Richard M. Arnold v. Reed A. Richardson

In the

United States Court of Appeals For the Seventh Circuit No. 20‐2701

RICHARD M. ARNOLD, Petitioner‐Appellant,

v.

REED A. RICHARDSON, Warden, Respondent‐Appellee.

Appeal from the United States District Court for the Eastern District of Wisconsin. No. 2:15‐cv‐01524‐NJ — Nancy Joseph, Magistrate Judge.

ARGUED APRIL 15, 2021 — DECIDED OCTOBER 6, 2021

Before KANNE, ROVNER, and HAMILTON, Circuit Judges. ROVNER, Circuit Judge. This case returns to us after we remanded to the district court for an evidentiary hearing on whether the petitioner could overcome the one‐year time bar to filing his petition for a writ of habeas corpus, under the actual innocence exception. After holding the hearing, the district court determined that Arnold failed to meet the rigorous

2 No. 20‐2701

standard for overcoming the time bar set in Schlup v. Delo, 513 U.S. 298 (1995), and dismissed the petition. We affirm. I. We assume familiarity with our prior opinion and focus only on the facts and procedural history relevant to the issues in this appeal. See Arnold v. Dittman, 901 F.3d 830 (7th Cir. 2018) (“Arnold I”). In 2008, a Wisconsin jury convicted Arnold of repeated sexual assault of the same child, his son, M.A., who was the principal witness against him at trial. Because Arnold was a persistent repeater, the court was required to sentence him to life in prison without the possibility of parole. The judgment was entered on August 12, 2008. In October 2011, the Wisconsin Court of Appeals affirmed both the conviction and the denial of Arnold’s first request for post‐conviction relief. The Wisconsin Supreme Court subsequently denied Arnold’s petition for review. In November 2011, after the Wisconsin Court of Appeals had entered its ruling denying Arnold any relief, M.A. signed a notarized affidavit in which he recanted the substance of his trial testimony. M.A. explained that, at the time he accused his father of sexual abuse, he had been under the supervision of a juvenile court and was participating in a program (hereafter “Program”) for children who sexually abused other children. He claimed that his trial testimony had been false, and that he had felt pressured to accuse his father in order to placate the person running that Program. Nearly two years after M.A. signed the affidavit, Arnold filed a pro se petition in state court seeking a new trial on the basis of his son’s recantation. The Wisconsin circuit court denied the petition without a hearing,

No. 20‐2701 3

and the Wisconsin Court of Appeals affirmed. No hearing was required, the appellate court concluded, because the motion was legally insufficient. The appellate court held that the affidavit was not newly discovered evidence but rather was cumulative of evidence presented at trial that M.A. had told others that his father had not assaulted him. The Wisconsin Supreme Court again denied review. On December 21, 2015, Arnold filed a federal petition for a writ of habeas corpus, pursuant to 28 U.S.C. § 2254. His pro se petition was founded on due process and a claim for actual innocence. On the State’s motion, the district court dismissed the petition as untimely. Under 28 U.S.C. § 2244(d)(1)(A), a petitioner must seek habeas relief within one year of the date that his conviction became final. Arnold’s conviction became final on April 23, 2012, and so he had until April 24, 2013 to file his petition, missing the deadline by more than two years. The district court considered whether Arnold’s claim for actual innocence could excuse the running of the limitations period, and concluded that it did not. See Schlup, 513 U.S. at 327 (noting that a petitioner must establish that it is more likely than not that no reasonable juror would have convicted him in light of the new evidence of his innocence in order to overcome the time bar). The district court reasoned that the state court had already conducted the Schlup analysis and concluded that the recantation affidavit would not have affected the jury’s verdict. On appeal, we vacated that ruling and remanded for a hearing. We noted that Arnold had filed his petition beyond the section 2244(d)(1) one‐year time limit, rendering his petition untimely unless he could establish that he qualified for

4 No. 20‐2701

an exception to the time bar. Arnold relied on his actual innocence claim to overcome the time bar, and the framework for evaluating that claim was established by Schlup: A claim of actual innocence must be both credible and founded on new evidence. Schlup, 513 U.S. at 324, 115 S.Ct. at 865. To be credible, the claim must have the support of “reliable evidence—whether it be exculpatory scientific evidence, trustworthy eyewitness accounts, or critical physical evidence.” Ibid. That evidence must also be new in the sense that it was not before the trier of fact. Ibid.; Gladney, 799 F.3d at 896, 898. The petitioner’s burden is to show that, in light of this new evidence, it is more likely than not that no reasonable juror would have found him guilty beyond a reasonable doubt. Schlup, 513 U.S. at 327, 115 S.Ct. at 867; see also id. at 329, 115 S.Ct. at 868. In evaluating the claim, the court is to conduct a comprehensive assessment that takes into account any reliable evidence probative of peti‐ tioner’s innocence or guilt, even evidence that was previously excluded; the court is not bound by the rules of evidence that would govern at trial. Id. at 327–28, 115 S.Ct. at 867. It is not the court’s role to determine independently what the petitioner likely did or did not do; rather, its task is to assess the likely impact of the new evidence on reasonable jurors. Id. at 329, 115 S.Ct. at 868. Although any delay or lack of diligence by the petitioner in pursu‐ ing his claim of actual innocence is not a bar to the claim, it is among the factors that the court may

No. 20‐2701 5

consider in assessing the merits of the claim. McQuiggin, 569 U.S. at 388–400, 133 S.Ct. at 1935–36. Arnold I, 901 F.3d at 836–37. Arnold presented a plausible claim of actual innocence because the principal witness against him recanted; this evidence was new in the sense that it was not before the jury that convicted him. But no court had ever considered the credibility and reliability of M.A.’s recantation, and then applied the Schlup analysis to assess the likely impact of the new evidence on reasonable jurors. We therefore remanded “the case to the district court for an evidentiary hearing at which the credibility of M.A.’s recantation can be assessed along with the probable impact that the recantation would have had on reasonable jurors. As Schlup makes clear, any reliable evidence bearing on the veracity of the recantation and on Arnold’s guilt or innocence may be considered in making these assessments.” Arnold I, 901 F.3d at 842. On remand, the district court held an evidentiary hearing over the course of two days, hearing testimony from M.A. himself, who was twenty‐eight years old at the time of the hearing; Dr. Mark Goldstein, a child psychologist and expert retained by Arnold on the subject of child sexual abuse; and Karen B., M.A.’s former counselor in the Program. Consistent with his affidavit, M.A. denied that his father assaulted him, and testified that he made the accusations at the time because he believed that was what his counselor wanted to hear. He asserted that his counselor had great power over his life and could (and did) arrange for him to be taken into custody for trivial reasons. He believed that she could cause revocation of his judicial supervision and that he would then go to prison for five years, and so he said the things that he believed she

6 No. 20‐2701

wanted him to say. He also claimed to have felt pressure from his father’s probation officer to accuse his father of abuse. Finally, he was angry at his father at the time for failing to give him attention. He did not know at the time he made the accusations that his father might face life in prison, and after getting his own life in order, wanted to correct what he had done in his father’s case. He explained that his father’s ex‐wife, Randi Shaw, drafted the affidavit based on conversations he had with her, and that she had driven him to a notary with two witnesses in order to sign the affidavit. Karen B., M.A.’s former counselor in the Program, recalled M.A.’s name but did not recall working with him. She ex‐ plained that she was a social worker and coordinator of the Program. In that role, she provided group and individual treatment to troubled juveniles. She did not seek out evidence of Program rules violations but did pass on some information about violations to the juveniles’ social workers in certain circumstances. She testified that she did not have authority to revoke supervision and that she worked to develop a therapeu‐ tic alliance with the juveniles in the Program. If she suspected that a juvenile had been a victim of a sexual assault, she would approach the subject “gingerly,” and generally waited for the child to raise it. She testified that she would not push a child to disclose abuse and denied that she would ever suggest to a child that a particular person had victimized the child. Finally, Dr. Goldstein testified generally about the statistics for children and adolescents who make false reports of sexual assault. Dr. Goldstein asserted that adolescents were more likely to falsely report sexual abuse than young children. He noted that many false accusations were tied to custody

No. 20‐2701 7

disputes, and he listed characteristics common to false reports as well as factors that might motivate a child to falsely report abuse. He explained that there were no studies on adults recanting accusations that they made when they were adoles‐ cents, and he described the reasons why people sometimes recanted. He had not been retained to offer an opinion on M.A.’s credibility and declined to offer an opinion. He pointed out aspects of M.A.’s recantation that were consistent with the research on false allegations, and also factors that could be consistent with the truth of the initial accusations. In an exhaustive and careful opinion, the district court assessed all of the evidence presented at Arnold’s original trial and at the evidentiary hearing. Applying the Schlup standard, the court found that the evidence from the hearing was new, and then assessed whether it was reliable and “whether it is more likely than not that no reasonable juror would convict Arnold in light of the totality of the evidence.” Arnold v. Richardson, 2020 WL 4569162, *22 (E.D. Wis. August 7, 2020) (“Arnold II”). The court began by noting several features of M.A.’s recantation that undermined its reliability, including that: (1) it was generated three years after the jury trial, which itself was three or four years after the claimed abuse, far from the time of the actual events; (2) M.A. had appeared at a court hearing during his father’s direct appeal only eight months before he signed the affidavit, and had opportunities to speak to the judge, the prosecutor and defense counsel, but he did not profess his father’s innocence to anyone and did not sign the affidavit until after his father lost his appeal; and (3) M.A. himself never initiated steps to exonerate his father, instead admitting that the recantation affidavit was the product of

8 No. 20‐2701

Randi Shaw’s initiative. The court also noted certain inconsis‐ tencies in M.A.’s explanation for the delay in generating the recantation affidavit, and noted possible motives for M.A. to falsely recant that were consistent with expert testimony at trial and at the evidentiary hearing. In analyzing each of these features and factors, the court noted that a reasonable juror could credit M.A.’s original trial testimony as more accurate than the later recantation, that a reasonable juror could question why M.A. did not exonerate his father when he had an opportunity to do so during the appeal process, and that a reasonable juror could conclude that if M.A. actually fabricated the original allegations and wanted to exonerate his father, he would have initiated the affidavit himself and then followed through with it instead of handing it off to his father’s ex‐wife. The court framed the analysis in terms of what a reasonable juror could conclude based on the evidence. The court then considered the totality of the evidence and found that, even in light of the new evidence, a reasonable juror was likely to find M.A.’s original trial testimony credible. Again, the analysis was exhaustive and included evidence both from the original trial and the evidentiary hearing as well as the affidavit itself. The court also acknowledged that, in light of the entire record, it is possible that a reasonable juror could acquit Arnold. But the court correctly noted that this possibility was not sufficient to meet the Schlup standard, which required Arnold to “show that it is more likely than not that no reason‐ able juror could convict him.” Arnold II, 2020 WL 4569162, *25 (E.D. Wis. August 7, 2020). Because Arnold had not met that burden, the court dismissed the petition as untimely. Id.

No. 20‐2701 9

II. On appeal, Arnold faults the district court for failing to make the credibility determinations that he claims we ordered on remand. In particular, he contends that the district court failed to determine whether M.A.’s recantation was reliable and credible. He asserts that the court distorted the standard by instead considering whether a reasonable juror could find M.A.’s recantation credible. Arnold argues that without a finding as to whether the court itself found M.A.’s recantation credible, there is no way for the district court or this court to decide the reliability of the recantation in the eyes of a reason‐ able juror. Arnold summed up his criticism of the district court’s ruling in his Reply Brief: [R]ather than making a credibility determination, the lower court talked about credibility in probabil‐ istic terms, saying repeatedly that a reasonable juror “could question why …” or “could conclude that …” or “[i]t would not be a bridge too far to conclude that …” or “[t]his would likely indicate to a reason‐ able juror that … .” (Id.) And yet, to probabilistically state what a jury could have thought about M.A.’s recantation does not tell us whether the court found that recantation to be credible. It is, instead, the application of the actual innocence test to Arnold’s case with an undecided but necessary factual predicate. Reply Brief, at 5–6.

10 No. 20‐2701

We review de novo the district court’s decision to dismiss Arnold’s petition for habeas relief as untimely. Simms v. Acevedo, 595 F.3d 774, 777 (7th Cir. 2010). When a district court has held an evidentiary hearing in a habeas proceeding, we review its factual findings for clear error. Pidgeon v. Smith, 785 F.3d 1165, 1171 (7th Cir. 2015). As we noted above, we re‐ manded “to the district court for an evidentiary hearing at which the credibility of M.A.’s recantation can be assessed along with the probable impact that the recantation would have had on reasonable jurors.” Arnold I, 901 F.3d at 842. Arnold’s primary complaint is that the district court failed to state whether the court itself found M.A.’s recantation credible and instead engaged in a probabilistic determination of whether a jury would have found it credible. But we made clear in our original order that “[i]t is not the court’s role to determine independently what the petitioner likely did or did not do[.]” Arnold I, 901 F.3d at 837. Rather, Arnold was required to demonstrate on remand that “it is more likely than not that no reasonable juror would have found him guilty beyond a reasonable doubt,” an inherently probabilistic determination. Arnold I, 901 F.3d at 837. See also Schlup, 513 U.S. at 330 (noting that the “more likely than not” standard reflects the probabilistic nature of the court’s inquiry). And that is exactly what the district court did, consistent with our remand order and with Schlup’s framework for evaluating a claim of actual innocence as a gateway to review of a habeas claim that would otherwise be foreclosed by untimeliness: The meaning of actual innocence … does not merely require a showing that a reasonable doubt exists in the light of the new evidence, but rather that no

No. 20‐2701 11

reasonable juror would have found the defendant guilty. It is not the district court’s independent judgment as to whether reasonable doubt exists that the standard addresses; rather the standard requires the district court to make a probabilistic determina‐ tion about what reasonable, properly instructed jurors would do. Thus, a petitioner does not meet the threshold requirement unless he persuades the district court that, in light of the new evidence, no juror, acting reasonably, would have voted to find him guilty beyond a reasonable doubt. Schlup, 513 U.S. at 329. Our recitation of the district court’s analysis demonstrates that the court properly applied that probabilistic standard on remand and concluded that Arnold fell short. It was not a part of the court’s charge on remand to independently determine whether it found M.A.’s recantation credible or reliable. The court employed the proper standard in viewing the new evidence through the eyes of a reasonable juror. The court made no error of law or fact in reaching its determination, and correctly dismissed the petition as time‐barred. AFFIRMED.