Ruben Robert Herrera, a state prisoner, brought a petition for habeas corpus relief under 28 U.S.C. § 2254 claiming that his constitutional rights were violated by the admission at trial of illegally obtained evidence. The district court denied relief, holding that the state court's determination of harmless error was entitled to a presumption of correctness and, alternatively, that the error was harmless under Brecht v. Abrahamson, 507 U.S. 619, 113 S. Ct. 1710, 123 L. Ed. 2d 353 (1993). A panel of this court concluded the district court erred in presuming correct the state court's harmless error analysis and in reaching its own harmless error determination without reviewing the entire state court record. See Herrera v. Lemaster, 225 F.3d 1176 (10th Cir.2000). We directed the district court on remand to assess whether the constitutional error was harmless under Brecht after an independent review of the state court record. Mr. Herrera sought en banc review, arguing that the Brecht harmless error standard does not apply to a habeas action governed by AEDPA1 when, as here, the state court did not perform its harmless error analysis under Chapman v. California, 386 U.S. 18, 87 S. Ct. 824, 17 L. Ed. 2d 705 (1967).

We granted rehearing en banc and directed the parties to address the following question:

After AEDPA, when a state court decides a constitutional issue contrary to Supreme Court authority, or unreasonably applies Supreme Court authority, should the habeas court on collateral review apply the harmless error standard of Chapman v. California, 386 U.S. 18, 87 S. Ct. 824, 17 L. Ed. 2d 705 (1967), to assess the constitutional violation, or should it apply the standard set out in Brecht v. Abrahamson, 507 U.S. 619, 113 S. Ct. 1710, 123 L. Ed. 2d 353 (1993).

Herrera v. Lemaster, No. 98-2060 (10th Cir. Feb. 26, 2002) (order granting rehearing en banc). Upon en banc consideration, we hold that the standard set out in Brecht is the appropriate one to use in these circumstances.

* Mr. Herrera was convicted in New Mexico state court of first degree murder and aggravated assault with a firearm. He was sentenced to life in prison for the murder, eighteen months for the aggravated assault, and one year for the firearm enhancement. On direct appeal, Mr. Herrera argued that his residence was searched and evidence found there was seized under an invalid warrant, and that the admission of the seized evidence at his trial therefore violated his rights under the Fourth Amendment. Although the New Mexico Supreme Court held that the warrant was constitutionally deficient, it concluded the admission of the illegally seized items was harmless error under State v. Moore, 94 N.M. 503, 612 P.2d 1314, 1315 (1980). See State v. Herrera, 102 N.M. 254, 694 P.2d 510, 513-15 (1985).

Mr. Herrera subsequently filed this petition for federal habeas relief, arguing the state supreme court erred in holding that the admission of the illegally seized evidence was harmless. The matter was referred to a magistrate judge, who issued proposed findings and conclusions recommending that relief be denied. In so doing, the magistrate conducted no hearings, ordered no briefs, proceeded without access to the trial record, and issued his proposed disposition without prior notice to the parties. In reaching his decision, the magistrate gave a presumption of correctness to the state court's harmless error determination, and alternatively concluded the error was harmless under the standard set out in Brecht.

Mr. Herrera filed objections to the report, contending the state court's harmless error determination was entitled to no deference because that court had not performed its analysis under the standard mandated by the Supreme Court in Chapman.2 Mr. Herrera further argued that because application of the Brecht standard of review is predicated upon a state court evaluation of harmlessness under Chapman, when the state court failed to apply the correct Chapman analysis, as here, the federal habeas court should assess harmlessness under Chapman rather than Brecht. The state filed no response, and the district court adopted the magistrate's recommendation that the petition be dismissed with prejudice. Mr. Herrera appealed.

We granted Mr. Herrera a certificate of appealability on his claims that the admission of the unlawfully seized evidence was not harmless error and that the district court erred in denying habeas relief without reviewing the state court record. On appeal, Mr. Herrera argued that Brecht is no longer good law after AEDPA, and that even if Brecht survives it does not apply when the state court has not performed a harmless error analysis under Chapman. In so doing, Mr. Herrera contended that the state court's failure to apply Chapman was contrary to clearly established federal law as determined by the Supreme Court, and that the harmless error issue is therefore a legal question governed by 28 U.S.C. § 2254(d)(1),3 to which the presumption of correctness does not apply.

The panel opinion agreed with Mr. Herrera in part. It held that under Williams v. Taylor, 529 U.S. 362, 120 S. Ct. 1495, 146 L. Ed. 2d 389 (2000), a state court decision is contrary to clearly established federal law within the meaning of section 2254(d)(1) if the state court applies a rule that contradicts the governing law set forth in the Supreme Court's cases. See Herrera, 225 F.3d at 1178 (quoting Williams, 529 U.S. at 405, 120 S. Ct. 1495). The panel further agreed that the state court's failure to assess the harmlessness of the Fourth Amendment violation under Chapman was contrary to clearly established federal law under section 2254(d)(1).4 It also pointed out that the presumption of correctness provided in AEDPA by 28 U.S.C. § 2254(e)(1)5 applies only to state court fact findings, and that the harmless error inquiry is a mixed question of law and fact to which the presumption does not apply. Accordingly, it held that the district court erred in presuming correct the state court's harmless error determination. Although the panel also agreed with Mr. Herrera that the district court erred in concluding alternatively that the admission of the illegally seized evidence was harmless without reviewing the entire state court record, it did not accept the argument that a federal habeas court undertakes a harmless error review in these circumstances under the Chapman standard. The panel remanded with directions that the district court make a harmless error determination under Brecht upon review of the entire state court record.

Mr. Herrera petitioned for rehearing en banc, renewing his arguments that federal habeas courts should assess the harmlessness of a constitutional error under Chapman rather than Brecht when the state courts fail to do so, and that the justification for the Brecht standard has disappeared upon the enactment of AEDPA. We granted rehearing en banc to address these issues.6

Before we turn to the en banc question, we address briefly the state's argument that no Fourth Amendment violation occurred in this case because the illegally seized evidence was admissible under the good faith exception to the exclusionary rule established in United States v. Leon, 468 U.S. 897 (1984). Leon was decided after the trial in this case and while Mr. Herrera's direct appeal was pending. The state filed a supplemental brief with the New Mexico Supreme Court arguing that the case did not require consideration of the good faith exception set out in Leon but that it could provide an alternative ground for sustaining Mr. Herrera's convictions. See rec. vol. I, doc. 7, exh. H. The court concluded it need not discuss the issue in view of its determination that the admission of the illegally seized evidence was harmless. See Herrera, 694 P.2d at 514 n. 1.

The state did not specifically raise the applicability of Leon in federal district court, either by asserting the issue in its answer or by way of objection to the magistrate's report and recommendation. The state did urge the applicability of Leon to the panel on appeal as part of its contention that the district court did not err in applying Brecht. In so doing, the state made the barebones argument that the affidavit was sufficient to allow the issuing magistrate to conclude the residence to be scrutinized was Mr. Herrera's residence.

When an issue has not been properly raised below, we generally do not address it on appeal. See Proctor & Gamble Co. v. Haugen, 222 F.3d 1262, 1270-71 (10th Cir.2000). Although some courts have considered the applicability of the Leon exception to be a mixed question of fact and law that can be addressed for the first time on appeal, see United States v. Maggitt, 778 F.2d 1029, 1034 (5th Cir.1985) (citing cases); United States v. Sager, 743 F.2d 1261, 1265-66 (8th Cir.1984), we conclude that such consideration is not appropriate here.

For the first time in its brief to the en banc court, the state contended the affidavit was facially sufficient on the basis of evidence presented at the state court suppression hearing. We will not consider new assertions presented for the first time on rehearing en banc, particularly those based on factual allegations. Accordingly, we decline the state's invitation to address the applicability of Leon.

In addressing the continued viability and applicability of Brecht after the enactment of AEDPA, we look first to the text of the decision itself and compare the concerns underlying its holding with the purposes Congress intended to further by enacting AEDPA. In Brecht, the Court held that the harmless error standard set out in Chapman, 386 U.S. at 24, 87 S. Ct. 824, under which a conviction must be set aside unless the constitutional error was harmless beyond a reasonable doubt, does not apply to federal habeas review of state court convictions. Instead, the Court adopted the harmless error standard set out in Kotteakos v. United States, 328 U.S. 750, 776, 66 S. Ct. 1239, 90 L. Ed. 1557 (1946), under which relief is granted when the constitutional error had substantial and injurious effect or influence on the jury's verdict. The Court supported its conclusion that a less stringent harmless error standard should apply on collateral review by looking first "to the considerations underlying our habeas jurisprudence," pointing to the primacy of the principal that "collateral review is different from direct review." Brecht, 507 U.S. at 633, 113 S. Ct. 1710. The Court set out four concerns that are furthered by applying different standards on habeas than would be applied on direct review.

The reason most frequently advanced in our cases for distinguishing between direct and collateral review is the State's interest in the finality of convictions that have survived direct review within the state court system. We have also spoken of comity and federalism. The States possess primary authority for defining and enforcing the criminal law. In criminal trials they also hold the initial responsibility for vindicating constitutional rights. Federal intrusions into state criminal trials frustrate both the States' sovereign power to punish offenders and their good-faith attempts to honor constitutional rights. Finally, we have recognized that [l]iberal allowance of the writ ... degrades the prominence of the trial itself, and at the same time encourages habeas petitioners to relitigate their claims on collateral review.

Id. at 635, 113 S. Ct. 1710 (citations and quotations omitted). After assessing the impact of the Chapman and Kotteakos standards on these interests, the Court concluded that the latter standard is "better tailored to the nature and purpose of collateral review and more likely to promote the considerations underlying our recent habeas cases." Id. at 638, 113 S. Ct. 1710.

We note that while the Brecht standard is less stringent than that in Chapman, it is still "appropriately demanding." Brecht, 507 U.S. at 641, 113 S. Ct. 1710 (Stevens, J., concurring). Justice Stevens, whose concurring opinion made the fifth vote for the majority in Brecht, wrote separately to emphasize this point. Justice Stevens stressed that the standard adopted in Brecht "places the burden on prosecutors to explain why those errors were harmless;7 requires a habeas court to review the entire record de novo in determining whether the error influenced the jury's deliberations; and leaves considerable latitude for the exercise of judgment by federal courts." Id. at 640-41, 113 S. Ct. 1710. As Justice Stevens observed, while the Brecht standard is less stringent than that in Chapman, "[g]iven the critical importance of the faculty of judgment in administering either standard, however, that difference is less significant than it might seem...." Id. at 643, 113 S. Ct. 1710.

The standard in Brecht must also be construed and applied in light of the Court's subsequent decision in O'Neal v. McAninch, 513 U.S. 432, 115 S. Ct. 992, 130 L. Ed. 2d 947 (1995). There the Court held that when a federal habeas judge finds a constitutional error in a state court trial but the record is evenly balanced such that the judge is in grave doubt about whether the error is harmless, "the uncertain judge should treat the error, not as if it were harmless, but as if it affected the verdict (i.e., as if it had a `substantial and injurious effect or influence in determining the jury's verdict.')." Id. at 435, 115 S. Ct. 992.

The Court rested its legal conclusion on three considerations. Id. at 437, 115 S. Ct. 992. First, the Court observed that both Chapman and Kotteakos apply the rule, id. at 437-40, 115 S. Ct. 992, and that it properly takes into account the stakes involved in a habeas proceeding, in which a person's custody, not mere civil liability, is at issue, id. at 440, 115 S. Ct. 992. Second, the Court concluded that its rule is consistent with the purposes underlying habeas relief. Id. at 442. After balancing the interests in protecting against unconstitutional convictions and in assuring fundamentally fair trials against the interests in the finality of state court judgments and state-federal comity, the Court concluded that "[o]n balance, we must doubt that the law of habeas corpus would hold many people in prison `in violation of the Constitution,' for fear that otherwise a smaller number, not so held may eventually go free." Id. at 443, 115 S. Ct. 992. Finally the Court pointed out the administrative virtues of its rule, noting that it "is consistent with the way that courts have long treated important trial errors." Id. After weighing the above considerations, the Court held that "in cases of grave doubt as to harmlessness the petitioner must win." Id. at 437, 115 S. Ct. 992.

The concerns that prompted the Court in Brecht to adopt a less stringent standard for harmless error assessment on collateral review are the same ones that led Congress to pass AEDPA. See Calderon v. Thompson, 523 U.S. 538, 554-55, 118 S. Ct. 1489, 140 L. Ed. 2d 728 (1998) (citing Brecht as example of habeas corpus jurisprudence consistent with objects of AEDPA). "AEDPA's purpose [was] to further the principles of comity, finality, and federalism. There is no doubt Congress intended AEDPA to advance these doctrines." Williams v. Taylor, 529 U.S. 420, 436, 120 S. Ct. 1479, 146 L. Ed. 2d 435 (2000). In enacting AEDPA "Congress wished to curb delays, to prevent `retrials' on federal habeas, and to give effect to state convictions to the extent possible under law." Williams, 529 U.S. 362, 404, 120 S. Ct. 1495, 146 L. Ed. 2d 389. Accordingly, applying Brecht to cases governed by AEDPA would not run counter to the Act's purpose.

On rehearing en banc, Mr. Herrera argues that after AEDPA, he is entitled to habeas relief without regard to whether the standard in Brecht is met when, as here, the state court decision was contrary to clearly established law. In essence he argues that the prerequisites for federal habeas relief set out in section 2254(d) replace the Brecht inquiry, contending the Court's justifications in Brecht for using a less demanding harmless error standard have disappeared with the enactment of AEDPA. In Mr. Herrera's view, if a state court fails to apply Chapman on direct review, the decision is contrary to Supreme Court authority under section 2254(d)(1) and the federal habeas court must therefore apply the proper Chapman standard. If, on the other hand, the state court properly applies Chapman, the federal habeas court assesses only whether the state court's application was unreasonable under section 2254(d)(2). He thus urges that the Brecht inquiry should have no application in either scenario.

We are not persuaded Congress intended to alter the Supreme Court's pre-existing federal habeas jurisprudence on harmless error analysis. As we have discussed, Congress' purpose in enacting AEDPA is entirely congruent with the rationale set out by the Court in Brecht for adopting a less demanding harmless error standard on collateral review. Congress intended that AEDPA raise the bar with respect to availability of federal habeas relief, an intent that does not permit us to construe AEDPA to frustrate that intent by broadening the availability of habeas relief. Moreover, we "generally assume Congress knows the law and legislates in light of federal court precedent." Jurado-Gutierrez v. Greene, 190 F.3d 1135, 1146 (10th Cir.1999); see also Goodyear Atomic Corp. v. Miller, 486 U.S. 174, 184-85, 108 S. Ct. 1704, 100 L. Ed. 2d 158 (1988). We therefore assume Congress was aware of and acted in light of the harmless error standard set out in Brecht when it enacted AEDPA.

Mr. Herrera contends that applying the Brecht standard on habeas review would reduce the constitutional protection provided by established Supreme Court precedent and defeat the remedial purpose of habeas relief. The Court, however, has decided otherwise. As discussed above, although the Court narrowed the availability of habeas relief in Brecht, in O'Neal the Court ameliorated the effect of Brecht by ruling that when a habeas court is in grave doubt as to the harmlessness of a constitutional error, the petitioner is entitled to relief. O'Neal, 513 U.S. at 435, 115 S. Ct. 992. In so doing, the Court balanced comity, finality, and deference to state court criminal proceedings against the need to protect the fundamental fairness of criminal trials and avoid grievous wrongs. We are not at liberty to reweigh the balance struck by the Court in establishing the Brecht/O'Neal standard.

We also are not persuaded by Mr. Herrera's argument that because the state court in Brecht had assessed harmlessness under the proper Chapman analysis, Brecht is only applicable in those circumstances and does not govern habeas petitions where the state did not apply Chapman. The broad language and the analysis employed by both the plurality and by Justice Stevens in his concurrence imply that the standard set out was intended to govern all federal habeas review of state court decisions. See Brecht, 507 U.S. at 623, 113 S. Ct. 1710 (Rehnquist, J., writing for the Court) ("The Kotteakos harmless-error standard is better tailored to the nature and purpose of collateral review than the Chapman standard, and application of a less onerous harmless-error standard on habeas promotes the considerations underlying our habeas jurisprudence."); id. at 643, 113 S. Ct. 1710 (Stevens, J., concurring) ("The Kotteakos standard that will now apply on collateral review is less stringent than the [Chapman] standard applied on direct review.").

Moreover, the Supreme Court has subsequently clearly indicated that a federal habeas court is to apply the Brecht standard to a habeas petition governed by AEDPA even when the state court has not assessed a constitutional error for harmlessness under Chapman. See Penry v. Johnson, 532 U.S. 782, 121 S. Ct. 1910, 150 L. Ed. 2d 9 (2001). In Penry, a petitioner convicted of capital murder in state court asserted that evidence was admitted at trial in violation of his Fifth Amendment privilege against self-incrimination. The state court on direct review disagreed and concluded that no constitutional violation had taken place. The petitioner then brought a proceeding for federal habeas relief, alleging the state court decision was contrary to or involved an unreasonable application of federal law. Although the Supreme Court concluded otherwise, the Court alternatively indicated:

Even if our precedent were to establish squarely that the prosecution's use of the Peebles report violated Penry's Fifth Amendment privilege against self-incrimination, that error would justify overturning Penry's sentence only if Penry could establish that the error "had substantial and injurious effect or influence in determining the jury's verdict."

Id. at 795, 121 S. Ct. 1910 (quoting Brecht, 507 U.S. at 637, 113 S. Ct. 1710).

In Penry no harmless error analysis took place in state court because that court determined that no constitutional error had occurred. Id. at 789-91, 121 S. Ct. 1910. Here, likewise, no proper harmless error analysis took place in state court because that court did not apply the Chapman standard mandated by the Supreme Court in conducting its harmless error review. In both instances, however, the result was the same: the petitioner never received a harmless error analysis under the proper standard in state court. The Supreme Court in Penry nonetheless stated that the appropriate standard by which to gauge harmlessness on collateral review was that supplied by Brecht. Id. at 795, 121 S. Ct. 1910. Even though no party in Penry asserted that Brecht was an inappropriate standard, we are not inclined to disregard this clear signal from the Court that Brecht applies to an AEDPA case even when no proper harmless error assessment occurred in state court.

In sum, we conclude that in cases governed by AEDPA, the habeas court is to apply the harmless error standard set out in Brecht when a state court decides a constitutional issue contrary to controlling Supreme Court authority or unreasonably applies that authority. Accordingly, we vacate the district court's judgment and remand with directions to assess the harmlessness of the Fourth Amendment violation under the Brecht standard in light of the entire state court record. In all other respects the panel opinion is reinstated.

REVERSED and REMANDED for further proceedings consistent with this opinion.