704. MCGRADY_FINAL.DOCX (DO NOT DELETE) 11/22/2011 9:07 AM 1606 SETON HALL LAW REVIEW [Vol. 41:1599 process in the United States would result in endless “retrials.”48 Seek- ing to finalize punishments and accord proper respect for state deci- sions, congressional representatives debated the best mechanism to accomplish their goals. Congress, through § 2254(d)(1), ultimately decided to restrict the reviewable federal law and heighten the de- gree of incorrectness required to overturn a state level decision. The legislative history of the Act reveals the intentions and con- cerns of the authors of § 2254(d)(1).49 Several members of Congress worried that federal courts would interpret the “unreasonable appli- cation” language as establishing a “wrong-but-reasonable standard.”50 For example, Senator Patrick Moynihan, a critic of the statute, warned, “We are about to enact a statute which would hold that con- stitutional protections do not exist unless they have been unreasona- bly violated, an idea that would have confounded the framers. Thus we introduce a virus that will surely spread throughout our system of laws.”51 Senator Edward Kennedy’s concerns were even more pointed. He interpreted the statute as requiring federal courts to simply “defer” to state-level decisions.52 In short, members of Con- gress predicted that incorrect applications of federal law would be protected behind the shield of “unreasonableness,” which would render even some patently wrong decisions unreviewable under ha- beas procedure. In this way, some congressional representatives were concerned—and as subsequent cases would show, rightly con- cerned—that AEDPA “would prevent federal courts from granting habeas petitions when a state court decision was wrong as a matter of federal law, but nonetheless reasonable (under some undisclosed construction of that term).”53 This last point indicates that it was not 48 See H.R. REP. NO. 104-518, at 111 (1996) (Conf. Rep.); 142 CONG. REC. S3454, 3463 (daily ed. April 17, 1996) (statement of Sen. Bradley); see also Stevenson, supra note 4, at 704. 49 As an aside, the Bill passed 91–8–1 in the Senate and 293–133–7 in the House of Representatives. 142 CONG. REC. S3454 (daily ed. April 17, 1996); 142 CONG. REC. H3618 (daily ed. April 18, 1996). 50 See infra notes 51–57 and accompanying text; see also Allan Ides, Habeas Stan- dards of Review Under 28 U.S.C. § 2254(d)(1): A Commentary on Statutory Text and Su- preme Court Precedent, 60 WASH. & LEE L. REV. 677, 693 (2003). 51 142 CONG. REC. S3438 (daily ed. April 17, 1996) (statement of Sen. Moynihan); see also Ides, supra note 50, at 693 n.29. 52 142 CONG. REC. S3458 (daily ed. April 17, 1996) (statement of Sen. Kennedy) (“[T]he proposal would unwisely require Federal courts to defer to State courts on issues of Federal constitutional law. A Federal court could not grant a writ habeas corpus based on Federal constitutional claims, unless the State court’s judgment was ‘an unreasonable application of Federal law.’”). 53 Ides, supra note 50, at 693. MCGRADY_FINAL.DOCX (DO NOT DELETE) 11/22/2011 9:07 AM 2011] COMMENT 1607 only the standard of “reasonableness” that troubled critics, but the potential danger arising from the fact that the standard was “unde- fined.” Even the supporters of AEDPA spurned a wrong-but-reasonable interpretation of § 2254(d)(1).54 For example, Senator Joseph Biden identified the potential difficulties implicated by a “wrong-but- reasonable” approach: [T]he bill seems to allow an exception to the general rule [against granting habeas] but one that is likely to be illusory be- cause a claim can be granted only if the State court’s application of Federal law to the facts [is] not merely wrong but unreasona- ble. This is an extraordinar[ily] deferential standard to the State courts, and I believe it is an inappropriate one. It puts the Feder- al courts in the difficult position of evaluating the reasonableness of a State court judge rather than simply deciding whether or not he correctly applied the law, not whether he did it reasonably. You can have a reasonable mistake. They could reasonably con- clude that on a constitutional provision, it should not apply, when 55 in fact the Supreme Court would rule it must apply. Senator Arlen Specter, a cosponsor of the bill, stated that he was “not entirely comfortable” with the “unreasonable application” standard but he ultimately concluded “that the standard in the bill will allow Federal courts sufficient discretion to ensure that convictions in State court have been obtained in conformity with the Constitution.”56 Senator Orin Hatch, also a cosponsor of the bill, rebutted concerns over the “wrong-but-reasonable” loophole by arguing that the stan- dard “enables the Federal court to overturn State court positions that clearly contravene Federal law. It further allows the Federal courts to review State court decisions that improperly apply clearly established Federal law.”57 Senator Hatch’s use of the term “improperly” instead of “unreasonably” suggests that even he, as a supporter, was using the terms interchangeably. This seems curiously close to equating “in- correct” with “unreasonable,” an interpretation that the Williams v. Taylor decision would later expressly reject.58 54 Id. at 694. 55 141 CONG. REC. S7842 (daily ed. June 7, 1995) (statement of Sen. Biden); see also Ides, supra note 50, at 694. 56 142 CONG. REC. S3472 (daily ed. April 17, 1996) (statement of Sen. Specter); see also Ides, supra note 50, at 695. 57 141 CONG. REC. S7846 (daily ed. June 7, 1995) (statement of Sen. Hatch) (em- phasis added); see also Ides, supra note 50, at 695. 58 See infra note 79. Other congressional representatives were even more explicit about equating “incorrect” with “unreasonable.” See, e.g., 142 CONG. REC. S3465 (dai- ly ed. April, 17, 1996) (statement of Sen. Levin) (“I believe the courts will conclude, MCGRADY_FINAL.DOCX (DO NOT DELETE) 11/22/2011 9:07 AM 1608 SETON HALL LAW REVIEW [Vol. 41:1599 Moreover, even President Clinton acknowledged the danger of allowing federal courts to defer to state decisions involving federal law. Upon signing the law, the President released the following statement: “I have signed this bill because I am confident that the Federal courts will interpret these provisions to preserve independent review of Federal legal claims and the bedrock constitutional principle of an independent judiciary.”59 Therefore, as Professor Ides has ar- gued, “to the extent that the legislative history is informative, it re- veals some concern that the ‘unreasonable application’ standard of review might be read to preclude federal court review of a state court decision that could be described as wrong-but-reasonable.”60 Yet if § 2254(d)(1) is unclear regarding “the scope of the review power em- bodied in the ‘unreasonable application’ principle, the . . . legislative history seems to establish . . . that such a radically innovative wrong- but-reasonable standard has no legitimate place within the sphere of inter- pretive possibilities.”61 After considering this legislative history, one wonders what these same members of Congress would have thought if they could have fo- reseen the holdings in later cases, such as Woodford v. Visciotti62 and Early v. Packer.63 Both opinions explicitly endorsed the “wrong-but- reasonable” interpretation of the statute. While the holdings un- doubtedly conflict with the aforementioned congressional senti- ments, Congress has yet to modify the statute to make it absolutely clear that such interpretations are incorrect. The Court is not obli- as they should, that a constitutional error cannot be reasonable and that if a State court decision is wrong, it must necessarily be unreasonable.”). 59 Statement on Signing the Antiterrorism and Effective Death Penalty Act of 1996, 32 WEEKLY COMP. PRES. DOC. 719 (April 26, 1996) (emphasis added), available at http://clinton6.nara.gov/1996/04/1996-04-24-president-statement-on- antiterrorism-bill-signing.html. 60 See Ides, infra note 50, at 697. 61 Id. (emphasis added). 62 537 U.S. 19 (2002). The Court reversed a grant of habeas relief in a per cu- riam decision. Id. at 20. The majority held that even if the state court was wrong, it is not enough to convince a federal habeas court that, in its independent judgment, the state-court decision applied Strickland incorrectly. Id. at 27. Note that while this per curiam decision appears lopsided, it is only because the case involved a denial of habeas corpus rights. When the independent judgment of the rest of the Court aligns with the conservative bloc’s blind deference to the state court, lopsided results will sometimes emerge. This, however, is not proof that the different sects of the court are applying the same standard. 63 537 U.S. 3 (2002). The Court reversed a grant of habeas relief in a per curiam decision. Id. at 4. Even if the Court agreed with the Ninth Circuit majority that there was jury coercion here, the Court stated that “it is at least reasonable to con- clude that there was not, which means that the state court’s determination to that effect must stand.” Id. at 11.