Upon Further Review

A Publication of the Philadelphia Bar Association

Home > Criminal Justice

Print | | Share Stumble Upon Facebook Delicious Digg Reddit Google

Third Circuit Determines Standard to Apply on Habeas Where State Knowingly Presented or Failed to Correct Perjured Testimony

Upon Further Review Staff on 8/2/2017
PRECEDENTIAL
UNITED STATES COURT OF APPEALS
FOR THE THIRD CIRCUIT
________________
No. 15-3427
________________
VANCE HASKELL,
Appellant
v.
SUPERINTENDENT GREENE SCI;
ATTORNEY GENERAL PENNSYLVANIA;
DISTRICT ATTORNEY ERIE COUNTY
________________
Appeal from the United States District Court for the Western District of Pennsylvania
(D.C. Civil Action No. 1-10-cv-00149)
District Judge: Honorable Susan Paradise Baxter
________________
Argued March 27, 2017

Before: AMBRO, VANASKIE, and
RESTREPO, Circuit Judges
(Opinion filed August 1, 2017)

Lisa B. Freeland, Esquire
Federal Public Defender
Elisa A. Long, Esquire (Argued)
Assistant Federal Public Defender
Office of Federal Public Defender
1001 Liberty Avenue
1500 Liberty Center
Pittsburgh, PA 15222
Counsel for Appellant

Mark W. Richmond, Esquire (Argued)
Erie County Office of District Attorney
140 West 6th Street
Erie, PA 16501
Counsel for Appellee

________________
OPINION OF THE COURT

________________
AMBRO, Circuit Judge

Article Image

A gunman murdered Darrell Cooley in a bar in Erie, Pennsylvania, in December 1994. Nearly four years later, the Commonwealth of Pennsylvania indicted and tried Vance Haskell for Cooley's murder. The primary issue at the trial was whether Haskell was the gunman. In addition to circumstantial evidence linking Haskell to the murder, the Commonwealth presented four eyewitnesses. But one of these eyewitnesses recanted his pre-trial testimony implicating Haskell and two had previously denied that they would be able to identify the shooter. The fourth eyewitness, Antoinette Blue, did provide consistent testimony claiming she could identify the shooter. What's more, she claimed to expect nothing in return from the Commonwealth in exchange for her testimony. But this last claim was untrue. Both Blue and the prosecutor knew that she expected to receive help in her own pending criminal matters in exchange for her testimony. The prosecutor failed to correct Blue's statement; he even went on to rely on it and vouch for Blue in his closing argument.

Haskell filed a habeas petition challenging his conviction as tainted by perjured testimony in violation of his Fourteenth Amendment right to due process. We must decide whether Haskell is entitled to relief once he has shown a reasonable likelihood the false testimony could have affected the judgment of the jury, Giglio v. United States, 405 U.S. 150, 154 (1972) (citing Napue v. Illinois, 360 U.S. 264, 271 (1959)), or whether he must also show Blue's perjured testimony caused him "actual prejudice" under the standard in Brecht v. Abrahamson, 507 U.S. 619, 637-38 (1993). We hold that Brecht does not apply when the State has knowingly presented or failed to correct perjured testimony. In those circumstances a petitioner carries his burden when he makes the reasonable likelihood showing required by Giglio and Napue. Because Haskell has done so here, we grant his petition.

  1. BACKGROUND

    In the early hours on December 10, 1994, a man entered a bar called Jethroe's Steakhouse in Erie, Pennsylvania with a semiautomatic weapon (described at trial as an "Uzi") and opened fire. He shot roughly 14 times, killing Darrell Cooley and wounding Kevin Twillie. The shooter fled the scene with another man, Curtis Mathis.

    Mathis was convicted in November 1995 for his role in these crimes (two counts of hindering apprehension of the shooter). He did not identify the shooter and received three to seven years in prison.

    A year into his sentence, Mathis, hoping that his cooperation would result in parole, communicated with Detective Sergeant James Skindell to cooperate in the ongoing investigation to identify the shooter. He provided a videotaped statement in which he named Vance Haskell (whom Mathis also called "Hakeem") as the shooter. Haskell was charged with Cooley's murder, aggravated assault of Twillie, unlawful carrying of a firearm, and several related crimes in November 1997. His trial began ten months later.

    As noted, the primary issue at trial was the identity of the shooter. The Commonwealth's prosecutor, Matthew R. Hayes, presented testimony from over 40 witnesses; only four-Mathis, Roseanna Wayne, Dorothea Roberts, and Blue-ever claimed to be able to identify Haskell as the shooter, and all except Blue had denied-either at trial or before-that they could do so.

    1. Trial Testimony

      Haskell is from Rochester, New York, but the Commonwealth presented evidence that he was in Erie, Pennsylvania around the time of the murder. Felicia Clark testified that Haskell and Mathis had been staying at her Erie apartment in Franklin Terrace in the weeks leading up to the shooting. The two drove from Rochester to Erie with Clark's brother, and evidence suggests that he drove Haskell away from Erie between December 9th and 11th: Clark's uncle testified at trial that he had loaned his car to her brother on December 9, 1994; when he got it back two days later, it had been driven 586 miles, and police later matched to Haskell fingerprints on beer bottles left in the car.

      Nine witnesses testified that the unknown shooter was wearing a puffy coat; although two of them recalled the coat being blue or black in color, the other seven described it as green. Eight witnesses testified that Mathis and an unidentified shooter were in Jethroe's together and fled after the shooting. One man who was in the parking lot during the shooting testified that he saw Mathis and Haskell running toward the alley behind Jethroe's. He noted that Haskell was wearing a "big fluffy jacket." Also, three witnesses testified that Mathis and someone else took a cab to Franklin Terrace. One of them was the cab driver, although his only recollection was of picking up two black men from Red's Tavern, which is not far from Jethroe's. A resident of Franklin Terrace picked Haskell out of a photo line-up and said he had been at her home after the shooting.

      Two witnesses testified they previously saw Haskell with a gun similar to the firearm recovered in an alley near Jethroe's. One said that he had seen Haskell several times at Felicia Clark's home with "a nine-millimeter pistol, 380 automatic, an Uzi, like, type machine gun." J.A. 745. He also identified the recovered Uzi as the gun he saw Haskell carrying. The other testified that she had seen Haskell with a black firearm "slightly bigger than your average handgun" four days before the shooting and that he was wearing a green down coat at that time. Id. at 683.

      In sum, these witnesses placed Haskell in Erie near Jethroe's around the time of the shooting and put two key items associated with the shooter in his possession: a large gun and a green, fluffy jacket. But none of them saw Haskell shoot the victims.

      Four individuals presented eyewitness testimony of the shooting. But each witness's testimony came with a few problems for the prosecution.

      The first, Mathis, who had put the Commonwealth on Haskell's trail and had already been convicted of assisting the shooter's escape, recanted his previous statements on the stand. In Mathis's videotaped statement, he said that he and Haskell went to Jethroe's together that night and, while he did not witness the shooting occur, he saw Haskell immediately after wearing a green "puffy" coat and holding a smoking gun. Mathis also stated that he fled the bar with Haskell, that Haskell threw the gun under a vehicle in an alley and tossed off his coat, and that they went to another nearby bar. From there they got into a cab to head toward Franklin Terrace. In the video Mathis was shown a photo line-up and identified Haskell.

      But at trial Mathis recanted his videotaped statement and testified instead that he was at Jethroe's at the time of the shooting but was not there with Haskell. He claimed he did not witness anyone with a gun; he left the bar in a car he drove himself and went to a place known as the "Holly" rather than the Franklin Terrace housing project.1 Mathis agreed that, when he gave the videotaped statement, his "only concern [was] getting out of jail[.]" Id. at 629. He related that he reviewed police reports of the murder (of which he had copies from his own involvement in the case) in order to tell the police and prosecutors what they wanted to hear-that is, what would make their case against Haskell.

      Second, Roseanna Wayne only stated that Haskell's appearance was not inconsistent with that of the shooter. Id. at 348 ("If the hair was down lower, the beard was off the face, the mustache was off the face … [,] [i]t look [sic] like the shooter."). But earlier in her testimony she said she was not sure she could identify the shooter. Id. ("Q. If you saw that person again, the person that was doing the shooting, do you think you'd recognize that person? A. No. I don't know."). At other moments in her testimony, Wayne appeared to be more confident, id. at 357 ("Sir, he look just like the man."), but she also admitted on cross-examination that she had never seen Haskell before the day she testified in court, id. at 357 ("Q. Never seen Mr. Haskell before today? A. No, sir.").

      The third eyewitness, Dorothea Roberts, testified in court that she was at Jethroe's on the night of the shooting and that she saw the shooter. She further testified that Haskell was the shooter and identified him in court. However, about three months after the shooting, Roberts had told Detective Skindell that she did not see the shooter. On cross-examination, Roberts denied this and said that if Detective Skindell wrote that in his report, he must have lied. Roberts also testified that she was currently in the Erie County Jail on charges of simple assault. Id. at 483.

      With three eyewitnesses who each made questionable identifications of Haskell, Prosecutor Hayes called Antoinette Blue to the stand. She testified that she saw Haskell shoot Cooley and had met Haskell before the shooting took place, strengthening the power of her identification. As context, Blue stated that she had seen Haskell around town for a few weeks, and, 20 minutes before the shooting, she smoked marijuana with him, Mathis, Felicia's Clark's brother, and a woman named Yolanda in Jethroe's parking lot. That night, Blue did not report to the police that she was able to identify the shooter.

      She never spoke with the police about the incident until three years later in February of 1998.

    2. Blue's Communications with the Commonwealth

      Blue was in the Erie County Jail when she finally spoke to the police about the shooting. Two warrants brought her there. One was issued for a parole violation following her conviction for disorderly conduct and resisting arrest. The second stemmed from her failure to appear for sentencing after pleading guilty to a charge of attempted theft.

      And Blue had other troubles in Mercer County. She was arrested there and charged with receiving stolen property, criminal conspiracy, unsworn falsification, three misdemeanor counts of retail theft, and four summary counts of retail theft. It was following this arrest that police transported her to the Erie County Jail because of that County's two outstanding warrants. Back in Erie, she reached out to Detective Skindell to cooperate in Haskell's case.

      Blue lied when she testified at Haskell's preliminary hearing on March 18, 1998. When asked on cross-examination whether she had "any criminal charges pending against [her,]" she left out her numerous pending charges in Mercer County and responded that she was "just [in jail] on a probation violation." Id. at 79.

      Blue also testified adamantly that she never discussed with anyone whether cooperating with Haskell's prosecution would help her get out of jail and that "it never occurred to [her]" that cooperation might be helpful to her. Id. at 109-10. But just two days after testifying at the preliminary hearing, Blue received sentences on her parole violations that (despite having picked up additional charges in Mercer County) resulted in her release from custody.

      Within weeks of the preliminary hearing, Detective Skindell informed Mercer County authorities that Blue was a cooperating witness in Erie County's case against Haskell. Skindell also told Blue's Mercer County defense attorney about her cooperation, and the attorney responded by sending a strongly worded letter to the Mercer County DA demanding a favorable outcome on Blue's pending charges due to her cooperation in the Haskell case. Finally, the prosecutor in Haskell's case, Hayes, reached out to the Mercer DA, who informed Hayes that the judge in Blue's case would be told of her cooperation at sentencing.2

      In September 1998, Blue testified at Haskell's trial. On cross-examination, Haskell's attorney pointed out that Blue was released from jail after she testified at Haskell's preliminary hearing. But Blue denied that the Erie County judge was aware of her cooperation and said that the timing of her release was "just a coincidence." Id. at 522.

      She also denied that she cooperated with the police in exchange for help with her criminal matters. When asked on what charges she was in jail at the time she communicated with Detective Skindell, Blue again mentioned only her probation violation and said nothing about her charges in Mercer County.

      Haskell's attorney then asked several questions aimed at revealing Blue's motivation to cooperate.

      Q. And did you contact the District Attorney's Office because you wanted some help to get out of jail?
      A. Get out for what? I wasn't facing a lot of time, what did I need help for?
      Q. So you didn't – this never came into your mind that you wanted to get help to get out of jail?
      A. No. Get out for what?
      Id. at 517.
      On re-direct, Prosecutor Hayes attempted to dispel the notion that Blue had agreed to cooperate in order to receive some benefit in her own criminal matters.
      Q… . Have you been promised anything by us to come in here and explain what you just explained?
      A. No.
      Q. Do you anticipate receiving any consideration for it?
      A. Do I what?
      Q. Do you expect to get something out of testifying?
      A. No, sir.
      Id. at 521.
      On re-cross, Blue again insisted that she would receive no benefit for testifying.
      Q. You didn't ask anybody to take [your testimony] into consideration?
      A. No, sir.
      Q. You don't think anybody was aware of that?
      A. No, sir.
      Id. at 522.

      In his closing argument, Prosecutor Hayes ridiculed the idea that Blue would benefit from her testimony and vouched for her credibility:

      Antoinette says that she sees Haskell over at the [sic] Felicia Clark's place. She also sees him out in the parking lot, and here she is the one that is trying to get all this benefit from this-this valuable testimony. And what she says she's doing out there, she's committing a crime. She's smoking marijuana. That should help her pretty well. Id. 1033-34.

      So, yes, she gives her statement [three] years later; yes, it's during the time she's in prison. Is it a lie? Of course not. It's not a lie… . She's not a liar, at least not about what happened here. And, if she's not a liar and if her information is good, here's your man. Id. 1036.

      Just over a month after Haskell's trial, Blue pled guilty in her Mercer County case to one count of retail theft and unsworn falsification. Before sentencing, Hayes sent the Court of Common Pleas of Mercer County a letter in which he explained that Blue gave "very important" testimony at Haskell's trial. Id. at 19.

      The Mercer County DA recommended a probationary sentence. Blue received a suspended sentence of one to four years in prison with 18 months of probation for the theft charge and a sentence of costs only on the unsworn falsification charge.

    3. Procedural History

      At trial Haskell was convicted of first-degree murder, unlawful carrying of a firearm, possessing an instrument of crime, aggravated assault, and reckless endangerment. He was sentenced to life imprisonment plus a 15-30 month period of incarceration consecutive to his life sentence.

      Haskell pursued his claim that Blue's testimony violated his right to due process in Pennsylvania's Post Conviction Relief Act ("PCRA") Court, which held that his perjured-testimony challenge was time barred. But once Haskell filed his habeas petition with the District Court under 28 U.S.C. § 2254, the Commonwealth expressly stated in its Answer that this claim was not procedurally defaulted and that the District Court must review it on the merits. Accordingly, it considered the merits, and because the PCRA Court had not reached them, it reviewed the claim de novo. On appeal, the Commonwealth makes no objection to the District Court's on-the-merits review of Haskell's perjured-testimony claim.

      The District Court held that Blue's testimony was false and that the prosecutor knew or should have known it was. However, it denied Haskell's request for relief because he failed to show that Blue's perjured testimony had a substantial and injurious effect or influence on the jury's verdict, which the Court believed he was required to demonstrate pursuant to Brecht, 507 U.S. at 627. Haskell requested, and we granted, a certificate of appealability.

  2. JURISDICTION AND STANDARD OF REVIEW

    The District Court had subject matter jurisdiction to consider Haskell's petition under 28 U.S.C § 2254, and we have jurisdiction to hear his appeal under 28 U.S.C. §§ 1291&2253. "Because the District Court did not hold an evidentiary hearing and, instead, based its decision on its review of the state court record, we apply a plenary standard of review of its decision and order." Branch v. Sweeney, 758 F.3d 226, 232 (3d Cir. 2014) (citing Duncan v. Morton, 256 F.3d 189, 196 (3d Cir. 2001)).

  3. ANALYSIS
    1. The Reasonable Likelihood Standard

      A state violates the Fourteenth Amendment's due process guarantee when it knowingly presents or fails to correct false testimony in a criminal proceeding. See Napue, 360 U.S. at 269; Giglio, 405 U.S. at 153; see also Lambert v. Blackwell, 387 F.3d 210, 242 (3d Cir. 2004). Consequently, "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 (1976), holding modified by United States v. Bagley, 473 U.S. 667 (1985).

      "[T]he same result obtains when the State, although not soliciting false evidence, allows it to go uncorrected when it appears." Giglio, 405 U.S. at 153 (quoting Napue, 360 U.S. at 269); see also Lambert, 387 F.3d at 242. A conviction must be set aside even if the false testimony goes only to a witness's credibility rather than the defendant's guilt. Napue, 360 U.S. at 270.

      The standard of review applicable to perjured testimony claims is "strict." Agurs, 427 U.S. at 104. This is so "not just because [those claims] involve prosecutorial misconduct, but more importantly because they involve a corruption of the truth-seeking function of the trial process." Id.

      Accordingly, in order to establish his claim, Haskell must show that (1) Blue committed perjury, (2) the Commonwealth knew or should have known that the testimony was false, (3) the false testimony was not corrected, and (4) there is a reasonable likelihood that the perjured testimony could have affected the judgment of the jury. Lambert, 387 F.3d at 242.

      Uncontested facts in the record demonstrate that Haskell has satisfied the first three elements. Blue lied when she testified-both at Haskell's preliminary hearing and at his trial-that she expected nothing in return for her testimony. She expected and eventually received favorable treatment at sentencing for her Mercer County charges. The Commonwealth, of course, knew Blue's testimony was false and failed to correct it. Turning to the final inquiry, we conclude that there is a reasonable likelihood that Blue's false testimony could have affected the jury's judgment. Blue was a key witness. All the other eyewitnesses had significant problems with their testimony. Mathis recanted on the witness stand and claimed that Haskell was not the shooter. Wayne and Roberts had both previously said that they would not be able to identify the shooter. Thus Blue, who claimed to have met Haskell before the shooting, provided strong evidence that Haskell was the shooter. As the Commonwealth put it in its closing argument at Haskell's trial, "this [is] valuable testimony." J.A. 1034. And the Commonwealth's decision to vouch for Blue's credibility only emphasizes her importance. Id. at 1036 ("It's not a lie… . She's not a liar, at least not about what happened here. And, if she's not a liar and if her information is good, here's your man.").

      Given her central role, knowledge of the benefit she received in exchange for her testimony-substantial help with her own pending criminal charges-poses a reasonable, and significant, likelihood of affecting the judgment of the jury. See Napue, 360 U.S. at 270 ("Had the jury been apprised of the true facts … it might well have concluded that [the witness] had fabricated testimony in order to curry the favor of the very representative of the State who was prosecuting the case in which [he] was testifying, for [he] might have believed that such a representative was in a position to implement (as he ultimately attempted to do) any promise of consideration.").

      Moreover, the facts of this case are in line with those in Napue and Giglio, the cases in which the Supreme Court initially articulated when false testimony requires relief. In Napue, as in Haskell's trial, a key witness falsely testified "that he had been promised no consideration for his testimony, and [] the Assistant State's Attorney handling the case had known this to be false." Napue, 360 U.S. at 267. And much like our case, Napue concerned the identification of assailants in a bar-room murder that the key witness was in a unique position to provide. Id. at 264 ("[The key witness's] testimony was extremely important because the passage of time and the dim light in the cocktail lounge made eyewitness identification very difficult and uncertain, and because some pertinent witnesses had left the state."). Finally, the benefit for which the Napue witness had bargained was the same as Blue's: "a recommendation for a reduction of his … sentence [in his own criminal matter] would be made and, if possible effectuated." Id. at 266.

      Giglio also involved a witness's false statement that he had been promised nothing in return for his testimony. 405 U.S. at 152. The witness there was promised immunity from prosecution by a government attorney but denied it on the stand. Id. And, like our case, the prosecution returned to and emphasized the false testimony in its summation of the case. Id. The Supreme Court emphasized that "whether the nondisclosure was a result of negligence or design, it is the responsibility of the prosecutor." Id. at 154. Recognizing the importance of this key witness's testimony, the Court found that his "credibility … 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. Because this fact was kept from the jury, due process required a new trial.

      Thus there is little doubt Haskell has met the standard set by Napue and Giglio.

    2. Brecht

      Meeting this standard, however, does not end our inquiry. Although the Supreme Court has held that the Giglio/Napue "materiality" standard discussed above is equivalent to the harmless-error standard articulated in Chapman v. California, 386 U.S. 18, 24 (1967) (requiring the beneficiary-the prosecution-of a constitutional error to demonstrate that it was harmless beyond a reasonable doubt), see Bagley, 473 U.S. at 680 n.93, the Commonwealth contends that Haskell must also meet the separate actual-prejudice standard of Brecht, 507 U.S. at 637. It held that when constitutional trial errors are raised in habeas proceedings, as opposed to on direct review, the petitioner is generally entitled to relief only if he can show "actual prejudice." Id. at 631. This occurs when the error had a "substantial and injurious effect or influence in determining the jury's verdict." Id. (quoting Kotteakos v. United States, 328 U.S. 750, 776 (1946)). "[I]f a judge has `grave doubt' about whether an error affected a jury in this way, the judge must treat the error as if it did so." O'Neal v. McAninch, 513 U.S. 432, 438 (1995) (internal quotation marks omitted).

      Brecht relied on three characteristics of habeas proceedings to ground the distinction between harmless error under Chapman applicable on direct review (putting the burden on the prosecution) and the heightened prejudice requirement it was announcing for habeas (that is, collateral) review (burdening the convicted petitioner seeking relief). First, "the State[] [has] [an] interest in the finality of convictions that have survived direct review within the state court system[,]" which review under Chapman's harmless-beyond-a-reasonable-doubt standard would undermine. Brecht, 507 U.S. at 635. Second, values of comity and federalism favor the Brecht standard because "[f]ederal intrusions into state criminal trials frustrate both the States' sovereign power to punish offenders and their good-faith attempts to honor constitutional rights." Id. (citing Engle v. Isaac, 456 U.S. 107, 128 (1982)). Third, "liberal 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. (internal citations, quotation marks, brackets, and ellipses omitted).

      But these concerns do not apply to all constitutional errors, and thus, there are a number of exceptions to Brecht's actual-prejudice requirement. The Court recognized in Brecht itself that structural constitutional errors, like denial of the right to counsel, are not subject to harmless-error review. Id. at 630. Moreover, it noted that, "in an unusual case, a deliberate and especially egregious error of the trial type, or one that is combined with a pattern of prosecutorial misconduct, might so infect the integrity of the proceeding as to warrant the grant of habeas relief, even if it did not substantially influence the jury's verdict." Id. at 638 n.9.

      The Court identified another exception in Kyles v. Whitley, 514 U.S. 419, 435 (1995), when it held that Brecht's standard does not apply when the state has violated Brady v. Maryland, 373 U.S. 83 (1963), by suppressing evidence favorable to the defendant. The Court explained that it "had previously rejected [substantial and injurious effect] as the standard governing constitutional disclosure claims[.]" Kyles, 514 U.S. at 436 (citing Agurs, 427 U.S. at 112). Once a petitioner demonstrates "`a reasonable probability that, had the evidence been disclosed to the defense, the result of the proceeding would have been different,'" the inquiry is over, and a petitioner meeting Brecht's substantial-and-injurious-effect (or actual-prejudice) standard is unnecessary. Id. at 435 (quoting Bagley, 473 U.S. at 682).

      Like the suppression of evidence, presentation of perjured testimony also violates Brady. See Agurs, 427 U.S. at 103 (describing three situations to which Brady applies: (1) the government's knowing presentation of or failure to correct false testimony; (2) its failure to provide requested exculpatory evidence; and (3) its failure to volunteer exculpatory evidence never requested). Cf. Bagley, 473 U.S. at 679 n.8 ("In fact, the Brady rule has its roots in a series of cases dealing with convictions based on the prosecution's knowing use of perjured testimony."). However, the Court in Kyles, a habeas case, explicitly declined to consider whether Brecht applies to collateral review of convictions tied to knowing use of perjured testimony. Kyles, 514 U.S. at 433 n.7 ("[W]e do not consider the question whether Kyles's conviction was obtained by the knowing use of perjured testimony and our decision today does not address [Brecht's applicability to] any claim under the first Agurs category.").

      Our Court also has not resolved whether Brecht applies to cases like the one before us. In a case preceding Kyles, we held that the Brecht standard does indeed apply to habeas petitions alleging a prosecutor's knowing use of perjured testimony. Robinson v. Arvonio, 27 F.3d 877, 885 (3d Cir. 1994), cert. granted, judgment vacated, 513 U.S. 1186 (1995). But the Supreme Court vacated our judgment in Robinson and remanded the case for further consideration in light of O'Neal, 513 U.S. at 438, which clarified the Brecht standard as requiring relief when "a judge has grave doubt about whether an error" "had substantial and injurious effect or influence upon the jury[.]"

      We, in turn, remanded the case back to the District Court and did not hear a subsequent appeal. See Order, Robinson v. Arvonio, et al., Case No. 92-5667 (Oct. 10, 1995). Because an order of the Supreme Court "vacating the judgment of the Court of Appeals deprives that court's opinion of precedential effect[,]" County of Los Angeles v. Davis, 440 U.S. 625, 634 n.6 (1979) (quoting O'Connor v. Donaldson, 422 U.S. 563, 577-78 n.12 (1975)), we have explicitly recognized that Robinson is no longer binding precedent. Hassine v. Zimmerman, 160 F.3d 941, 960 n.30 (3d Cir. 1998).

      Thereafter we have discussed (without mention of Brecht) the proper standard to apply to habeas petitions involving perjured-testimony claims. In Lambert, 387 F.3d at 242, we noted that when "the prosecution's case includes perjured testimony and the prosecution knew, or should have known, of the perjury … [or] when the government, although not soliciting false evidence, allows it to go uncorrected when it appears at trial[,] … the conviction must be set aside if there is any reasonable likelihood that the false testimony could have affected the judgment of the jury." We went on to quote Bagley's statement that although this rule is stated in terms that treat the knowing use of perjured testimony as error subject to harmless error review, it 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. Lambert, 387 F.3d at 242 (quoting Bagley, 473 U.S. at 679). The implication is that, for perjured-testimony claims raised in habeas proceedings, the materiality and harmless-error standards are one and the same.

      Accordingly, Brecht's standard would not apply. But, as noted, Lambert made no reference to Brecht. Perhaps this was because it concluded that the statements cited were not perjured. See id. at 243, 245, 252 & 266. In any event, Lambert does not resolve by a holding whether habeas petitioners must meet Brecht's actual-prejudice hurdle.

      Our sister Circuits are split on the question. Relying on Kyles, the Ninth Circuit has rejected application of Brecht to perjured-testimony cases. Hayes v. Brown, 399 F.3d 972, 984 (9th Cir. 2005) ("Even though this case comes to us on habeas review, we do not conduct an additional harmless error analysis under Brecht[.]'"). Analogizing to claims brought under Strickland v. Washington, 466 U.S. 668 (1984), the Ninth Circuit noted that "federal courts do not conduct a separate Brecht analysis in ineffective assistance of counsel claims." Id. at 985. Whenever the applicable test is "derived from the Agurs materiality standard" (e.g., claims involving suppression of evidence, ineffective assistance of counsel, or perjured testimony), Brecht does not apply. Id. ("When the Supreme Court has declared a materiality standard, as it has for this type of constitutional error, there is no need to conduct a separate harmless error analysis.").

      The First, Sixth, Eighth, and Eleventh Circuits have disagreed, applying Brecht to habeas petitions raising perjured testimony claims. See Gilday v. Callahan, 59 F.3d 257, 268 (1st Cir. 1995); Rosencrantz v. Lafler, 568 F.3d 577, 587-90 (6th Cir. 2009); United States v. Clay, 720 F.3d 1021, 1026-27 (8th Cir. 2013); Trepal v. Sec'y, Florida Dep't of Corr., 684 F.3d 1088, 1111-13 (11th Cir. 2012).

      The First Circuit, which issued its decision prior to the Ninth Circuit's ruling in Hayes, reasoned that "[a]pplying th[e] [reasonable-likelihood] standard in most cases involving perjury or its equivalent will likely result in a finding of constitutional error." Gilday, 59 F.3d at 268. "Scaling that lower materiality hurdle, however, still will leave the petitioner facing the Brecht harmless error inquiry into whether the perjured testimony in fact had a substantial and injurious effect or influence on the jury's verdict." Id. Brecht does not apply to evidentiary-withholding claims, by contrast, because those claims already require "a court to find an impact on the jury verdict sufficiently substantial to satisfy the Brecht harmless error test." Id.

      The Sixth Circuit agreed with the First and rejected the Ninth Circuit's intervening opinion, asserting that the Hayes Court "erred in failing to distinguish false-testimony claims from Brady withholding claims." Rosencrantz, 568 F.3d at 590. The Eleventh Circuit followed suit three years later. Trepal, 684 F.3d at 1113 (holding Brecht applied because "the more lenient Giglio materiality standard leaves room for the possibility that perjured testimony may be material under Giglio but still be harmless under Brecht").

      Four years after Rosencrantz, the Eighth Circuit joined the majority, applying Brecht because "the materiality standard for false testimony is lower, more favorable to the defendant, and hostile to the prosecution as compared to the standard for a general Brady withholding violation." Clay, 720 F.3d at 1026 (internal quotation marks and citation omitted).

      We favor-and therefore adopt-the Ninth Circuit's approach. As that Court recognized, Kyles suggests that for the three types of due-process violations discussed in Agurs there is no need to perform a separate harmless-error analysis under Brecht. Hayes, 399 F.3d at 985 (citing Kyles, 514 U.S. at 436). This is because for these violations the materiality and harmless-error standards merge. See Bagley, 473 U.S. at 678-80. That is, the test for materiality supplies the test for harmlessness and there is no need to look to Brecht to supply a harmless-error standard. Hayes, 399 F.3d at 985.

      More importantly, the concerns behind Brecht do not reach claims of perjured testimony presented by the state. To repeat, the Supreme Court's imposition of Brecht's harmless error standard was motivated by three concerns: (1) "the State[] [has] [an] interest in the finality of convictions that have survived direct review within the state court system"; (2) "[f]ederal intrusions into state criminal trials frustrate both the States' sovereign power to punish offenders and their good-faith attempts to honor constitutional rights"; and (3) "liberal 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[.]" Brecht, 507 U.S. at 635 (internal citations, quotation marks, brackets, and ellipses omitted). These are weighty interests no doubt, but they do not reach the facts before us.

      The Supreme Court has long counseled that "a deliberate deception of court and jury by the presentation of testimony known to be perjured … is [] inconsistent with the rudimentary demands of justice[.]" Mooney v. Holohan, 294 U.S. 103, 112 (1935). Put differently, "[it is a] well-established rule that a conviction obtained by the knowing use of perjured testimony is fundamentally unfair[.]" Bagley, 473 U.S. at 678-79. In Brecht itself the Court recognized "the writ of habeas corpus has historically been regarded as an extraordinary remedy, a bulwark against convictions that violate fundamental fairness." 507 U.S. at 633 (internal quotation marks omitted). Thus it is difficult to see how concerns of finality would trump rudimentary demands of justice and fundamental fairness when those are precisely the values the writ of habeas corpus is intended to protect.

      Second, when the state knowingly presents perjured testimony, we are not presented with a "good-faith attempt[] to honor constitutional rights," Id. at 635, but instead with a bad-faith effort to deprive the defendant of his right to due process and obtain a conviction through deceit. After all, courts apply Napue's "strict standard of materiality" to perjured-testimony cases "not just because they involve prosecutorial misconduct, but more importantly because they involve a corruption of the truth-seeking function of the trial process" by the state itself. Agurs, 427 U.S. at 104.

      Third, there is little chance that excluding perjured testimony claims from Brecht analysis will "degrade[] the prominence of the trial itself[.]" Brecht, 507 U.S. at 635. A defendant will usually be unable to litigate his claims of perjured testimony at "the trial itself" because the trial is where the perjury occurs. And it is possible, even likely, that petitioners will not know of the prosecution's use of perjured testimony until after the opportunity for direct review has passed.

      Finally, the First and Sixth Circuits note that, without Brecht review, perjured testimony faces a lower bar than suppression claims. Gilday, 59 F.3d at 268; Clay, 720 F.3d at 1026. But to us that seems to be a feature, not a bug. If suppression of evidence (and thereby, the truth) is a serious constitutional error, its fabrication is a greater error still. That is why the Supreme Court set out differing materiality standards for the three types of error that implicate Brady: (1) the government's knowing presentation of or failure to correct false testimony, (2) its failure to provide requested exculpatory evidence, and (3) its failure to volunteer exculpatory evidence never requested. See Agurs, 427 U.S. at 103-06. Presenting false testimony cuts to the core of a defendant's right to due process. It thus makes sense that "the materiality standard for false testimony is lower, more favorable to the defendant, and hostile to the prosecution as compared to the standard for a general Brady withholding violation." Clay, 720 F.3d at 1026.

      At root is how can a defendant possibly enjoy his right to a fair trial when the state is willing to present (or fails to correct) lies told by its own witness and then vouches for and relies on that witness's supposed honesty in its closing? As the Supreme Court recited in Napue, [i]t is of no consequence that the falsehood bore upon the witness' credibility rather than directly upon defendant's guilt. A lie is a lie, no matter what its subject, and, if it is in any way relevant to the case, the district attorney has the responsibility and duty to correct what he knows to be false and elicit the truth. 360 U.S. at 269-70 (quoting People v. Savvides, 136 N.E.2d 853, 854-55 (N.Y. 1956)) (internal ellipses omitted).

      For these reasons, we hold that the actual-prejudice standard of Brecht does not apply to claims on habeas that the state has knowingly presented or knowingly failed to correct perjured testimony. A reasonable likelihood that the perjured testimony affected the judgment of the jury is all that is required.

      Haskell has demonstrated that there is a reasonable likelihood that Blue's false testimony could have affected the judgment of the jury. Hence he is entitled to relief. He need not go on to show that this error had a substantial and injurious effect or influence in determining the jury's verdict because, when the state has corrupted the truth-seeking function of the trial by knowingly presenting or failing to correct perjured testimony, the threat to a defendant's right to due process is at its apex and the state's interests are at their nadir. Accordingly, we grant Haskell's habeas petition and remand for further proceedings consistent with this opinion.

1 The Commonwealth presented a great deal of testimony to rehabilitate Mathis's videotaped testimony. Among the more important details of his initial story corroborated by other witnesses were that Mathis and the shooter left the bar together, that a gun recovered from under an old truck left for many years in the alley near Jethroe's resembled the gun used by the shooter, and that Mathis and another man went to the home of Felicia Clark's neighbor in Franklin Terrace after the shooting.

2 I am aware that Ms. Blue faces a misdemeanor retail theft charge in Mercer County. I spoke with the prosecutor in that case and he explained he had already arrived at a Plea Agreement in her case… . I also explained that Ms. Blue was assisting in this prosecution. He indicated to me that this assistance would not alter his approach to his prosecution. He indicated he would make the assistance known at the time of her sentencing in Mercer County… . The only understanding I am aware of is for Ms. Blue's cooperation. We would make the sentencing Judge aware of this cooperation.
J.A. 1720 (Letter From Hayes to Haskell's Defense Counsel dated April 30, 1998).

3 [It is a] well-established rule 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. Although this rule is stated in terms that treat the knowing use of perjured testimony as error subject to harmless-error review, it 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. The Court in Agurs justified this standard of materiality on the ground that the knowing use of perjured testimony involves prosecutorial misconduct, and more importantly, involves a corruption of the truth-seeking function of the trial process.
Bagley, 473 U.S. at 678-80 (internal citations and quotation marks omitted).

Print | | Share Stumble Upon Facebook Delicious Digg Reddit Google

Add Comment

Newsletter Sign Up

Get the latest info delivered right to your inbox. Enter your email address below to subscribe.

Become a Contributor

You can submit your own articles to be considered for publication on Upon Further Review. LEARN MORE