Peter Ventura v. Attorney General, State of Florid

419 F.3d 1269, 2005 U.S. App. LEXIS 16605, 2005 WL 1869122

• Court: Court of Appeals for the Eleventh Circuit
• Decided: August 9, 2005
• Docket: 04-14564
• Status: Published

Opinion

MARCUS, Circuit Judge:

In this capital case, Peter Ventura has petitioned for federal habeas corpus relief on the ground that the state prosecutor’s knowing failure to correct a key government witness’s false testimony that he received no consideration in exchange for his testimony violated Ventura’s due process rights, as established in Giglio v. United States, 405 U.S. 150, 92 S.Ct. 763, 31 L.Ed.2d 104 (1972). The Supreme Court of Florida denied post-conviction relief, reasoning that Giglio’s materiality element was unsatisfied, since there was no reasonable likelihood that the false testimony could have affected the judgment of the jury.

After careful review of the record, we are convinced that the Florida court’s disposition of Ventura’s Giglio claim was neither contrary to nor an unreasonable application of clearly established federal law, and therefore we affirm the district court’s denial of Ventura’s petition.

I.

Peter Ventura was convicted by a Florida jury of the 1981 murder of Robert Clemente and sentenced to death. Clem-ente’s body was discovered on April 15, 1981, in a rural area of Volusia County off state route 44. The body was found in a truck bearing the logo of Clemente’s employer, Crow’s Bluff Marina. Clemente had been shot four times, beaten, and possibly stabbed.

Clemente’s former employer, Jerry Wright, had taken out a “key man” insurance policy on Clemente’s life, without Clemente’s knowledge, when Clemente worked at Wright’s tire store. Wright then hired Jack McDonald to murder Clemente in exchange for half of the insurance proceeds. McDonald, in turn, hired the petitioner, Ventura, to commit the murder.

On June 25, 1981, Ventura was arrested in Chicago and McDonald was arrested in Florida. Both were indicted for first-degree murder on June 30, 1981. Ventura was released on bond in Chicago while awaiting extradition, and failed to appear at his extradition hearing on August 18, 1981. He remained a fugitive until June 11, 1986, when he was apprehended in Austin, Texas. In the meantime, the state determined that it could not successfully prosecute McDonald without Ventura’s cooperation. McDonald had remained incarcerated without trial until the expiration of Florida’s speedy-trial deadline, which requires the state to commence trial within six months of indictment. Thereafter, by operation of Florida’s law, McDonald became immune from prosecution for Clem-ente’s murder. 1

After being released unconditionally by Florida, however, McDonald was indicted by a federal grand jury sitting in the Northern District of Illinois in 1983 on charges arising out of his involvement in a Chicago bank fraud scheme. He pled guilty and was sentenced to three consecutive five-year terms of imprisonment. After sentencing, McDonald was released on bond, and subsequently failed to report for his incarceration. He remained a fugitive until he was apprehended in Georgia some four years later on September 23, 1987.

On January 11, 1988, Ventura was brought to trial for Robert Clemente’s murder in the Circuit Court for the Seventh Judicial Circuit in Volusia County. McDonald, by then back in federal custody, testified against Ventura. Among other things, the state prosecutor, Stark, asked McDonald on direct examination, “Any promises been made to you concerning your testimony here?” to which McDonald replied, “None whatsoever.”

Then, on redirect, Stark asked McDonald, “And what is your motivation for testifying here today?” McDonald explained: “Well, I’m nearing sixty years of age. This is probably, undoubtedly, the most horrendous thing I have ever been involved in, and I think it is about time we cleared the air and it might give Mr. and Mrs. Clemente a little peace of mind knowing exactly what happened.”

Ventura’s counsel, Cass, followed up on this on reeross-examination, in this exchange with McDonald:

Q. You have just said that your motive for testifying is simply to clear the air, and bring the truth out; is that correct?

A. That’s correct.

Q. And for no other reason?

A. None.

Q. And because you’re in a position where it doesn’t really cost you anything to say it?

A. That’s correct.

A jury convicted Ventura of Clemente’s murder and then recommended the death penalty. The trial judge adopted the jury’s recommendation and sentenced Ven-tura to die on January 21, 1988. The Supreme Court of Florida affirmed both Ventura’s conviction and the death sentence on direct review. Ventura v. State, 560 So.2d 217 (Fla.1990).

Thereafter, it came to light that McDonald had, in fact, been testifying pursuant to a deal with the prosecution. A series of letters document this arrangement. First, on December 19,1986 — while McDonald was still a fugitive — Stark wrote the U.S. Attorney’s Office in Chicago (the office that prosecuted the bank-fraud case against McDonald) explaining the Clem-ente case and asking for assistance in securing McDonald’s testimony. Stark’s letter stated, in part:

While I understand the obvious reluctance of the sentencing court to show lienience [sic] upon Mr. McDonald especially in light of the fact that he took advantage of the court’s prior lieniency [sic] by failing to report to prison after having been given time to get his affairs in order, I feel that the interests of justice could be better served by having Mr. McDonald on lengthy probation with a short jail term if necessary, available to testify at the trial of Peter Ven-tura and possibly Jerry Wright (in the event he is indicted).

I would appreciate any consideration your office could give in the effort to locate Jack McDonald, or coax him out of hiding.

On March 6, 1987, the U.S. Attorney replied to Stark, explaining that it was too late to modify McDonald’s sentence in the bank fraud case, but that if McDonald testified against Ventura, “his cooperation and truthful testimony in that case can be made known to the Federal Parole Board at his first parole hearing.” The letter further stated: “Should Mr. McDonald surrender to federal authorities and also appear as a witness at Mr. Ventura’s trial, this office will consider the nature of Mr. McDonald’s cooperation and truthful testimony in evaluating whether to pursue further prosecution of Mr. McDonald on bond jumping charges.” No promise was made, however, at that time concerning the prosecution of McDonald on federal bond-jumping charges.

Several months later, on September 8, 1987, McDonald wrote a letter to one of the investigative agents, Detective Hudson of the Volusia County Sheriffs Office, stating:

If anyone is interested in my testimony at this point in time it will have to be a two for one trade. In other words I will cooperate fully provided I am released by court order from all federal charges including the IRS.... If you can get a court order to this effect out in front I will promptly turn myself in and cooperate fully.... Should it be decided that the U.S. Gov’t and the State of Florida can handle this I will give you whatever time is necessary to clear this up. But again, I can only do it as a free man. If by some fluke I am apprehended without any deal being made I will rot in hell before I would give any testimony on anything. This is a promise.

Soon thereafter, on September 23, McDonald was taken into federal custody. Two days later, Stark again wrote the U.S. Attorney’s Office, in an effort to secure a deal for McDonald on possible federal bond-jumping charges in exchange for his testimony at the Ventura trial. This letter stated, in pertinent part:

Pursuant to our telephone conversation of today’s date, I would like to formally request that you consider dismissing the bond jumping charges against Jack McDonald .... On September 24th, David Hudson, the lead investigator on the Ventura case[,] interviewed Mr. McDonald and was assured of McDonald’s cooperation with us in the prosecution of Peter Ventura and others involved in the murder of one Robert Clemente. Needless to say, Mr. McDonald is a crucial witness in both cases. The case of the

State of Florida vs. Peter Ventura is presently scheduled for trial in Circuit Court, Volusia County on October 12, 1987. Mr. McDonald’s cooperation is essential.

The U.S. Attorney then agreed not to prosecute McDonald for bond jumping. 2 An October 5, 1987 letter from the U.S. Attorney to Stark stated:

Pursuant to your request, my office will not pursue bond-jumping charges against Jack McDonald as long as he cooperates fully with your office in the upcoming murder case referred to in your letter of September 25, 1987. Should Mr. McDonald fail to testify truthfully in that case or in some other way fail to cooperate with your office, we will then be free to pursue bond-jumping charges.

Moreover, this agreement does not affect Mr. McDonald’s obligation to serve the federal sentence which has been imposed for his prior criminal conduct in this district.

Immediately after Ventura’s trial, on January 20, 1988, Stark wrote another letter to the U.S. Attorney’s Office explaining that “[wjhile there were no promises made to Mr. McDonald in return for his testimony,” he “fe[lt] a compelling obligation to advise [the U.S. Attorney’s Office] and the Court in Chicago of the assistance provided by Mr. McDonald to the State of Florida in the prosecution of those persons involved in the homicide of Robert G. Clemente.” Among other things, Stark said:

While I realize that Jack McDonald is equally responsible for the death of Robert Clemente, along with his two co-defendants, I also recognize the fact that he did not have to provide information to the State of Florida detailing each player’s participation in the overall scheme to commit murder. He has testified in Mr. Ventura’s trial and I anticipate calling him in Mr. Wright’s trial. He has previously stated in deposition and at trial that his motive for giving his testimony has been to clear the air and set the record straight so that the family of the victim can have some peace of mind. He has also stated that “this is the most horrendous crime he has ever been involved with.”

I believe I understand Mr. McDonald’s motivation to testify and I further believe that his testimony has been extremely valuable in the prosecution of his co-defendants. Whatever consideration can be given him at any future hearings in his two Federal cases in return for this assistance would, in my opinion, be in the interest of justice.

Corporal David Hudson of the Volusia County Sheriffs Office and I would appreciate the courtesy of a telephone call regarding the scheduling of any future hearings to be held for Mr. McDonald so that we can make arrangements to be heard by the Court considering Mr. McDonald’s cooperation in Florida.

Finally, Stark wrote the U.S. Attorney’s Office again on October 31, 1988, explaining that McDonald was cooperating in the case against Jerry Wright, and that “[a]ny considerations [sic] that the Federal Courts could show Mr. McDonald for his efforts in this regard for his cooperation to date and in the future would be appreciated.” No record evidence indicates that the prosecutor at any time communicated McDonald’s cooperation either to the federal district court in Chicago or to the United States Parole Commission. McDonald was not indicted for federal bond jumping.

On March 2, 1992, Ventura filed his first motion for postconviction relief in the state trial court pursuant to Florida Rule of Criminal Procedure 3.850. Ventura claimed that he was unable to file a proper postconviction motion because several agencies had not satisfied his public records requests, and he listed the claims he intended to raise once the requests were fulfilled. The trial court dismissed all of Ventura’s claims, but the Supreme Court of Florida reversed the dismissal as premature, directing the trial court to permit Ventura to amend his postconviction motion after the public records issues were resolved. Ventura v. State, 673 So.2d 479 (Fla.1996).

On August 19, 1996, Ventura filed an amended postconviction motion in the state trial court, this time raising fifteen claims, including a Giglio claim alleging, based on the series of letters described above, that prosecutor Stark knowingly offered false testimony that McDonald had received no consideration in exchange for his trial testimony.

The trial court summarily dismissed ten of Ventura’s claims and, on June 1, 1998, held an evidentiary hearing on the remaining claims, including the Giglio issue. At this hearing, Ventura’s counsel introduced the above-described letters, and attorneys Stark and Cass both testified. Stark offered no explanation for allowing the false testimony, claiming to have little recollection of the relevant events. He could not definitively state whether the letters were disclosed to Ventura. He said that his office maintained an “open-file policy,” which made their entire file available to the defense, but could not recall whether the letters in question were part of the file.

Defense counsel Cass testified that he did not recall ever being told about a deal for McDonald in exchange for his testimony, that he was unaware of any communications between the state attorney and the U.S. Attorney’s Office, and that he remembered the frustration of not being able to impeach McDonald based on his motive for testifying.

Following the hearing, the trial court denied Ventura’s remaining claims, including his Giglio claim, with little explanation. Order, July 28, 1998. The trial court apparently interpreted the Giglio claim as an ineffective assistance of counsel claim, rejecting it on the ground that Ventura failed to demonstrate a reasonable probability of a different result had counsel performed adequately. Id. at 1-2. In reaching this conclusion, the trial court lumped the Giglio claim together with several others, finding that “the other issues are not worthy of any specific discussion.” Id. at 5.

The Supreme Court of Florida, by vote of 4-8, affirmed the denial of postconviction relief. Ventura v. State, 794 So.2d 553 (Fla.2001). As to Ventura’s Giglio claim, the supreme court found that the state had indeed secured a deal for McDonald in exchange for his testimony, and that the prosecutor knowingly presented false testimony to the contrary, but that the testimony was not material. Because McDonald had been significantly impeached at trial, and because McDonald’s testimony was extensively corroborated by other pieces of compelling evidence, the Florida Supreme Court reasoned, evidence of the deal was not “material” under Gig-lio. Id. at 562-65. The United States Supreme Court subsequently denied cert. Ventura v. Florida, 535 U.S. 1098, 122 S.Ct. 2296, 152 L.Ed.2d 1054 (2002).

Ventura then petitioned for a writ of habeas corpus in the United States District Court for the Middle District of Florida. After careful review, the district court denied the petition. Order, Ventura v. Moore, No. 6:02-cv-1159-Orl-19KRS (M.D.Fla. July 30, 2004). The district court expressed “grave doubts” about whether the state supreme court had applied the correct materiality standard in evaluating Ventura’s Giglio claim, since the state court had asked whether there was “a reasonable probability that the false evidence may have affected the judgment of the jury,” id. at 18 (quoting Ventu-ra, 794 So.2d at 563), rather than whether “the false testimony could ... in any reasonable likelihood have affected the judgment of the jury,” id. (quoting Giglio, 405 U.S. at 154, 92 S.Ct. 763). Nevertheless, although the state court had “misstated the Giglio rule,” the district court found that “the decision itself’ was neither contrary to nor an unreasonable application of clearly established federal law. Id. at 20. Applying Giglio’s “any reasonable likelihood” standard, the district court agreed with the state supreme court’s conclusion that the substantial impeachment of McDonald at trial, as well as the extensive corroboration of his testimony, established that “presentation of the correct facts could not have altered the outcome of the trial,” and thus that the false testimony was immaterial. Id. at 21.

On October 28, 2004, the district court granted a certificate of appealability only as to Ventura’s Giglio claim. That claim is the only one now before us.

II.

Giglio error is a species of Brady error that occurs when “the undisclosed evidence demonstrates that the prosecution’s case included perjured testimony and that the prosecution knew, or should have known, of the perjury.” United States v. Agurs, 427 U.S. 97, 103, 96 S.Ct. 2392, 49 L.Ed.2d 342 (1976). “If false testimony surfaces during a trial and the government has knowledge of it, ... the government has a duty to step forward and disclose.” Brown v. Wainwright, 785 F.2d 1457, 1464 (11th Cir.1986). “In order to prevail on a Giglio claim, a petitioner must establish that the prosecutor knowingly used perjured testimony, or failed to correct what he subsequently learned was false testimony, and that the falsehood was material.” Tompkins v. Moore, 193 F.3d 1327, 1339 (11th Cir.1999).

The origins of the Giglio doctrine lie in the Supreme Court’s decision in Napue v. Illinois, 360 U.S. 264, 79 S.Ct. 1173, 3 L.Ed.2d 1217 (1959), which held that a prosecutor’s failure to correct false testimony by the principal state witness that he had received no promise of consideration in return for his testimony violated the defendant’s Fourteenth Amendment due process rights and required a reversal of the judgment of conviction. The Court explained that “it is established that a conviction obtained through use of false evidence, known to be such by representatives of the State, must fall under the Fourteenth Amendment.” Id. at 269, 79 S.Ct. 1173 (citing Mooney v. Holohan, 294 U.S. 103, 55 S.Ct. 340, 79 L.Ed. 791 (1935)). “The same result obtains when the State, although not soliciting false evidence, allows it to go uncorreeted when it appears.” Id. This principle, the Court observed, “does not cease to apply merely because the false testimony goes only to the credibility of the witness,” since “[t]he jury’s estimate of the truthfulness and reliability of a given witness may well be determinative of guilt or innocence, and it is upon such subtle factors as the possible interest of the witness in testifying falsely that a defendant’s life or liberty may depend.” Id. Reversal was required because “the false testimony used by the State in securing the conviction may have had an effect on the outcome of the trial.” Id. at 272, 79 S.Ct. 1173.

■ Subsequently, in Giglio v. United States, 405 U.S. 150, 92 S.Ct. 763, 31 L.Ed.2d 104 (1972), the Supreme Court held that the government’s failure to correct false testimony that its key witness (the defendant’s coconspirator) had received no promise of nonprosecution in exchange for his testimony, as well as the prosecutor’s false statement to this effect in closing argument, required that the defendant be granted a new trial. The Court explained that “deliberate deception of a court and jurors by the presentation of known false evidence is incompatible with rudimentary demands of justice.” Id. at 153, 92 S.Ct. 763 (citation and internal quotation marks omitted).

The Giglio Court made clear, however, that such errors do not require automatic reversal, and articulated a “materiality” standard to guide the determination of whether a new trial is warranted:

We do not ... automatically require a new trial whenever a combing of the prosecutors’ files after the trial has disclosed evidence possibly useful to the defense but not likely to have changed the verdict. A finding of materiality of the evidence is required under Brady. A new trial is required if “the false testimony could ... in any reasonable likelihood have affected the judgment of the jury.”

Id. (citations and internal quotation marks omitted) (quoting Napue, 360 U.S. at 271, 79 S.Ct. 1173). Because “the Government’s case depended almost entirely on [the falsely testifying witness’s] testimony,” the Court reasoned, his “credibility as a witness was therefore an important issue in the case, and evidence of any understanding or agreement as to a future prosecution would be relevant to his credibility and the jury was entitled to know of it.” Id. at 154-55, 92 S.Ct. 763. Accordingly, the Court reversed the judgment of conviction.

Since its decisions in Napue and Giglio, the Supreme Court “has consistently held that a conviction obtained by the knowing use of perjured testimony is fundamentally unfair, and must be set aside if there is any reasonable likelihood that the false testimony could have affected the judgment of the jury.” United States v. Agurs, 427 U.S. 97, 103, 96 S.Ct. 2392, 2397, 49 L.Ed.2d 342 (1976) (footnote omitted) (emphasis added); see also Kyles v. Whitley, 514 U.S. 419, 433 & n. 7, 115 S.Ct. 1555, 131 L.Ed.2d 490 (1995); United States v. Bagley, 473 U.S. 667, 677, 105 S.Ct. 3375, 87 L.Ed.2d 481 (1985); accord Brown, 785 F.2d at 1465-66. 3

The “any reasonable likelihood” standard differs from the materiality standard applicable to other types of Brady violations because of the nature of the error. As the Supreme Court has explained, “the Court has applied a strict standard of materiality [to Giglio violations], not just because they involve prosecutorial misconduct, but more importantly because they involve a corruption of the truth-seeking function of the trial process.” Agurs, 427 U.S. at 104, 96 S.Ct. 2392; accord United States v. Alzate, 47 F.3d 1103, 1110 (11th Cir.1995).

As we have explained, Ventura’s Giglio claim arises out of his post-trial discovery that the state prosecutor had worked out a deal with the U.S. Attorney’s Office to immunize McDonald from prosecution on federal bond-jumping charges in exchange for his testimony at Ventura’s trial, and that McDonald’s testimony that he had received no promises in exchange for his testimony was therefore false. The State does not contest the Florida Supreme Court’s factual determination that this deal indeed existed, and that McDonald’s testimony to the contrary was false. See Ven-tura, 794 So.2d at 562. The only issue the parties have briefed and the only one now before us is whether the error was “material.”

III.

We review the district court’s denial of Ventura’s habeas petition de novo. Harrell v. Butterworth, 251 F.3d 926, 930 (11th Cir.2001). However, under the Anti-terrorism and Effective Death Penalty Act (“AEDPA”), which governs Ventura’s petition, we may grant the writ only if we find that the Florida Supreme Court’s resolution of Ventura’s Giglio claim “resulted in a decision that was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States.” 28 U.S.C. § 2254(d)(1).

A.

The phrase “clearly established Federal law,” as used in § 2254(d)(1), encompasses only the holdings of the United States Supreme Court. Williams v. Taylor, 529 U.S. 362, 412, 120 S.Ct. 1495, 146 L.Ed.2d 389 (2000). Section 2254(d)(1) ex pressly “restricts the source of clearly established law to [the Supreme] Court’s jurisprudence.” Id. As we have previously explained, “§ 2254(d)(1) provides a measuring stick for federal habeas courts reviewing state court decisions. That measuring stick is ‘clearly established Federal law.’ Clearly established federal law is not the case law of the lower federal courts, including this Court. Instead, in the habeas context, clearly established federal law ‘refers to the holdings, as opposed to the dicta, of [the Supreme Court’s] decisions as of the time of the relevant state court decision.’ ” Putman v. Head, 268 F.3d 1223, 1241 (11th Cir.2001) (alteration in original) (quoting Williams, 529 U.S. at 362, 120 S.Ct. 1495).

As far as Giglio materiality is concerned, the clearly established law of the Supreme Court is simply that reversal of a conviction is required when “there is any reasonable likelihood that the false testimony could have affected the judgment of the jury.” The Supreme Court has repeated this standard numerous times, but has not elaborated further on its meaning; indeed, no Supreme Court case since Giglio itself has squarely addressed a Giglio claim. Accordingly, the proper inquiry for our Court is whether the Florida Supreme Court’s treatment of Ventura’s Giglio claim was contrary to or an unreasonable application of the “any reasonable likelihood” standard. 4

B.

The “contrary to” and “unreasonable application” clauses of § 2254(d)(1) are separate bases for reviewing a state court decision. Williams, 529 U.S. at 412-13, 120 S.Ct. 1495; see also Putman, 268 F.3d at 1241. “Under the ‘contrary to’ clause, a federal habeas court may grant the writ if the state court arrives at a conclusion opposite to that reached by [the Supreme] Court on a question of law or if the state court decides a case differently than this Court has on a set of materially indistinguishable facts.” Williams, 529 U.S. at 412-13, 120 S.Ct. 1495; see also Bell v. Cone, — U.S.-, 125 S.Ct. 847, 851, 160 L.Ed.2d 881 (2005) (“A state court’s decision is ‘contrary to ... clearly established Federal law’ if the state court applies a rule that contradicts the governing law set forth in our cases, or if the state court confronts facts that are materially indistinguishable from a relevant Supreme Court precedent and arrives at a result opposite to ours.” (citation and internal quotation marks omitted)).

Thus, there are two possible bases for concluding that the Florida Supreme Court’s resolution of Ventura’s Giglio claim was contrary to clearly established federal law: either (1) it reached a conclusion of law that contradicts one reached by the Supreme Court; or (2) it reached a result opposite one reached by the Supreme Court on a set of materially indistinguishable facts.

The second possibility is more easily dismissed. Whether a Giglio violation is material is a highly fact-dependent inquiry, and we can find no Supreme Court case whose facts may fairly be characterized as “materially indistinguishable” from those of Ventura’s case. Whether the false testimony offered in a particular case could in any reasonable likelihood have affected the judgment of the jury must be analyzed in light of a number of highly context-specific factual considerations, including the importance of the testimony of the falsely testifying witness to the government’s case, the nature and significance of the falsehood, and, notably, to what extent the witness’s testimony is substantially corroborated by other evidence.

The context-specific nature of this inquiry is evident from the Supreme Court’s analysis in Giglio and Napue — the only two cases in which the Supreme Court has squarely addressed a Giglio claim — as well

as from our own analysis in Giglio eases. In Giglio, the Supreme Court found reversal of the defendant’s conviction required because “the Government’s case depended almost entirely on [the falsely testifying witness’s] testimony; without it there could have been no indictment and no evidence to carry the case to the jury.” Giglio, 405 U.S. at 154, 92 S.Ct. 763. In Napue, the Court again based its conclusion that reversal was required on its “own evaluation of the record” before it. Na-pue, 360 U.S. at 272, 79 S.Ct. 1173. Likewise, all of our cases reversing convictions on the basis of Giglio error — even those involving false testimony that a witness did not cut a deal in exchange for his testimony — have conducted careful analyses of the trial record to render a case-specific conclusion as to whether the violation might have made a difference. See, e.g., United States v. Rivera Pedin, 861 F.2d 1522 (11th Cir.1988); Haber v. Wainwright, 756 F.2d 1520 (11th Cir.1985); DeMarco v. United States, 928 F.2d 1074 (11th Cir.1991); United States v. Sanftlippo, 564 F.2d 176 (5th Cir.1977).

Neither the Supreme Court nor our Court has ever adopted a per se rule that the government’s failure to correct a witness’s false testimony that he did not receive any consideration in exchange for his testimony requires reversal of the conviction. We can find no case on all fours with Ventura’s, and therefore cannot conclude that the Florida Supreme Court’s treatment of Ventura’s Giglio claim was “contrary to” clearly established federal law in the sense of reaching an opposite result on materially indistinguishable facts.

The second inquiry — whether the state court reached a conclusion of law that contradicts one reached by the Supreme Court — requires a careful analysis of the Florida Supreme Court’s decision. In analyzing Ventura’s Giglio claim, the Florida court began its analysis by correctly noting: “To establish a violation of Giglio Ventura must show: ‘(1) that the testimony was false; (2) that the prosecutor knew the testimony was false; and (3) that the statement was material.’ ” Ventura, 794 So.2d at 562 (quoting Robinson v. State, 707 So.2d 688, 693 (Fla.1998)). The court also accurately explained “that ‘[t]he thrust of Giglio and its progeny has been to ensure that the jury know the facts that might motivate a witness in giving testimony, and the prosecutor not fraudulently conceal such facts from the jury.’ ” Id. (quoting Robinson, 707 So.2d at 693).

However, the Florida Supreme Court’s phrasing of Giglio’s materiality standard deviates from the formulation favored by the United States Supreme Court. As the Florida court explained it: “Under Giglio, a statement is material if ‘there is a reasonable probability that the false evidence may have affected the judgment of the jury.’ ” Id. (quoting Routly v. State, 590 So.2d 397, 400 (Fla.1991)). Routly cites Giglio and Napue, but nevertheless formulates the standard in this altered manner. The Florida Supreme Court’s formulation differs from Giglio’s in three respects: it uses “a” instead of “any”; “probability” instead of “likelihood”; and “may” instead of “could.”

The last distinction is immaterial, since “may” and “could” are both words used to express “possibility.” Oxford English Dictionary (2d ed.1989). Indeed, “can” — the present tense of “could” — is “sometimes used interchangeably with may.” Merriam-Webster’s Collegiate Dictionary (10th ed.1994); accord Oxford English Dictionary (2d ed.1989) (listing “may” as a synonym of “can”). Nor do we see a significant difference between “a” and “any.” “Any” is simply an “indeterminate derivative” of “a” or “an,” “in which the idea of unity ... is subordinated to that of indifference as to the particular one or ones that may be selected.” Oxford English Dictionary (2d ed.1989). For purposes of articulating the Giglio standard, this distinction is of no consequence.

The key issue is the Florida court’s use of the term “probability” instead of “likelihood.” After careful review, we conclude that because the court proceeded to describe the Giglio standard accurately, and because AEDPA plainly requires us to give state courts the benefit of the doubt on habeas review, the Florida court’s single use of the phrase “a reasonable probability” does not render its decision contrary to clearly established Supreme Court law. Insofar as the district court suggested that the Florida Supreme Court erroneously conflated Giglio’s standard with Brady’s, we disagree, since the Florida court understood the critical difference between the two and correctly applied the more lenient Giglio standard.

The Florida Supreme Court explicitly acknowledged the essential difference between the Giglio standard that applied to Ventura’s claim and the “reasonable probability” standard that governs more general Brady claims — namely, that Giglio sets the materiality bar lower than Brady. The Florida Supreme Court observed the distinction this way:

In denying Ventura’s claim, the trial court incorrectly relied on the materiality standard appropriate to Brady claims. See United States v. Alzate, 47 F.3d 1103, 1109-10 (11th Cir.1995) (“Where there has been a suppression of favorable evidence in violation of Brady v. Maryland, the nondisclosed evidence is material: ‘if there is a reasonable probability that, had the evidence been disclosed to the defense, the result of the proceeding would have been different.’ ... A different and more defense-friendly standard of materiality applies where the prosecutor knowingly used perjured testimony, or failed to correct what he subsequently learned was false testimony.”) (citations omitted).

Ventura, 794 So.2d at 563.

Indeed, the Alzate case on which the Florida court relied recites the Giglio standard in exactly the terms the United States Supreme Court has repeatedly used:

[T]he falsehood is deemed to be material “if there is any reasonable likelihood that the false testimony could have affected the judgment of the jury.” United States v. Agurs, 427 U.S. 97, 103, 96 S.Ct. 2392, 2397, 49 L.Ed.2d 342 (1976) (emphasis added); accord Giglio v. United States, 405 U.S. 150, 154, 92 S.Ct. 763, 766, 31 L.Ed.2d 104 (1972); Nannie v. Illinois, 360 U.S. 264, 271, 79 S.Ct. 1173, 1178, 3 L.Ed.2d 1217 (1959).

Alzate, 47 F.3d at 1110. 5

The Florida Supreme Court concluded that “based on this record of ample impeachment and corroboration, we hold the evidence of the deal immaterial under Gig-lio.” Ventura, 794 So.2d at 565. In reaching this result, the court cited two cases — its own decisions in Routly v. State, 590 So.2d 397 (Fla.1991), and White v. State, 729 So.2d 909 (Fla.1999) — that recited the standard in terms of a “reasonable probability,” but also another case — Unit ed States v. Petrillo, 821 F.2d 85 (2d Cir.1987) — -that used the well-established “any reasonable likelihood” formulation. Petril-lo explained: “Where the government uses perjured testimony in obtaining a conviction, and knew or should have known of the perjury, the defendant’s due process right to a fair trial is violated ‘if there is any reasonable likelihood that the false testimony could have affected the judgment of the jury.’ ” Id. at 87 (quoting Agurs, 427 U.S. at 103, 96 S.Ct. 2392).

Against this backdrop, we cannot say that the state court’s imprecision in explaining the Giglio standard for materiality means that it “arrive[d] at a conclusion opposite to that reached by [the Supreme] Court on a question of law,” Williams, 529 U.S. at 412, 120 S.Ct. 1495, or — as this part of the “contrary to” standard has alternatively been stated — that it “ap-plie[d] a rule that contradicts the governing law set forth in [Supreme Court] cases,” Bell, 125 S.Ct. at 851.

To be sure, a “reasonable probability” standard differs from a “reasonable likelihood” standard. See, e.g., Alzate, 47 F.3d at 1110 n. 7 (noting that “the district court applied the ‘reasonable probability of a different result’ standard,” which “is substantially more difficult for a defendant to meet than the ‘could have affected’ standard we apply”); Stephens v. Hall, 407 F.3d 1195, 1206 (11th Cir.2005) (“This reasonable likelihood standard imposes a ‘con siderably less onerous’ burden on [the petitioner] than the Brady standard.”).

However, the fact that the state court failed to track precisely the language used by the Supreme Court does not mean that it applied the wrong standard here. Indeed, the fact that the state court explicitly recognized the critical difference between the Giglio and Brady standards strongly suggests that it did indeed apply the more lenient Giglio standard. The Florida Supreme Court criticized the state trial court for improperly applying Brady’s standard rather than Giglio’s, and cited our Alzate case for the proposition that the Giglio standard is more defense friendly than the Brady standard.

Moreover, neither the U.S. Supreme Court nor this Court has ever attempted to articulate precisely what the difference is between the Brady standard and the Giglio standard, and we doubt whether the distinction between a “reasonable probability” and a “reasonable likelihood” is readily susceptible of quantification. It has always sufficed for our purposes to observe that one standard is appreciably more stringent than the other. Alzate, for example, called the Brady standard “substantially more difficult for a defendant to meet than the ‘could have affected’ standard we apply.” Alzate, 47 F.3d at 1110 n. 7. Our opinion in Stephens v. Hall called Giglio’s standard “considerably less onerous” than Brady’s. Stephens, 407 F.3d at 1206 (quoting Kirkpatrick v. Whitley, 992 F.2d 491, 497 (5th Cir.1993)). And the Florida Supreme Court plainly captured the essence of this inexact distinction when it observed that the trial court’s reliance “on the materiality standard appropriate to Brady claims” was “incorrect[ ],” and cited Alzate for the proposition that a “different and more defense-friendly standard of materiality applies” to Giglio claims. Ventura, 794 So.2d at 563.

We add that the phrase “contrary to,” as the Supreme Court explained in Williams v. Taylor, means “diametrically different,” “opposite in character or nature,” or “mutually opposed.” 529 U.S. at 405,120 S.Ct. 1495. The Court offered the following example of what would constitute a decision “contrary to” governing law:

Take, for example, our decision in Strickland v. Washington, 466 U.S. 668, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984). If a state court were to reject a prisoner’s claim of ineffective assistance of counsel on the grounds that the prisoner had not established by a preponderance of the evidence that the result of his criminal proceeding would have been different, that decision would be “diametrically different,” “opposite in character or nature,” and “mutually opposed” to our clearly established precedent because we held in Strickland that the prisoner need only demonstrate a “reasonable probability that ... the result of the proceeding would have been different.” Id., at 694, 466 U.S. 668, 104 S.Ct. 2052, 80 L.Ed.2d 674.

Id. at 405-06, 120 S.Ct. 1495. In contrast, a state court decision that is merely “contrary to the federal court’s conception of how Strickland ought to be applied in that particular case ... is not ‘mutually opposed’ to Strickland itself.” Id. at 406, 120 S.Ct. 1495.

A misstatement of a standard whose substance a court applies correctly falls within the latter category. When a court recognizes the substantive difference between two competing standards but conflates their language, the resulting decision is not “diametrically different” from clearly established precedent. A habeas court’s focus is properly on the substance rather than the form of the state court’s decision. See, e.g., Mitchell v. Esparza, 540 U.S. 12, 16, 124 S.Ct. 7, 157 L.Ed.2d 263 (2003) (“A state court’s decision is not ‘contrary to ... clearly established Federal law1 simply because the court did not cite our opinions. We have held that a state court need not even be aware of our precedents, ‘so long as neither the reasoning nor the result of the state-court decision contradicts them.’” (citation omitted) (quoting Early v. Packer, 537 U.S. 3, 7-8, 123 S.Ct. 362, 154 L.Ed.2d 263 (2002))); Bell, 125 S.Ct. at 853.

Finally, the Supreme Court has explained that “§ 2254(d) requires that state-court decisions be given the benefit of the doubt. Readiness to attribute error is inconsistent with the presumption that state courts know and follow the law.” Holland v. Jackson, 542 U.S. 649, 124 S.Ct. 2736, 2739, 159 L.Ed.2d 683 (2004) (alteration, citations, and internal quotation marks omitted). Thus, the Court has several times reversed grants of federal habeas relief on the ground that the state court’s incorrect articulation of the Strickland standard — which requires a defendant claiming ineffective assistance of counsel to demonstrate “a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different,” Strickland, 466 U.S. at 694, 104 S.Ct. 2052 (emphasis added)— was “contrary to” clearly established Supreme Court precedent. See, e.g., Woodford v. Visciotti, 537 U.S. 19, 123 S.Ct. 357, 154 L.Ed.2d 279 (2002); Holland v. Jackson, 542 U.S. 649, 124 S.Ct. 2736, 2738-39, 159 L.Ed.2d 683 (2004). 6

These decisions stand in contrast to Williams v. Taylor, where the Court found that the state supreme court’s application of an incorrect Strickland standard was “contrary to” clearly established law, since the state court’s decision “turned on” its erroneous conception of the relevant standard. The Supreme Court explained: “While [the state court] may also have conducted an ‘outcome determinative’ analysis of its own, it is evident to us that the court’s decision turned on its erroneous view that a ‘mere’ difference in outcome is not sufficient to establish constitutionally ineffective assistance of counsel.” Williams, 529 U.S. at 397, 120 S.Ct. 1495 (emphasis added) (citation omitted).

These eases suggest that a misstatement of the law may be insufficient to find a state court decision contrary to clearly established law; rather, the state court’s decision must have resulted from its applica tion of an incorrect standard. 7 In this case, the Florida Supreme Court understood the law and applied it correctly. The Florida court’s decision is undeniably entitled to deference and is not contrary to clearly established Supreme Court precedent. The only remaining question is whether the state court decision was an “unreasonable application” of such precedent.

C.

Under § 2254(d)’s “unreasonable application” clause, “a federal habeas court may grant the writ if the state court identifies the correct governing legal principle from [the Supreme] Court’s decisions but unreasonably applies that principle to the facts of the prisoner’s case.” Williams, 529 U.S. at 412-13, 120 S.Ct. 1495. However, “a federal habeas court may not issue the writ simply because that court concludes in its independent judgment that the state-court decision applied [the governing legal principle] incorrectly. Rather, it is the habeas applicant’s burden to show that the state court applied [that principle] to the facts of his case in an objectively unreasonable manner. An unreasonable application of federal law is different from an incorrect application of federal law.” Woodford, 537 U.S. at 24-25, 123 S.Ct. 357 (citations and internal quotation marks omitted).

Put differently, the reviewing court may not “substitute[ ] its own judgment for that of the state court.” Id. at 25, 123 S.Ct. 357; see also, e.g., Mitchell, 540 U.S. at 18, 124 S.Ct. 7 (“We may not grant respondent’s habeas petition, however, if the state court simply erred in concluding that the State’s errors were harmless; rather, habeas relief is appropriate only if the [state court] applied harmless-error review in an ‘objectively unreasonable’ manner.”); Lockyer v. Andrade, 538 U.S. 63, 75, 123 S.Ct. 1166, 155 L.Ed.2d 144 (2003) (“The gloss of clear error fails to give proper deference to state courts by conflating error (even clear error) with unreasonableness.”).

In this case, the Florida Supreme Court identified the correct governing legal principle and did not apply it unreasonably. Its conclusion that there was no reasonable likelihood that the false testimony presented at Ventura’s trial could have affected the judgment of the jury was not “objectively unreasonable” in light of the extensive and powerful corroborating evidence introduced and the substantial impeachment of McDonald at trial. 8

As for corroboration, the state’s evidence linking Ventura to Clemente’s murder consisted of McDonald’s testimony; the testimony of three other witnesses: Joseph Pike, Reginald Barrett, and Timothy Arview; and documentary evidence including motel receipts and a record of a wire transfer placing Ventura — then a resident of Chicago — in the vicinity of the murder at the time Clemente was killed.

McDonald testified at length about the details of Clemente’s murder. He explained that Wright was having financial problems and recruited him to kill Clem-ente in exchange for half of the proceeds of a key man life insurance policy Wright held on Clemente. McDonald then brought in Ventura to commit the murder in exchange for half of McDonald’s share of the proceeds. McDonald said that he met with Ventura in Atlanta in April 1981, a few days before the murder, to work out the details. They planned that Ventura would travel to Daytona Beach, where McDonald would meet him. McDonald supplied Ventura with some money for expenses, and later wired him an additional $100 in DeLand, Florida, when Ventura ran out of funds.

McDonald testified that he met Ventura outside the Days Inn on April 15, 1981, and they drove to Barnett Bank in De-Land. Ventura had called Clemente, who was a boat salesman at Crow’s Bluff Marina, and, posing as a potential customer, arranged to meet him at the bank. Clem-ente arrived in a black pick-up truck and left with Ventura. McDonald followed along at a distance in his own car. McDonald and Ventura had planned that Ven-tura would ask Clemente to pull off the road at a prearranged spot along state route 44 so that Ventura could relieve himself, and Ventura would then shoot Clemente. Clemente’s truck pulled off the road at the designated spot, and McDonald waited nearby. About ten minutes later, Ventura returned to McDonald’s car. McDonald dropped Ventura off at a restaurant and proceeded to Daytona Beach to see Wright about getting $2000 for Ventu-ra, who wanted money to travel. According to McDonald, Ventura planned to go to California to see his daughter.

McDonald collected $2000 and met Ven-tura at the Atlanta airport to give him part of it. At that meeting, McDonald said, he informed Ventura that it would take thirty to sixty days to collect the insurance proceeds. McDonald returned to Daytona Beach around June 20 when he learned that Wright was about to be paid by the insurance company, and was arrested for the murder upon his arrival.

Interlocking physical and testimonial evidence introduced at trial corroborated most of the critical aspects of McDonald’s account of the murder. First, McDonald’s testimony that Jerry Wright had taken out a key man insurance policy on Clemente and that the purpose of this murder-for-hire was to collect the proceeds is corroborated on numerous levels. A representative of Midwestern National Life in Cleveland, Ohio confirmed that Jerry Wright held a key man life insurance policy on Robert Clemente issued by Midwestern National Life on July 2, 1980, and that he filed a claim on the policy shortly after the murder, on April 25, 1981. Police searching Ventura’s residence in Maywood, Illi nois after his arrest found a piece of paper inside Ventura’s “Weekly Minder” with the name “Midwestern Life” written on it. The record evidence affords no innocent explanation for this physical evidence linking Ventura to the insurance company that issued the policy for which Clemente was killed.

Moreover, Reginald Barrett, a neighbor and acquaintance of Ventura’s, testified that in February 1981 — approximately two months prior to the murder — -the petitioner, Ventura, asked him to contact Midwestern Life Insurance in Ohio “to inquire about a certain insurance called ‘keyman insurance.’ ” Specifically, Ventura wanted to know how long it would take to pay a claim, and, notably, “if an employee ... was insured under a keyman policy, but was to leave the place of employment, would this insurance policy still cover him after he had left.” (Clemente had left Wright’s employ and worked for Crow’s Bluff Marina at the time of his death.) Barrett never placed the call, but a representative of Midwestern National Life confirmed at trial that the fact Clemente no longer worked for Wright at the time of his death did not affect Wright’s ability to make a claim under the policy. Again, the record evidence offers no benign explanation for why Ventura would ask his Mend Barrett, some two months before the murder, to inquire of the insurance company whether it would honor a key man insurance policy after the employee had left his place of employment.

Joseph Pike, still another acquaintance and one-time business partner of Ventura, provided further independent corroboration of McDonald’s account of the murder. Pike testified that on May 6, 1981 — shortly after the murder — Ventura told Pike about a plan that Ventura, McDonald, and another acquaintance of McDonald’s had devised to murder a man and collect a key man insurance policy on his life. When Pike directly asked what Ventura’s role was, Ventura explained that he “handled the extermination.”

McDonald’s testimony that he met with Ventura in Atlanta in April 1981, and arranged to meet him shortly thereafter in Daytona Beach, was also corroborated by documentary and testimonial evidence. Among other things, Barrett testified that in late February or early March of 1981, Ventura asked him for a gun, which he said he needed for “Jack in Atlanta.” (Barrett in fact provided Ventura with a gun, but not the one used to kill Clem-ente.) Then, during the first week of April, just before the murder, Ventura told Barrett that he needed some money “to go to Atlanta to meet with Jack.” Barrett helped Ventura to get a job that paid him $250, and the next day Ventura told Barrett “that he was on his way to Atlanta” because, “Jack wanted him to come down and burn someone,” which Barrett took to mean “to murder someone.”

Subsequently, around April 10 or 11, Ventura left a message asking Barrett to call him back. When Barrett returned the call, he learned that the phone number Ventura had left was for a motel in Dayto-na or DeLand, Florida. At that time, Ventura asked Barrett to take some money to Ventura’s son, explaining that he was still doing business and would call when he was finished.

Motel records introduced at trial further corroborated McDonald’s testimony that Ventura traveled to the Atlanta area in early April and then continued on to the Daytona area, where he was staying when Clemente was murdered on April 15. Indeed, motel records place Ventura in the Atlanta area from April 5-7 and 9-10, and then in Volusia County from April 10-18. Records from Motel 1 in Chamblee, Georgia' — just outside Atlanta' — -showed that Ventura had checked in April 5, departing April 7, and that he checked in again April 9, departing April 10. Another motel registration card established that Ventura had registered at the Boulevard Motel in De-Land the nights of April 10, 11, and 12, 1981. Records from a Days Inn in the Daytona area established that Ventura had stayed there the nights of April 13 and 14, checking out April 15,1981. More records from Motel 1 in Chamblee showed that Ventura checked back into that motel on April 16, departing April 18. On this record, we can find no benign explanation for why Ventura, a resident of the Chicago area, had been in Atlanta or Volusia County during this period.

The records from the Days Inn, where Ventura stayed from April 13-15, also corroborate McDonald’s testimony that he picked up Ventura in front of the Days Inn on April 15, the day of Clemente’s murder. In addition, Corporal David Hudson of the Volusia County Sheriffs Department testified that he checked out a phone number from which Reginald Barrett had received a call around 9:00 a.m. on April 15, and traced the number to a pay phone outside the same Days Inn.

The records indicating that Ventura stayed at the Boulevard Motel the nights of April 10-12 interlock with Barrett’s testimony that he reached a motel in Daytona or DeLand when he dialed the return number Ventura had left when he called Barrett around April 10 or 11. Corporal Hudson further verified that his investigation revealed that Barrett had placed a return call to Ventura at the Boulevard Motel, although he could not remember the precise date of the call.

Additional documentary evidence corroborated McDonald’s testimony that he wired $100 to Ventura in DeLand, Florida, when Ventura ran out of expense money. Western Union records introduced at trial showed that on April 13, Peter Ventura picked up a $100 money order, which had been sent on April 12 by Jack McDonald, at an office located in DeLand’s Greyhound bus station.

In addition, McDonald’s account of the details of the Clemente murder was entirely consistent with the forensic evidence presented at trial and with the testimony of several of Clemente’s co-workers. Several law enforcement officers dispatched to the crime scene reported that Clemente’s body was recovered in a secluded area off state route 44; that it was found in a black pick-up truck bearing the Crow’s Bluff Marina insignia; and that Clemente had been shot to death. Denise Jorgenson, the owner of Crow’s Bluff Marina, confirmed that Clemente had told her on the day of his murder that he was meeting a client at Barnett Bank, an occurrence Jorgenson considered unusual since clients typically came to the marina. Charles W. Coccia, a mechanic at the marina, testified that he lent Clemente his company truck to go meet a client on the day of Clemente’s murder. Coccia identified the truck in which Clemente’s body was found as the one he had given Clemente.

The delay McDonald recounted in collecting the insurance proceeds after the murder was also corroborated by the testimony of both Pike and Barrett. Pike testified that when Ventura told him about the murder in early May, “there were indications that there were delays in collecting,” and Ventura “was expecting to collect it within a few weeks.” Barrett testified that in early May, upon Ventura’s return from a trip to California, Ventura mentioned “that he would be getting his payment in about thirty days,” when “Mr. McDonald was to receive some funds from an insurance proceed in Florida.” Then, in late May, Ventura told Barrett that “Jack should be coming to Chicago within just a few days to make a payment” for the job Ventura had done. Ventura ex plained that “there was some delay” in receiving the money that “Jack ... was going to collect ... from someone who had an insurance policy.” A representative of the insurance company, Midwestern National Life, confirmed that the key man insurance claim had been filed on April 25, 1981, and that payment had been issued initially on June 19, 1981, but that payment on the check had been stopped and that the claim had not been paid until September 14,1981.

Barrett also confirmed McDonald’s testimony that Ventura intended to travel to California after the murder. Ventura called Barrett around April 19 or 20, 1981, and said that he would be going to California as soon as his job was finished. Ventura called Barrett again in early May “indicating that he had gotten back from California,” and Barrett picked him up at the bus station.

A letter Ventura sent to Barrett further corroborates Ventura’s involvement in the murder with McDonald. Around May 4 or 5, when Barrett returned from a trip, he had received a letter addressed to him at his office in Maywood, Illinois, postmarked April 17. The return address read, “P. Ventura, Atlanta, Georgia.” The letter, which was introduced at trial, stated, “In the event something should happen to me! Jack McDonald is responsible.” The letter then supplied McDonald’s phone number. Included in the mailing envelope along with the letter was an envelope bearing the return address of Motel 1 — the motel in Chamblee, Georgia where Ventu-ra was staying on April 17. On that envelope was written, “Jack lives near here.”

Ventura’s subsequent flight after his indictment for the Clemente murder is further evidence of his role in the crime. Ventura was released on $50,000 bond after his indictment in late July, 1981, and failed to appear at a hearing scheduled for August 18, 1981. In spite of the efforts of state and federal law enforcement to track him down, Ventura remained a fugitive for five years. He lived under several assumed names, including Juan Gadaya and Juan Contras, until he was apprehended in Austin, Texas in June 1986. As we have noted repeatedly, “evidence of resistance to arrest and flight is admissible to demonstrate consciousness of guilt and thereby guilt.” United States v. Wright, 392 F.3d 1269, 1277 (11th Cir.2004) (quoting United States v. DeParias, 805 F.2d 1447, 1454 (11th Cir.1986)); see also, e.g., United States v. Frazier, 387 F.3d 1244, 1266 n. 20 (11th Cir.2004) (en banc) (finding that defendant’s flight from police “was strong evidence of consciousness of guilt”); United States v. Beard, 775 F.2d 1577, 1581 (11th Cir.1985) (“There is no question that evidence of flight can raise a permissive inference of consciousness of guilt of the crime charged.”).

Finally, Ventura’s role in the murder was independently corroborated by the trial testimony of Timothy Arview, a young man who became acquainted with Ventura while Ventura was living in Austin, Texas as a fugitive. Arview, who did not know McDonald, Barrett, or Pike, testified at trial that he first met Ventura, who was using the name Juan Gadaya, in December 1985 or January 1986, and worked for him for six or seven months doing construction. Around April or May of 1986, Arview and Ventura were wrestling and “playing around” until Ventura “got upset and started playing roughly.” Ventura then specifically told Arview that he had killed a man in Florida five years earlier, and that it was a contract killing. Lieutenant Juan Gonzales of the Austin Police Department confirmed that Arview reported this information — which matched the date, location, and nature of the Clemente killing exactly — to Austin police, which in turn led to Ventura’s rearrest.

In addition to this extensive chain of corroborating evidence, the Florida Supreme Court took into account the fact that McDonald was substantially impeached in a variety of ways at trial. First, McDonald was impeached by his admitted involvement in the Clemente murder. McDonald testified at trial that he had been indicted for the murder along with Ventura, but that he had escaped prosecution because of Florida’s speedy trial rule.

Second, McDonald was impeached by evidence of his prior convictions. McDonald conceded on direct examination that after his release on speedy trial grounds, he had been indicted by the federal government and had pled guilty, in 1983, to four counts of postal theft. He was sentenced in May or June of 1983 to a term of imprisonment of fifteen years, then fled before he started serving his sentence, and was not taken back into custody until September 23, 1987. On cross-examination, McDonald also acknowledged another prior conviction, for embezzlement involving bank fraud, in 1964. Finally, McDonald admitted that he had been involved in the illegal sale of a load of stolen truck tires to Mr. Wright, although he was apparently never prosecuted for this offense.

Third — and perhaps most significantly— McDonald was extensively impeached on cross-examination with his own prior inconsistent deposition testimony. McDonald was deposed at the Sheraton Inn at the Atlanta Airport on May 4, 1983, by the prosecutor, Stark. At that deposition, McDonald said that he had not been with Ventura on the day of the murder, in direct contradiction of his trial testimony. At trial, McDonald admitted that in spite of being sworn to testify truthfully, he “did not tell the truth” in that deposition. McDonald explained that his “intention was just to leave some clouds in the issue,” and that perhaps he “did not want to face the reality of being virtually, you might say, at the scene.” Defense counsel chipped further at McDonald’s credibility by insinuating that he had lied in the deposition “because [he was], in fact, at the scene and did commit the murder,” a suggestion that McDonald emphatically denied.

Finally, the defense attacked McDonald’s motive for testifying by suggesting that McDonald harbored “some feeling of rancor towards Mr. Ventura as a result of the Federal bank scam that resulted in [McDonald’s] conviction.” Again, McDonald flatly denied this. Defense counsel then asked whether McDonald had agreed to testify because he was angry that Wright never paid him for his participation in the murder, another suggestion that McDonald strongly denied.

We are bound by the state court’s decision unless it was not only erroneous but “objectively unreasonable.” Woodford, 537 U.S. at 24-25, 123 S.Ct. 357. As the Supreme Court has admonished, a federal habeas court is not free to “substitute[ ] its own judgment for that of the state court.” Id. at 25, 123 S.Ct. 357. We cannot, in this case, characterize as unreasonable the state court’s determination that the corroboration and impeachment were so substantial that there was no reasonable likelihood McDonald’s false testimony could have affected the judgment of the jury.

Virtually every step of McDonald’s testimony as he took the jury through the planning, execution, and aftermath of Clemente’s murder was corroborated by documentary and testimonial evidence. The multiple admissions Ventura made over many years to Barrett, Pike, and Arview, as well as the testimony of Clem-ente’s co-workers and of police who responded to the crime scene, mirrored McDonald’s account of the crime. Motel receipts, Western Union records, and phone calls placed Ventura — a Chieago-area resident — in Atlanta and in Volusia County at precisely the times McDonald said he was there, including the time of Clemente’s murder. A conversation between Ventura and Barrett and the note found in Ventura’s Weekly Minder directly linked Ventura to the insurance company that issued the key man policy on Clemente’s life. A subsequent note in which Ventura suggested McDonald might have it in for him, as well as Ven-tura’s flight after his indictment, further indicated Ventura’s involvement in the murder.

In addition, McDonald was extensively impeached by evidence of his own involvement in the Clemente murder, his prior felony fraud convictions and bond jumping, and his prior false deposition testimony in the Clemente case, as well as by the implication that he was motivated to testify by ill will toward Ventura and Wright. 9

The substantial volume of interlocking corroborating and impeachment evidence soundly supports the Florida Supreme Court’s conclusion that a jury would not have discounted McDonald’s testimony even if it knew he had been promised a deal on potential bond-jumping charges in exchange for his testimony, and thus that there was no reasonable likelihood that McDonald’s false testimony could have affected the judgment of the jury. We therefore agree with the district court that the Florida Supreme Court’s disposition of Ventura’s Giglio claim was not an unreasonable application of clearly established federal law.

IV.

The conduct of the prosecutor in this case was inexcusable, and wholly discordant with his paramount obligation to seek the truth. Nevertheless, the Florida Supreme Court’s determination that McDonald’s false testimony was not “material,” as required by Giglio, was neither contrary to nor an unreasonable application of clearly established federal law. We, therefore, affirm the district court’s denial of Ventura’s petition for a writ of habeas corpus.

AFFIRMED.

1 . Jerry Wright was tried and convicted of the first-degree murder of Robert Clemente in February, 1990, and sentenced to life imprisonment.

2 . In 1983, when McDonald jumped bond, bond jumping carried a penalty of a fine of up to $5000, and/or up to five years' incarceration. The statute in effect at that time, 18 U.S.C. § 3150 (which was enacted in 1966 and repealed in 1984), stated:

Whoever, having been released pursuant to this chapter, willfully fails to appear before any court or judicial officer as required, shall, subject to the provisions of the Federal Rules of Criminal Procedure, incur a forfeiture of any security which was given or pledged for his release, and, in addition, shall (1) if he was released in connection with a charge of felony, or while awaiting sentence of pending appeal or certiorari after conviction of any offense, be fined not more than $5,000 or imprisoned not more than five years, or both, or (2) if he was released in connection with a charge of misdemeanor, be fined not more than the maximum provided for such misdemeanor or imprisoned for not more than one year, or both, or (3) if he was released for appearance as a material witness, shall be fined not more than $1,000 or imprisoned for not more than one year, or both.

Section 3150 did not require the trial judge to run the bond-jumping sentence consecutively to the sentence for the underlying charges. However, under the current version of the statute, 18 U.S.C. § 3146, the sentences must run consecutively. Id. § 3146(b)(2).

3 . Giglio error does not fall within the category of constitutional error that the Supreme Court has characterized as "structural,” requiring automatic reversal of a tainted conviction. A structural error is a “defect affecting the framework within which the trial proceeds, rather than simply an error in the trial process itself.” Arizona v. Fulminante, 499 U.S. 279, 310, 111 S.Ct. 1246, 113 L.Ed.2d 302 (1991). The Supreme Court has found structural error “only in a very limited class of cases,” including those involving complete denial of counsel, a biased trial judge, racial discrimination in selection of a grand jury, denial of self-representation, denial of a public trial, and a defective reasonable doubt instruction. Neder v. United States, 527 U.S. 1, 8, 119 S.Ct. 1827, 144 L.Ed.2d 35 (1999) (citation and internal quotation marks omitted).

4 . Our Court has held, on several occasions, that Giglio's "any reasonable likelihood” standard is equivalent to the harmless error inquiry of Chapman v. California, 386 U.S. 18, 87 S.Ct. 824, 828, 17 L.Ed.2d 705 (1967), which provides that a conviction must be reversed if constitutional trial error occurs, unless the beneficiary of the error establishes that the error was "harmless beyond a reasonable doubt.” Id. at 24, 87 S.Ct. 824. We have stated at least twice in cases before us on direct review, see Alzate, 47 F.3d at 1110; United States v. Rivera Pedin, 861 F.2d 1522, 1529 n. 13 (11th Cir.1988), and at least once in a case before us on habeas review, see Carr v. Schofield, 364 F.3d 1246, 1255 (11th Cir.2004), that the harmless-error inquiries required by Giglio and by Chapman are indistinguishable.

Although this is the law of our Circuit, it is not clearly established federal law as determined by the Supreme Court, since no majority has ever held that the Giglio standard is equivalent to the Chapman standard, and since the Supreme Court's decision in Brecht v. Abrahamson, 507 U.S. 619, 113 S.Ct. 1710, 123 L.Ed.2d 353 (1993), casts substantial doubt on whether Chapman is ever the appropriate standard for harmless-error analysis conducted on collateral review.

In concluding that the two standards are equivalent, we have drawn on Justice Black-mun's opinion in United States v. Bagley, 473 U.S. 667, 105 S.Ct. 3375, 87 L.Ed.2d 481 (1985), which observed that Giglio's "any reasonable likelihood” standard "may as easily be stated as a materiality standard under which the fact that testimony is perjured is considered material unless failure to disclose it would be harmless beyond a reasonable doubt.” Id. at 678-80, 105 S.Ct. 3375. Thus, Justice Blackmun concluded, "the standard of review applicable to the knowing use of perjured testimony is equivalent to the Chapman harmless-error standard.” Id. at 680 n. 9, 105 S.Ct. 3375. Although Justice Blackmun clearly equated the Giglio and Chapman standards, this portion of his opinion was joined only by Justice O'Connor and, consequently, does not constitute a holding of the Court.

The Chapman Court itself suggested some similarity between the standards, stating: "There is little, if any, difference between our statement in Fahy v. State of Connecticut about 'whether there is a reasonable possibility that the evidence complained of might have contributed to the conviction’ and requiring the beneficiary of a constitutional error to prove beyond a reasonable doubt that the error complained of did not contribute to the verdict obtained.” Chapman, 386 U.S. at 24, 87 S.Ct. 824 (quoting Fahy v. Connecticut, 375 U.S. 85, 84 S.Ct. 229, 11 L.Ed.2d 171 (1963)).

This statement does not, however, suffice to render it "clearly established” that the Chapman and Giglio standards are indistinguishable. For one thing, this statement in Chapman was dictum. Moreover, Fahy involved a different kind of constitutional error — admission of unconstitutionally obtained evidence— and the Fahy standard, while very similar to the Giglio standard, is not identical. (Fahy asks “whether there is a reasonable possibility that the evidence complained of might have contributed to the conviction"; Giglio asks whether “there is any reasonable likelihood that the false testimony could have affected the judgment of the jury.”) Indeed, when Giglio was decided nine years after Fahy, the Court made no mention of Fahy.

Because the Supreme Court has never definitively stated that Giglio 's “any reasonable likelihood” inquiry is the same as the Chapman "harmless beyond a reasonable doubt” inquiry, the only clearly established law governing Giglio materiality analysis is that Gig-lio error requires reversal when there is any reasonable likelihood that the false testimony could have affected the judgment of the jury. As the Supreme Court recently emphasized, “[a] federal court may not overrule a state court ... when the precedent from this Court is, at best, ambiguous.” Mitchell v. Esparza, 540 U.S. 12, 17, 124 S.Ct. 7, 157 L.Ed.2d 263 (2003). Accordingly, the "any reasonable likelihood” formulation is the only one the Florida court was obligated to apply in resolving Ventura's Giglio claim.

Moreover, even if Chapman is seen as establishing the proper materiality standard for a Giglio claim on direct review, Brecht casts substantial doubt on whether this standard has any application on habeas review. Brecht held that Chapman's "harmless beyond a reasonable doubt” standard is not the proper harmless-error inquiry on collateral review; rather, habeas courts must apply the test laid out in Kotteakos v. United States, 328 U.S. 750, 66 S.Ct. 1239, 90 L.Ed. 1557 (1946), which asks whether the error “had substantial and injurious effect or influence in determining the jury’s verdict.” Brecht, 507 U.S. at 637, 113 S.Ct. 1710 (quoting Kotteakos, 328 U.S. at 776, 66 S.Ct. 1239). The Court explained that “an error that may justify reversal on direct appeal will not necessarily support a collateral attack on a final judgment,” id. at 634, 113 S.Ct. 1710 (citations and internal quotation marks omitted), citing “the State's interest in the finality of convictions that have survived direct review within the state court system,” "comity and federalism” concerns, and the integrity of the trial process as reasons for distinguishing between direct and collateral review. See id. at 633-36, 113 S.Ct. 1710. In addition, "[sjtate courts are fully qualified to identify constitutional error and evaluate its prejudicial effect on the trial process under Chapman, and state courts often occupy a superior vantage point from which to evaluate the effect of trial error.” Id. at 636, 113 S.Ct. 1710. Thus, "it scarcely seems logical to require federal habeas courts to engage in the identical approach to harmless-error review that Chapman requires state courts to engage in on direct review.” Id.

As several of our sister Circuits have observed, Brecht continues to apply on habeas review even after AEDPA when a state court either fails to conduct any harmless error review because it unreasonably concludes that no constitutional error has occurred, or conducts harmless error review applying the wrong standard (i.e. when the state court's decision is contraiy to or an unreasonable application of clearly established federal law). See, e.g., Aleman v. Sternes, 320 F.3d 687 (7th Cir.2003); Herrera v. Lemaster, 301 F.3d 1192 (10th Cir.2002) (en banc). Thus, even when a state court improperly fails to use the Chapman standard in conducting harmless error review, a federal habeas court reviews the constitutional error only under the more lenient Kotteakos standard. Cf. Penry v. Johnson, 532 U.S. 782, 795, 121 S.Ct. 1910, 150 L.Ed.2d 9 (2001) ("Even if our precedent were to establish squarely that the prosecution[ ] ... violated Penry’s Fifth Amendment privilege against self-incrimination, that error would justify overturning Penry’s sentence [on federal habeas review] only if Penry could establish that the error 'had substantial and injurious effect or influence in determining the jury’s verdict.' " (quoting Brecht, 507 U.S. at 637, 113 S.Ct. 1710)). Chapman, therefore, has little application to a case before us on collateral review.

5 . The Florida Supreme Court also suggested that in determining whether the Giglio standard is satisfied, "courts must focus on whether the favorable evidence could reasonably be taken to put the whole case in such a different light as to undermine confidence in the verdict." Ventura, 794 So.2d at 563 (quoting White v. State, 729 So.2d 909, 913 (Fla.1999)). Although no Supreme Court case explains the Giglio standard in these terms, it bears noting that our Court has, on at least one occasion, explained the Giglio standard using nearly identical terms. In United States v. Dickerson, 248 F.3d 1036, 1041 (11th Cir.2001), a panel of this Court stated: "The materiality element [of Giglio ] is satisfied if the false testimony 'could reasonably be taken to put the whole case in such a different light as to undermine confidence in the verdict.’ ” Id. at 1041 (quoting Strickler v. Greene, 527 U.S. 263, 290, 119 S.Ct. 1936, 144 L.Ed.2d 286 (1999)) (internal quotation marks omitted); see also Carr v. Schofield, 364 F.3d 1246, 1255 (11th Cir.2004) (explaining the standard as whether "the defendant failed to receive 'a trial resulting in a verdict worthy of confidence'”) (quoting Kyles, 514 U.S. at 434, 115 S.Ct. 1555).

6 . In Woodford, the Court observed that the state court's "occasional shorthand reference to that standard by use of the term 'probable' without the modifier may perhaps be imprecise, but if so it can no more be considered a repudiation of the standard than can this Court’s own occasional indulgence in the same imprecision." Id. at 23-24, 123 S.Ct. 357 (citing Mickens v. Taylor, 535 U.S. 162, 166, 122 S.Ct. 1237, 152 L.Ed.2d 291 (2002) (using the phrase "probable effect upon the outcome”), and Williams, 529 U.S. at 393, 120 S.Ct. 1495 (using the phrase "probably affected the outcome”)).

This imprecision, the Court held, did not warrant the conclusion that the state court’s decision was contrary to clearly established law, and indeed the Court criticized the federal appeals court for not giving the benefit of the doubt to the state court:

The Court of Appeals made no effort to reconcile the state court's use of the term "probable" with its use, elsewhere, of Strickland's term "reasonably probable,” nor did it even acknowledge, much less discuss, the California Supreme Court's proper framing of the question as whether the evidence "undermines confidence” in the outcome of the sentencing proceeding. This readiness to attribute error is inconsistent with the presumption that state courts know and follow the law. It is also incompatible with § 2254(d)'s "highly deferential standard for evaluating state-court rulings,” Lindh v. Murphy, 521 U.S. 320, 333, n. 7, 117 S.Ct. 2059, 138 L.Ed.2d 481 (1997), which demands that state-court decisions be given the benefit of the doubt.

Id. at 24, 123 S.Ct. 357.

7 . This conclusion is consistent with the numerous cases in our Court concluding that federal courts are required to defer to state court decisions even when they offer no explanation or reasoning as to how they reached their result. As we have explained: "The plain language of § 2254(d)(1) requires only that the federal claim have been 'adjudicated on the merits in State court proceedings' and have 'resulted in a decision' that is neither contrary to nor involves an unreasonable application of Supreme Court precedent. That is all the text of the provision requires.” Wright v. Moore, 278 F.3d 1245, 1254 (11th Cir.2002). "The statutory language focuses on the result, not on the reasoning that led to the result, and nothing in that language requires the state court adjudication that has resulted in a decision to be accompanied by an opinion that explains the state court’s rationale. Accordingly, all that is required is a rejection of the claim on the merits, not an explanation.” Id. at 1255 (citations omitted). Accord Hening v. Sec'y, Dep’t of Corrs., 397 F.3d 1338, 1347 (11th Cir.2005) ("To be entitled to deferential review under 28 U.S.C. § 2254, it is not necessary for the state court to explain its decision. Even a summary, unexplicated rejection of a federal claim qualifies as an adjudication entitled to deference under § 2254(d).”); Parker v. Sec’y, Dep’t of Corrs., 331 F.3d 764, 775-76 (11th Cir.2003); Isaacs v. Head, 300 F.3d 1232, 1259-60 (11th Cir.2002).

8 . Ventura makes a great deal of the fact that prosecutor Stark, in the state-court hearing on Ventura's 3.850 motion, referred to McDonald numerous times as the state's "star witness.” Indeed, the State concededly felt that it could not proceed with the prosecution of either McDonald or Ventura without the other. Nevertheless, the State’s case against Ventura was compelling even in light of McDonald's false testimony, since the numerous sources of powerful corroboration make it implausible that the jury would have wholly discarded McDonald’s testimony had they known that he was testifying pursuant to a deal. We cannot substitute one participant’s conclusory characterization of a witness’s testimony for a thorough review of all of the evidence presented at trial.

9 . The effect of these other means of impeachment was not, however, to render evidence of McDonald's deal cumulative. See, e.g., United States v. Sanfilippo, 564 F.2d 176, 178 (5th Cir.1977) ("The fact that the history of a witness shows that he might be dishonest does not render cumulative evidence that the prosecution promised immunity for testimony. A jury may very well give great weight to a precise reason to doubt credibility when the witness has been shown to be the kind of person who might perjure himself.”); accord Brown, 785 F.2d at 1466. Rather, the other sources of impeachment are simply relevant to the determination of whether there is any reasonable likelihood that evidence of McDonald’s deal could have affected the judgment of the jury.