In the Supreme Court of the United States
JOHN FRANCIS FRY, PETITIONER
CHERYL K. PLILER, WARDEN
ON WRIT OF CERTIORARI
TO THE UNITED STATES COURT OF APPEALS
FOR THE NINTH CIRCUIT
BRIEF FOR THE UNITED STATES
AS AMICUS CURIAE SUPPORTING RESPONDENT
PAUL D. CLEMENT
Counsel of Record
ALICE S. FISHER
Assistant Attorney General
MICHAEL R. DREEBEN
Deputy Solicitor General
JONATHAN L. MARCUS
Assistant to the Solicitor
JOEL M. GERSHOWITZ
Attorney Department of Justice
Washington, D.C. 20530-0001
The United States will address the following question:
Whether the "substantial and injurious effect" harm less-error standard for constitutional trial errors adop ted in Brecht v. Abrahamson, 507 U.S. 619 (1993), for habeas proceedings under 28 U.S.C. 2254 applies re gardless of whether the state appellate court recognized the constitutional error and reviewed it for harmless ness under Chapman v. California, 386 U.S. 18 (1967).
In the Supreme Court of the United States
JOHN FRANCIS FRY, PETITIONER
CHERYL K. PLILER, WARDEN
ON WRIT OF CERTIORARI
TO THE UNITED STATES COURT OF APPEALS
FOR THE NINTH CIRCUIT
BRIEF FOR THE UNITED STATES
AS AMICUS CURIAE SUPPORTING RESPONDENT
INTEREST OF THE UNITED STATES
This case presents the question whether the harm less-error standard for constitutional trial errors set out in Brecht v. Abrahamson, 507 U.S. 619 (1993), for ha beas proceedings under 28 U.S.C. 2254 applies whether or not the state appellate courts recognized the constitu tional error and reviewed it under the harmless-error standard of Chapman v. California, 386 U.S. 18 (1967). Although this case arises from a habeas petition filed by a state prisoner pursuant to Section 2254, the question presented also arises in cases brought by federal prison ers pursuant to 28 U.S.C. 2255, when an intervening change in the law prompts a federal prisoner to raise in a Section 2255 motion a constitutional claim that was either not raised or found to be meritless on direct review.1 Because the question presented implicates the harmless-error standard applicable to constitutional trial errors asserted in Section 2255 proceedings, the United States has a substantial interest in its correct resolution.
1. In the early morning of October 27, 1992, the bod ies of James and Cynthia Bell were discovered in a vehi cle parked on the shoulder of Interstate 505 in northern California. Cynthia, who was seated on the driver's side, had been shot in the head. James, who was seated in the front passenger seat, had been shot eight times in his head, chest, back and wrist. J.A. 26-27. At least four witnesses saw an older-model pickup truck with a light- colored camper shell parked at the crime scene next to the victim's car during the evening of October 26, 1992. J.A. 28. Petitioner was charged with the murders. After his first two trials ended in hung juries, petitioner was convicted of both murders at his third jury trial and sen tenced to life imprisonment without the possibility of parole. J.A. 67-68.
Numerous witnesses testified at petitioner's trial that Cynthia owed petitioner money for drugs, and that petitioner had been angered when he caught her roam ing about his house, because he believed she was trying to steal from him. At least three witnesses testified that they heard petitioner threaten to kill Cynthia in the days immediately preceding the murders. Several wit nesses, including petitioner's brother, testified that when petitioner returned home on the night of the mur ders, he was spotted with blood. Petitioner's handgun also had blood on it and contained four spent shells.2 That same night, petitioner made incriminating state ments to his brother and several other witnesses, saying he "did" Cynthia and James Bell and describing in vivid detail the impact of a .357 magnum bullet on a human head. Expert testimony established that one of the bul lets recovered from the victims had definitely been fired from petitioner's gun. When the police attempted to take petitioner into custody, he fled from the police in a high-speed chase. He confessed to another inmate that he had committed the murders, and tried to arrange to have his brother injured in order to persuade him to change his testimony. J.A. 71.
Petitioner testified at trial. While denying that he committed the murders, petitioner admitted that at the time of the murders he owned a 1977 powder blue pickup truck with a white camper shell and a .357 magnum. He admitted that Cynthia Bell dealt drugs for him and owed him money at the time she was murdered. Petitioner further admitted driving an acquaintance-who had tes tified that petitioner believed she (the acquaintance) had stolen money from him and had threatened to kill her (J.A. 29)-into the country two days before the murders, taking out his gun and shooting it twice out the window. He admitted that when he arrived home on the night of the murders, he "pulled into the driveway and sat there for a while," "trying to maintain control" and "get a fo cus." J.A. 53. Petitioner admitted going into the bath room upon entering the house, but denied changing his clothes or having blood on them. He admitted that he removed the camper shell from his pickup truck the morning after the murders, but claimed that he did so to save gas mileage. Petitioner also admitted lying to the police when asked when he had last seen his .357. Fi nally, while petitioner did not admit that he had ar ranged to have an inmate harm his brother, petitioner admitted that he gave the inmate his brother's address. J.A. 49-55.
2. At trial, petitioner sought to establish that one or more other individuals committed the murders, includ ing Anthony Hurtz. He introduced evidence that Hurtz had admitted his involvement to several witnesses. See J.A. 61-64. The trial court excluded the testimony of an additional witness, Pamela Maples, that she overheard Hurtz state that he committed murders that resembled the circumstances of the murder of the Bells. J.A. 17.
Maples testified at trial that, while at her sister's house in Vallejo in April 1994, she "overheard" Hurtz discussing some homicides. The prosecution objected to the testimony, and the defense made an offer of proof. Maples testified outside the presence of the jury that, during a conversation among her sister, Hurtz, and "a guy named Steve," she heard Hurtz state that he had killed a man and a woman in a car. Maples was "in and out of the room" at the time, only heard "bits and pieces" of the conversation, and did not participate in it. The conversation was already underway when she came in, and she did not know how long the participants dis cussed the murders, the context of the discussion, or whether the discussion was "serious" or not. Maples testified that Hurtz said that he shot the girl in the head first and then reached over and shot the guy. Hurtz said the victims were in a parking place, but he did not de scribe either the location of the murders or the type of car the victims were in. Maples further testified that Hurtz stated that he was "full of blood" after the mur ders, and that he ran through a field, threw mud on him self, and phoned someone to pick him up. Maples did not hear Hurtz state when the murders took place. Under cross-examination, Maples acknowledged that she had no idea whether the murders about which Hurtz spoke took place one week or ten years before the conversa tion, or whether they occurred in California or else where. The trial court sustained the State's objection to this testimony, concluding that the defense had provided insufficient evidence to connect the incident described by Hurtz to the murder of the Bells. J.A. 7-17, 94-95.
3. Petitioner took an appeal to the California Court of Appeals, contending, among other things, that the exclusion of Maples's testimony violated his constitu tional right to present a defense and to a fair trial. The court of appeals concluded that the trial court did not abuse its discretion in excluding the testimony. The court explained that, "[g]iven the absence of any actual linkage to the charged offenses, or any other evidence supporting an inference that the killings allegedly de scribed by Hurtz were those of Cynthia and James Bell, Maples' testimony simply did not tend to show [peti tioner's] innocence." J.A. 97. The court added that "no possible prejudice" could have resulted from the exclu sion of the evidence because Maples's testimony was "cumulative" of the testimony of other defense witnesses that Hurtz had confessed to shooting two unidentified people at close range while the victims were seated in a parked car. J.A. 97 n.17. In finding that the exclusion of the testimony was nonprejudicial, the court did not specify the harmless-error standard it was applying. Thereafter, petitioner filed a petition for review in the Supreme Court of California, which was denied. J.A. 113. Petitioner did not file a petition for a writ of certio rari seeking direct review of his conviction.
4. Petitioner moved in the United States District Court for the Eastern District of California for a writ of habeas corpus under 28 U.S.C. 2254. J.A. 114. He raised several issues, including the exclusion of Maples's testi mony at his trial. The magistrate judge concluded that the exclusion of Maples's testimony violated the Sixth and Fourteenth Amendments because "the evidence of third-party culpability . . . . was sufficiently 'crucial or reliable' to outweigh the state's interest in exclusion of untrustworthy evidence." J.A. 180 (citing Chambers v. Mississippi, 410 U.S. 284 (1973); Washington v. Texas, 388 U.S. 14 (1967)). The magistrate judge explained that, "[a]lthough the offer of proof regarding Maples is in parts imprecise, certain significant points match or closely parallel some of the key circumstances of the case at issue"-to wit, "[b]oth victims had been shot to death; victim Cynthia Bell had been shot in the head; the victims were in a car, parked in a turnout area; evidence showed she was in the driver's seat and James Bell was in the passenger seat and that the shooter was standing outside the vehicle's driver side; testimony revealed that the shooter would have been covered in blood." J.A. 179. The magistrate judge also took issue with the state ap pellate court's statement that no other evidence sup ported the inference that Hurtz had described the mur der of the Bells, observing that seven other witnesses "specifically link[ed] Hurtz to the deaths of the Bells." Ibid. The magistrate judge held that "[f]or the state appellate court not to recognize the trial court's error was an unreasonable application of clearly established law as set forth by the Supreme Court." J.A. 180.
The magistrate judge then addressed whether the constitutional error was harmless. J.A. 180-182. The magistrate judge concluded that the error was harmless under the standard set forth in Brecht v. Abrahamson, 507 U.S. 619, 623 (1993), under which an error is deemed harmful if it "had substantial and injurious effect or in fluence in determining the jury's verdict." See J.A. 180 (quoting Brecht). Although the magistrate judge did not agree with the state appellate court that "no possible prejudice" could have resulted from the exclusion of Ma ples's testimony, he found that there had been an "in sufficient showing" that the exclusion of the testimony had a "substantial and injurious effect" on the verdict. J.A. 181-182. The magistrate judge explained that the jury heard other defense witnesses implicate Hurtz in the murders in "far more damning detail" and that Ma ples's testimony "could not have emerged completely un scathed by a cross-examination that would have pointed up [its] infirmities." J.A. 181. Having thus found the constitutional error to be harmless, the magistrate judge recommended that habeas relief be denied. J.A. 182.
The district court conducted a de novo review of the case, adopted the findings and recommendations of the magistrate judge, and denied the habeas petition. J.A. 208-209.
5. Petitioner took an appeal to the Ninth Circuit. That court agreed with the district court that the exclu sion of Maples's testimony entailed an unreasonable application of clearly established federal law because the testimony was "material and would have substantially bolstered [petitioner's] claims of innocence." J.A. 212 (quoting Chia v. Cambra, 360 F.3d 997, 1003 (9th Cir. 2004)). The court, however, also agreed with the district court that the error was harmless under the Brecht st andard. Ibid. The court rejected petitioner's claim that the Brecht harmless-error standard was inapplicable because the state appellate court had "failed to conduct a meaningful prejudice review." J.A. 212 n.3. The court explained that, under its decisions, "the Brecht standard applies uniformly in all Federal habeas corpus cases under § 2254 regardless of the error standard, if any, applied by the state court." Ibid. (citing Bains v. Cam bra, 204 F.3d 964, 976 (9th Cir.), cert. denied, 531 U.S. 1037 (2000), and Inthavong v. LaMarque, 420 F.3d 1055, 1059 (9th Cir. 2005), cert. denied, 126 S. Ct. 1660 (2006)).
Judge Rawlinson dissented. J.A. 213-214. Applying the Brecht standard, Judge Rawlinson concluded that the exclusion of Maples's testimony was not harmless. J.A. 214.
SUMMARY OF ARGUMENT
Harmless-error review under the standard set out in Brecht v. Abrahamson, 507 U.S. 619 (1993), applies in a federal habeas proceeding under 28 U.S.C. 2254 when ever the court finds a constitutional trial error, regard less of whether the state appellate courts reviewed that error under the harmless-error standard of Chapman v. California, 386 U.S. 18 (1967). Brecht adopted the less onerous harmless-error standard of Kotteakos v. United States, 328 U.S. 750 (1946), for federal constitutional trial errors asserted in habeas proceedings under 28 U.S.C. 2254 because of concerns about finality and fed eralism. The scope and rationale of the Court's opinion justify the application of the Brecht standard across the board.
Petitioner contends that Brecht's scope is limited to cases in which the state appellate court had found (as the Wisconsin Supreme Court had in Brecht) the consti tutional error to be harmless beyond a reasonable doubt under Chapman. Petitioner's contention lacks merit.
First, the majority opinion in Brecht made clear that it found the Chapman standard unsuitable for collateral review for a number of reasons, none of which related to the content of the state appellate court's legal determi nations. The Court mentioned only one possible excep tion to its "hold[ing] that the Kotteakos harmless-error standard applies in determining whether habeas relief must be granted because of constitutional error of the trial type," 507 U.S. at 638 (footnote omitted), and that potential exception (for intentional errors) does not turn on the nature of the state courts' appellate review and is far more narrow than the one for which petitioner advo cates.
Second, all the reasons the Brecht Court identified for rejecting Chapman on collateral review apply with the same force to cases in which the state courts on direct appeal did not conduct Chapman review. The Brecht Court first cited the States' interest in the final ity of their criminal judgments. That interest arises regardless of the precise content of the state appellate courts' legal determinations in upholding the conviction. The Brecht Court next cited the intrusion federal habeas review visits on "the States' sovereign power to punish offenders and their good-faith attempts to honor consti tutional rights." 507 U.S. at 635 (quoting Engle v. Isaac, 456 U.S. 107, 128 (1982)). The federal habeas courts' infringement on state sovereignty is no less severe in cases in which the state courts found no federal constitu tional error. The Brecht Court also cited the historic purpose of the writ to provide relief only to those prison ers "grievously wronged." Id. at 637. A state prisoner is no more or less "wronged" by a state court decision erroneously finding no constitutional error than he is by a state court decision erroneously determining that the error was harmless. Finally, the Court cited the "social costs" of granting habeas relief, ibid., pointing out, among other things, the difficulties in retrying defen dants many years after the case was brought. Those costs do not vary based on the reason the state courts denied relief.
Petitioner's proposed approach not only finds no sup port in Brecht, but it would also produce inequitable results. Similarly situated prisoners whose constitu tional rights were violated should not face different out comes in a Section 2254 proceeding as a result of the state appellate courts making different errors in their cases, one with respect to whether there was a constitu tional error at all, and the other with respect to whether the error was harmless under Chapman. Petitioner's approach would also be difficult to administer in prac tice, because a state court will often not make clear what harmless-error standard it is applying. This case it self-in which the state appellate court found "no possi ble prejudice" (J.A. 97 n.17)-illustrates that point.
Because the court of appeals here correctly applied Brecht rather than Chapman and correctly concluded that the error in excluding Maples's testimony did not have a "substantial and injurious effect or influence in determining the jury's verdict," 507 U.S. at 623 (quoting Kotteakos v. United States, 328 U.S. 750, 776 (1946)), its judgment should be affirmed.
A STATE PRISONER IN A HABEAS PROCEEDING UNDER 28 U.S.C. 2254 MAY NOT OBTAIN RELIEF FOR A CONSTITU TIONAL TRIAL ERROR UNLESS IT HAD A SUBSTANTIAL AND INJURIOUS EFFECT ON THE JURY'S VERDICT
In Brecht v. Abrahamson, 507 U.S. 619 (1993), this Court held that a state prisoner seeking a writ of habeas corpus under 28 U.S.C. 2254 cannot obtain relief on the basis of a federal constitutional trial error unless that error "had substantial and injurious effect or influence in determining the jury's verdict." 507 U.S. at 631 (quoting Kotteakos v. United States, 328 U.S. 750, 776 (1946)). In adopting the Kotteakos standard, which the Court had established for review of non-constitutional errors on direct appeal from federal convictions, see id. at 764-765, the Brecht Court expressly rejected applying the harmless-error standard of Chapman v. California, 386 U.S. 18 (1967), under which a constitutional trial error triggers reversal of a conviction on direct appeal unless the error "was harmless beyond a reasonable doubt," id. at 24. The Brecht Court chose the "less oner ous" Kotteakos standard because "the costs of applying the Chapman standard on federal habeas outweigh the additional deterrent effect, if any, that would be derived from its application on collateral review." 507 U.S. at 636-637.
Following Brecht's teaching, the court of appeals here applied the Kotteakos standard and concluded that the constitutional error was harmless. J.A. 212. Peti tioner contends (Br. 24-27) that the court of appeals erred in applying that standard because in Brecht, un like here, the state appellate court applied Chapman on direct review, and Brecht's holding and rationale are limited to that circumstance. As we explain below, peti tioner's contention lacks merit.
A. Brecht Is Not Limited To Habeas Cases In Which A State Court Applied Chapman Review On Direct Appeal
In Brecht, the Court addressed whether the Seventh Circuit on collateral review correctly applied the Kott eakos harmless-error standard rather than the Chap man harmless-error standard to the violation of the peti tioner's due process right under Doyle v. Ohio, 426 U.S. 610 (1976), not to have his post-Miranda3 silence used against him at his trial. In deciding that the Seventh Circuit was correct to apply the Kotteakos standard, the Court explained that "[t]he imbalance of the costs and benefits of applying the Chapman harmless-error stan dard for collateral review counsels in favor of applying a less onerous standard on habeas review of constitu tional error." Brecht, 507 U.S. at 637. The Court fur ther explained that the "less onerous [Kotteakos] stan dard" was "better tailored to the nature and purpose of collateral review and more likely to promote the consid erations underlying our recent habeas cases." Id. at 637-638. Those observations prompted the Court to "hold that the Kotteakos harmless-error standard ap plies in determining whether habeas relief must be granted because of constitutional error of the trial type." Id. at 638 (footnote omitted).
The categorical nature of the Court's observations and holding refutes petitioner's assertion (Pet. Br. 24- 27) that the Court adopted a standard for collateral re view of constitutional error that would apply only when a state court on direct review recognized that error and reviewed it for harmlessness under Chapman's stan dard. The Brecht Court held out the possibility that the Kotteakos standard might not apply "in an unusual case" in which "a deliberate and especially egregious error of the trial type, or one that is combined with a pattern of prosecutorial misconduct, might so infect the integrity of the proceeding as to warrant the grant of habeas re lief, even if it did not substantially influence the jury's verdict." 507 U.S. at 638 n.9. That potential exception does not aid petitioner, however, because it is focused on the conduct of the trial and does not turn on the nature of the state courts' appellate review. In addition, peti tioner's proposed approach would limit Brecht's applica tion in a much more radical way that finds no support in the decision.
Petitioner cites for support (Pet. Br. 23) a single pas sage in Brecht:
State courts are fully qualified to identify constitu tional error and evaluate its prejudicial effect on the trial process under Chapman, and state courts often occupy a superior vantage point from which to evalu ate the effect of trial error. For these reasons, it sc arcely seems logical to require federal habeas courts to engage in the identical approach to harmless-error review that Chapman requires state courts to en gage in on direct review.
507 U.S. at 636 (citation omitted). That passage does not support petitioner's contention that Brecht applies only in cases in which a state court applied Chapman on direct review. The Court's observation about a state court's capacity to identify and evaluate constitutional errors is a general one that applies regardless of what specific determinations a state court actually makes in any given case. Indeed, nothing in the passage sug gests, let alone establishes, that a state court's applica tion of Chapman is a prerequisite to application of the Kotteakos standard on federal collateral review. Rath er, the point is that a federal court in a Section 2254 pro ceeding is not similarly situated to a state appellate court directly reviewing a state conviction and thus should not apply the harmless-error standard that the Constitution "requires" the state appellate court to ap ply, whether or not the state court in fact found any con stitutional error or, if it did, explicitly applied Chapman in the particular case.
The categorical nature of the majority's holding is confirmed by all of the Justices who wrote separately. Neither Justice Stevens, who authored a concurring opinion, nor the dissenting Justices read the majority opinion to be limited to cases in which Chapman review was applied by the state courts. See Brecht, 507 U.S. at 643 (Stevens, J., concurring) (referring to the "Kott eakos standard that will now apply on collateral review" of state convictions); id. at 644 (White, J., dissenting) ("If * * * the state courts erroneously concluded that no violation had occurred or (as is the case here) that it was harmless beyond a reasonable doubt, * * * the majority would foreclose relief on federal habeas re view" when the petitioner's conviction was "tainted by a constitutional violation that, while not harmless be yond a reasonable doubt, did not have 'substantial and injurious effect or influence in determining the jury's verdict.'" (emphasis added) (quoting Kotteakos, 328 U.S. at 776); id. at 651 (O'Connor, J., dissenting) (observing that, under the majority decision, Chapman "no longer applies to any trial error asserted on habeas").
Consistent with Brecht's clear holding, the Court's subsequent decisions have applied Brecht categorically. This Court's decision in Penry v. Johnson, 532 U.S. 782 (2001), is illustrative, confirming that the Brecht stan dard applies even when a state court does not conduct Chapman review. In Penry, the Texas Court of Crimi nal Appeals had rejected on the merits the petitioner's Fifth Amendment claim that the admission of a psychia trist's report violated his Fifth Amendment privilege against self-incrimination and thus had no occasion to conduct harmless-error review. Although the Court ruled first that the Texas court's rejection of the peti tioner's Fifth Amendment claim was not "objectively unreasonable," id. at 795, under 28 U.S.C. 2254(d)(1), the Court held in the alternative that, assuming a Fifth Amendment violation, "that error would justify over turning [the petitioner's] sentence only if [he] could es tablish that the error 'had substantial and injurious ef fect or influence in determining the jury's verdict,'" ibid. (quoting Brecht, 507 U.S. at 637 (quoting Kott eakos, 328 U.S. at 776)). The Court concluded that, un der Brecht, it could not disturb the state court's rejec tion of the Fifth Amendment claim, because it had "con siderable doubt" that admission of the report "had a 'sub stantial and injurious effect' on the verdict." Penry, 532 U.S. at 796 (quoting Brecht, 507 U.S. at 637). Penry thus makes clear that Brecht applies to cases in which the state appellate courts did not conduct Chapman review because they found no constitutional error.4
Calderon v. Coleman, 525 U.S. 141, 143-146 (1998) (per curiam), also indicates that whether the state appel late court applied Chapman has no bearing on the appli cability of the Kotteakos standard on collateral attack. There, the California Supreme Court found that a jury instruction violated state law but concluded that the er ror was not prejudicial, without specifying the harmless- error standard. Id. at 143; see People v. Coleman, 759 P.2d 1260, 1281-1282 (Cal. 1998). On habeas review, the Ninth Circuit held that the instructional error violated the federal Constitution and was not harmless under Boyde v. California, 494 U.S. 370, 380 (1990). This Court assumed that there was constitutional error and ruled that the Ninth Circuit erred in failing to apply Brecht's "substantial and injurious effect" standard in determining whether that error was harmless. Cole man, 525 U.S. at 145. Without inquiring into whether the state court's application of harmless-error principles to the state law violation amounted to or substituted for harmless-error review of the analogous federal claim under the Chapman standard, the Court explained that "once the [Ninth Circuit] determined that the giving of the [jury] instruction was constitutional error, it was bound to apply the harmless-error analysis mandated by Brecht." Id. at 146.5
B. The Brecht Court's Rationale For Rejecting Chapman's Standard Applies Regardless Of Whether The State Courts Found A Constitutional Violation Or Applied Chapman
The Brecht Court identified several reasons for re quiring federal courts on collateral review to apply a harmless-error standard "less onerous" (507 U.S. at 637) than Chapman, and none of them turns on the state court having found a constitutional violation and then having found it harmless under Chapman on direct re view. The Brecht Court rejected the petitioner's argu ment that applying the Chapman standard on collateral review was necessary to deter state courts from over looking constitutional errors and prosecutors from com mitting them. The Court concluded that "the costs of applying the Chapman standard on federal habeas out weigh the additional deterrent effect, if any, that would be derived from its application on collateral review." 507 U.S. at 636.
First, the Court concluded that applying Chapman would "undermine the States' interest in finality." Brecht, 507 U.S. at 637. That finality interest is equally robust when the state appellate courts have concluded that the petitioner's trial was free of constitutional error as when they have found constitutional error but con cluded that it was harmless under Chapman. In either case, the state courts have decided that the conviction should not be disturbed, and it is that ultimate decision, as opposed to a subsidiary appellate legal determination, that triggers the "presumption of finality" that "atta ches to the conviction and sentence." Id. at 633. Thus, the cost that collateral review imposes on the State's interest in finality is as high in the context presented here as it was in Brecht.6
Second, the Court found that applying Chapman would "infringe upon [the States'] sovereignty over criminal matters." Brecht, 507 U.S. at 637. Petitioner contends (Pet. Br. 25) that the "comity concerns" that arise when a federal court repeats the approach taken by the state courts "do not exist" when the state courts fail to apply Chapman. State court determinations con cerning the federal harmless-error standard do not lie any closer to the core of state sovereignty (or otherwise implicate the State's comity interests) than any other application of federal law by state courts. Contrary to petitioner's suggestion, the States have a sovereign in terest in protecting judgments of conviction obtained in their criminal courts, regardless of the specific determi nations that their appellate courts may have made in upholding them. When the Brecht Court referred to the States' "sovereignty over criminal matters," 507 U.S. at 637, it meant the States' sovereign interest in protecting "final and presumptively correct convictions" for viola tions of state criminal law, ibid. (emphasis added), not some discrete state appellate legal determination con cerning federal law. See id. at 635 ("Federal intrusions into state criminal trials frustrate both the States' sov ereign power to punish offenders and their good- faith attempts to honor constitutional rights.") (empha sis added) (quoting Engle v. Isaac, 456 U.S. 107, 128 (1982)).
Third, the Court in Brecht observed that applying Chapman "is at odds with the historic meaning of ha beas corpus-to afford relief to those whom society has 'grievously wronged,'" Brecht, 507 U.S. at 637, because Chapman affords relief "merely because there is a rea sonable possibility that trial error contributed to the verdict," ibid. (internal quotation marks omitted). In the Brecht Court's view, the "substantial and injurious effect" test better reflects the historical basis for award ing habeas relief because it requires "actual prejudice." Ibid. Here too, the likelihood of grievous wrong does not turn on whether the state appellate court found a consti tutional error that was harmless under Chapman or no constitutional error at all. Grievous wrong can arise in either situation, but only Brecht and Kotteakos, not Chapman, demand a showing of actual prejudice that can identify cases of grievous wrong.7
Finally, the Brecht Court concluded that applying Chapman on collateral review would increase the num ber of convictions that would be overturned and thereby "impose significant 'social costs,' including the expen diture of additional time and resources for all the parties involved, the 'erosion of memory' and 'dispersion of wit nesses' that accompany the passage of time and make obtaining convictions on retrial more difficult, and the frustration of 'society's interest in the prompt adminis tration of justice.'" Brecht, 507 U.S. at 637 (quoting United States v. Mechanik, 475 U.S. 66, 72 (1986)). Those social costs, of course, would be incurred regard less of the basis for the state appellate courts' refusal to undo the petitioner's conviction.8
In light of all of the foregoing costs of applying Chap man on collateral review, the Brecht Court concluded that "[t]he imbalance of the costs and benefits of apply ing the Chapman harmless-error standard on collateral review counsels in favor of applying a less onerous stan dard on habeas review of constitutional error." 507 U.S. at 637. A state court's failure to apply Chapman on di rect review does not have any effect on that balance, much less tip the scales in favor of applying the strin gent Chapman standard on collateral review.9
Petitioner contends (Pet. Br. 26) that if the Kott eakos standard applies regardless of whether the state courts have applied Chapman, then it would apply even if a State eliminated direct review altogether. But if a State were to do that, then a defendant would be de prived not just of the benefit of the Chapman standard, but of all of the other benefits of direct review that the Court has held to be unavailable on collateral review, including retroactive application of new rules, see Teague v. Lane, 489 U.S. 288 (1989), the right to coun sel, see Pennsylvania v. Finley, 481 U.S. 551 (1987), and review under the plain-error standard for unpreserved claims, see United States v. Frady, 456 U.S. 152 (1982). See Brecht, 507 U.S. at 634-635. If the possibility that a State would eliminate direct review justifies applica tion of Chapman on collateral review, so too would it justify the application of all the other rights and stan dards applicable on direct review, thereby erasing the distinction between direct and collateral review that "resounds throughout [the Court's] habeas jurispru dence." Id. at 633. It is thus not surprising that, with one exception, every circuit in the wake of Brecht that addressed Brecht's scope held that Brecht applies on collateral review regardless of whether the state courts applied Chapman review. See, e.g., Hassine v. Zimmer man, 160 F.3d 941, 951-953 (3d Cir. 1998) (Brecht ap plies regardless); Hogue v. Johnson, 131 F.3d 466, 499 (5th Cir. 1997) (same), cert. denied, 523 U.S. 1014 (1998); Sherman v. Smith, 89 F.3d 1134, 1140-1141 (4th Cir. 1996) (same); Brewer v. Reynolds, 51 F.3d 1519, 1529 (10th Cir. 1995) (same); Tyson v. Trigg, 50 F.3d 436, 446-447 (7th Cir. 1995) (same), cert. denied, 516 U.S. 1041 (1996); Horsley v. Alabama, 45 F.3d 1486, 1492 n.11 (11th Cir. 1995); but see Orndorff v. Lockhart, 998 F.2d 1426, 1430 (8th Cir. 1993) (Brecht applies only when the state courts applied Chapman). 10
C. Petitioner's Approach Would Produce Inequitable Re- sults And Would Be Difficult To Administer
Petitioner contends (Pet. 25) that applying Brecht in the absence of state court Chapman review "com pound[s] the state court's error," because it means that the state prisoner whose trial was beset by constitu tional error never gets the benefit of Chapman review. But, of course, Brecht already means that a habeas peti tioner can be denied relief even if he never had the bene fit of error-free Chapman review. Here again petitioner attributes too much significance to the reason the state court denied relief, as opposed to the fact that it denied relief. From a fairness standpoint, it is difficult to see why it would matter to the habeas petitioner whether the state court (1) erroneously failed to find the asserted constitutional error or (2) correctly recognized the error but mistakenly concluded that the error was harmless beyond a reasonable doubt under Chapman. Either way, the habeas petitioner will desire relief he should have obtained on direct review. Yet under petitioner's proposed approach, the prisoner who is denied relief in state court on the mistaken ground that a recognized constitutional error was harmless beyond a reasonable doubt will not obtain habeas relief if the error had no substantial and injurious effect on the verdict, while the prisoner who fails to convince the state courts that the constitutional error was made will obtain habeas relief if harm is merely reasonably possible under Chapman. The anomalous nature of that result illustrates the flaw in petitioner's pegging the appropriate harmless-error standard on collateral review to the specific basis on which the state appellate court denied relief on the pris oner's constitutional claim rather than to the state court's ultimate conclusion that the conviction should stand.
Another problem with petitioner's position is that state appellate courts frequently find trial error to be "harmless" or "nonprejudicial" without making clear the specific standard they have applied, as this case illus trates. That is particularly likely to occur in a situation like this when the harmless-error determination is an alternative holding. Because habeas courts would have to parse and interpret many different permutations of language under petitioner's approach, it would add a layer of complexity to collateral review by relying on a criterion-the actual use of Chapman on direct re view-that is unrelated to the interests of either the State or the prisoner in the Section 2254 proceeding. Cf. O'Neal v. McAninch, 513 U.S. 432, 443 (1995) (citing the "administrative virtues" of the rule the Court chose to adopt).
D. The Court Of Appeals Correctly Held That Brecht Ap plies And That The Error Did Not Have A "Substantial And Injurious Effect" On The Jury's Verdict
For the reasons stated above, the court of appeals correctly applied Brecht and used the Kotteakos stan dard in affirming the district court's decision denying petitioner habeas relief. J.A. 211-212. Petitioner was not entitled to habeas relief unless the error had a "sub stantial and injurious effect or influence in determining the jury's verdict." Brecht, 507 U.S. at 637 (quoting Kot teakos, 328 U.S. at 776).11
The error could not have had such an effect or influ ence on the jury's verdict in this case. First, the evi dence of petitioner's guilt was powerful. Numerous wit nesses testified that Cynthia Bell owed petitioner money for drugs and that petitioner had threatened to kill her in the days immediately preceding the murders. Several witnesses, including petitioner's brother, testified that petitioner admitted killing the Bells and described the killing of Cynthia in graphic detail. These witnesses also testified that when petitioner returned home on the night of the murders, he was spotted with blood. Expert testimony established that one of the bullets recovered from the victims had definitely been fired from peti tioner's gun. Petitioner also confessed to another in mate that he had committed the murders, and tried to arrange to have his brother injured in order to persuade him to change his testimony. J.A. 71.
Petitioner contends that many of these witnesses were unreliable or threatened by the police (Pet. Br. 5-6 n.2). Petitioner himself, however, made numerous ad missions that corroborate significant parts of the State's case. Petitioner admitted that at the time of the mur ders he owned a 1977 powder blue pickup truck with a white camper shell (which resembled the vehicle wit nesses described seeing near the victim's vehicle the night of the murders) and a .357 magnum. He admitted that Cynthia Bell dealt drugs for him and owed him money at the time she was murdered. Petitioner further admitted driving an acquaintance-who testified that petitioner believed she (the acquaintance) had stolen money from him and had threatened to kill her (J.A. 29)-into the country two days before the murders, tak ing out his gun and shooting it twice out the window. He admitted that when he arrived home on the night of the murders, he "pulled into the driveway and sat there for a while," "trying to maintain control" and "get a focus." J.A. 53. Petitioner admitted going into the bathroom upon entering the house, but denied changing his clothes or having blood on them. He admitted that he removed the camper shell from his pickup truck the morning af ter the murders, but claimed that he did so to save gas mileage. Petitioner also admitted lying to the police when asked when he had last seen his .357. Finally, pe titioner admitted that he gave an inmate (who testified that he had offered to injure petitioner's brother) his brother's address. J.A. 49-55. Although petitioner de nied committing the murders, his own testimony lent force to the State's witnesses.
Second, while Maples's testimony would have possi bly implicated a third party, Anthony Hurtz, in the mur ders, numerous other witnesses testified to the same effect and in "far more damning detail." J.A. 181. Peti tioner asserts that Maples would have been the only witness to implicate Hurtz who was not biased against him, but, as the magistrate judge found, cross-examina tion would have exposed "the infirmities of her putative testimony," ibid., including the fact that she only over heard "bits and pieces" of the conversation in which Hurtz allegedly implicated himself. J.A. 95.
Finally, petitioner relies heavily (Pet. Br. 30) on the fact that the jury deliberated in the case for many weeks before reaching a verdict. The jury ultimately rejected the defense suggestion that the testimony implicating Hurtz exonerated petitioner, however. Given the str ength of the State's case, petitioner's inculpating testi mony, the significant weaknesses in Maples's proffered testimony, and the fact that her testimony would not have put any new theory before the jury, the error in excluding her testimony did not have a "substantial and injurious effect or influence in determining the jury's verdict."
The judgment of the court of appeals should be affirmed.
PAUL D. CLEMENT
ALICE S. FISHER
Assistant Attorney General
MICHAEL R. DREEBEN
Deputy Solicitor General
JONATHAN L. MARCUS
Assistant to the Solicitor
JOEL M. GERSHOWITZ
1 See, e.g., United States v. Dago, 441 F.3d 1238, 1245-1246 (10th Cir. 2006); United States v. Owen, 407 F.3d 222, 229 (4th Cir. 2005), cert. denied, 126 S. Ct. 1026 (2006); United States v. Rivera, 347 F.3d 850, 852 (10th Cir. 2003), cert. denied, 540 U.S. 1210 (2004); United States v. Montalvo, 331 F.3d 1052, 1057-1058 (9th Cir. 2003), cert. denied, 541 U.S. 1011 (2004); Ross v. United States, 289 F.3d 677, 682 (11th Cir. 2002) (per curiam), cert. denied, 537 U.S. 1113 (2003); Santana-Madera v. United States, 260 F.3d 133, 140 (2d Cir. 2001), cert. denied, 534 U.S. 1083 (2002); Murr v. United States, 200 F.3d 895, 906 (6th Cir. 2000).
2 The State also presented testimony that petitioner sat in his truck for some time before entering the house and that once he came inside, petitioner went into the bathroom, from which the sound of running water was heard. J.A. 37.
3 Miranda v. Arizona, 384 U.S. 436 (1966).
4 In Mitchell v. Esparza, 540 U.S. 12 (2003) (per curiam), the Court held that, in cases where the state courts have in fact found a constitu tional error harmless, habeas relief may be obtained under the Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA), Pub. L. No. 104-132, 110 Stat. 1214, only when the state court's "conclusion [that the error was harmless] was * * * objectively unreasonable." Id. at 18. Two circuits have read Esparza to implicitly reject Brecht in circumstances in which the state courts have concluded that the error was harmless. See Eddleman v. McKee, 471 F.3d 576, 583 (6th Cir. 2006) ("AEDPA replaced the Brecht standard with the standard of Chapman plus AEDPA deference when * * * a state court made a harmless-error determination."); Gutierrez v. McGinnis, 389 F.3d 300, 306 (2d Cir. 2004) ("The Supreme Court [in Esparza] implicitly rejected Brecht as the proper lens for examining the harmlessness of constitu tional errors on collateral review, at least where the state explicitly adjudicated a federal claim on harmless error grounds."). Esparza, however, concluded that the state court's harmless-error decision was not "objectively unreasonable," 540 U.S. at 18, and the Court thus had no occasion to decide whether the Brecht test also had to be satisfied to warrant habeas relief. Other circuits have continued to require that a federal constitutional error meet the Brecht standard before granting habeas relief under AEDPA in cases where the state courts found the error harmless. See Inthavong v. LaMarque, 420 F.3d 1055, 1061 (9th Cir. 2005) ("[B]oth the Brecht and the AEDPA/Esparza tests must be satisfied with respect to harmless error before relief can be granted."), cert. denied, 126 S. Ct. 1660 (2006); Jones v. Polk, 401 F.3d 257, 264-266 (4th Cir. 2005) (finding that the state court unreasonably applied Chapman but that the error did not have a substantial and injurious effect on the jury's verdict). In all events, Brecht would continue to govern cases, like Penry, where the state courts did not reach the harmless-error issue and all proceedings under 28 U.S.C. 2255.
5 Similarly, in O'Neal v. McAninch, 513 U.S. 432 (1995), the Court applied the Kotteakos standard adopted in Brecht, even though the state appellate court, having found no constitutional error, did not apply Chapman review. See Resp. Br. at 4, O'Neal, supra (No. 93-7407) (noting that the state appeals court "reject[ed] [the] [p]etitioner's claims" and held "that the challenged instructions were in accord with Ohio law"). See also Tuggle v. Netherland, 516 U.S. 10, 15 (1995) (Scalia, J., concurring) (because the state appellate court failed to perform its task of "determining whether there was reasonable doubt as to whether the constitutional error contributed to the jury's de cision, * * * a remand is appropriate to allow the Fourth Circuit to review the case under the harmless-error standard appropriate to collateral review," citing Brecht); Duncan v. Henry, 513 U.S. 364, 371- 372 & n.1 (1995) (Stevens, J., dissenting) (stating, in a case in which the state court, on direct appeal, had applied a variant of the Kotteakos standard in reviewing a due process violation for harmless error, that "the Brecht standard would apply in the federal habeas proceeding"). The Court's more recent statements about Brecht reflect the same view. See United States v. Dominguez Benitez, 542 U.S. 74, 81 n.7 (2004) ("When the government has the burden of showing that constit utional trial error is harmless because it comes up on collateral review, the heightened interest in finality generally calls for the Government to meet the more lenient Kotteakos standard.") (citing Brecht); United States v. Vonn, 535 U.S. 55, 63 (2002) (citing Brecht as "the variant of harmless-error review applicable on collateral attack").
6 The California Court of Appeals held in the alternative that, even assuming the exclusion of Maples's testimony was an abuse of discre tion, "no possible prejudice could have occurred," because her testi mony was "merely cumulative." J.A. 97 n.17. Thus, the state appellate court did conclude, in the alternative, that any error was harmless. Had the court invoked the Chapman standard, it is unclear whether petitioner would view this case as subject to Brecht. See J.A. 212 n.3. It also is unclear why, even under petitioner's approach, a state court should have to invoke Chapman expressly when the finding that it made-i.e., "no possible prejudice"-is tantamount to a finding of harmlessness under Chapman's phraseology "harmless beyond a reasonable doubt." 386 U.S. at 24. See ibid. (observing that "[t]here is little, if any, difference between [the harmless beyond a reasonable doubt standard and] our statement in Fahy v. Connecticut about 'whether there is a reasonable possibility that the evidence complained of might have contributed to the conviction'"). See, e.g., Richardson v. Bowersox, 188 F.3d 973, 979 (8th Cir. 1999) (observing that Missouri Supreme Court's review of constitutional claim for prejudice and its conclusion that "the jury's verdict would have been no different if [the] out-of-court statement had been excluded" might qualify as Chapman review). Certainly, if the standard required by Chapman (or an even more defendant-protective standard) is, in fact, applied, it could be neither error nor a basis for rejecting Brecht for a court simply to omit a citation to Chapman.
7 In O'Neal, supra, the Court held that the necessary showing is made when "the record is so evenly balanced that a conscientious judge is in grave doubt" about whether a constitutional error "substantially influenced the jury's decision." 513 U.S. at 436-437.
8 Petitioner's amicus contends (Inn. Net. Br. 7) that AEDPA has "largely alleviated" the Brecht Court's concerns about the finality of state convictions because it "sets strict time limitations on habeas petitions." Petitioner's amicus underestimates the duration of state appellate and federal habeas proceedings. In this AEDPA proceeding, the jury found petitioner guilty on June 8, 1995, and the Ninth Circuit did not issue its decision affirming the denial of habeas relief until nearly eleven years later. See J.A. 1-2.
9 Petitioner mistakenly suggests (Pet. Br. 23 n.17) that California v. Roy, 519 U.S. 2 (1996) (per curiam), supports his position. There, the Court vacated a Ninth Circuit decision on collateral review that had applied an exacting special harmless-error standard to a jury instruc tion error. The Court concluded that the Ninth Circuit's decision was inconsistent with Brecht, which made clear that "the Kotteakos standard did apply to habeas review of what the Court called 'trial errors.'" Id. at 5. Although the Court mentioned that the state appellate court had found the error harmless under Chapman, see id. at 3, 5, the Court hardly "stress[ed]" (Pet. Br. 23 n.17) that fact, let alone tied application of the Kotteakos standard to it. Rather, the Court held that it followed Brecht and O'Neal (where the state appellate court did not conduct Chapman review, see note 5, supra) in holding that Kotteakos applied "for reasons related to the special function of habeas courts." 519 U.S. at 6.
10 In Section 2255 proceedings, the federal courts have applied Brecht regardless of whether Chapman was applied on direct review. See United States v. Dago, 441 F.3d 1238, 1246 (10th Cir. 2006) ("We have rejected the theory that the Brecht standard applies only when reviewing convictions that have previously been reviewed under the Chapman standard."); United States v. Montalvo, 331 F.3d 1052, 1057- 1058 (9th Cir. 2003) ("Brecht's harmless error standard applies to habeas cases under section 2255, just as it does to those under section 2254."), cert. denied, 541 U.S. 1011 (2004); Ross v. United States, 289 F.3d 677, 682 (11th Cir. 2002) (per curiam) (stating that it has "rejected the theory that the Brecht standard only applies to convictions which have been previously reviewed under the Chapman standard"), cert. denied, 537 U.S. 1113 (2003); Murr v. United States, 200 F.3d 895, 906 (6th Cir. 2000) (holding that Brecht applies in Section 2255 cases to "constitutional error that implicates trial procedures").
11 Because this case comes to the Court on the assumption that there was constitutional error, this Court need not address the standard that a defendant must meet to establish that a trial court's abuse of discretion in excluding a piece of evidence under a valid rule rises to the level of a due process or compulsory process violation.