NOT RECOMMENDED FOR PUBLICATION File Name: 20a0230n.06
No. 19-3067
UNITED STATES COURT OF APPEALS FOR THE SIXTH CIRCUIT
JAMES RUSSELL, )
FILED
) Apr 28, 2020 Petitioner-Appellee, ) DEBORAH S. HUNT, Clerk )
v. ) ON APPEAL FROM THE UNITED ) STATES DISTRICT COURT FOR THE LYNEAL WAINWRIGHT, Warden, ) SOUTHERN DISTRICT OF OHIO )
Respondent-Appellant. )
BEFORE: DAUGHTREY, CLAY, and GRIFFIN, Circuit Judges.
MARTHA CRAIG DAUGHTREY, Circuit Judge. Following his initial state trial in 2005 and a retrial in 2010, James Russell appealed his convictions for aggravated robbery and murder, which resulted from his shooting Phillip Troutwine in an effort to take the victim’s wallet. On direct appeal before the Ohio Court of Appeals in October 2010, Russell contended that the trial court erred in failing to merge his sentences for aggravated robbery and murder. The appeals court rejected this claim, reasoning that under the then-current state of law laid out in State v. Rance, the crimes were of “dissimilar import” because of the legislature’s intent in punishing them as separate crimes and, therefore, not subject to merger. But the appeals court did order a remand to accommodate a Batson hearing. By the time of Russell’s next appeal, Rance had been overturned by State v. Johnson, a case that rejected analysis of crimes in the abstract for the purpose of merging sentences. However, Russell’s appellate counsel, the same one who had represented him on Russell’s previous appeal, failed to re-raise the sentencing merger issue before
2
the Ohio Court of Appeals. After unsuccessful efforts to reopen the appeal to add the overlooked sentencing issue, Russell filed a habeas corpus petition in federal court, which eventually granted him relief conditioned upon a ruling by the Ohio Court of Appeals on Russell’s claim that his appellate counsel’s failure to re-raise the sentencing merger argument after the change in law constituted ineffective assistance of counsel.
The Warden now appeals the district court’s grant of a conditional writ on this issue, primarily on the ground that the law has developed since the counsel’s failure to re-raise the issue to such an extent that Russell would be unsuccessful in securing a merger of his sentences under the current state of the law. We disagree and affirm the judgment of the district court.
I. BACKGROUND
This case involves a robbery gone wrong. Candace Hargrove, James Russell’s girlfriend, arranged to have sex with Philip Troutwine for money. Prior to Troutwine’s arrival at their residence, however, Hargrove changed her mind, and she and Russell planned to rob him instead. Russell used a gun to threaten Troutwine and demanded money from him. Hargrove, from the next room, heard a struggle between Troutwine and Russell and then heard a gunshot, at which point Russell emerged from the kitchen saying, “Oh my God, I shot him. I didn’t mean to shoot him. I didn’t mean to shoot him.” Russell and Hargrove then used a tent to wrap up the victim’s dead body, put the body in the back of the victim’s car, and drove the car away, eventually leaving it in a parking lot. Police discovered it there three weeks later. Authorities eventually traced Russell and Hargrove to California, where they were arrested.
Russell was indicted on multiple counts, and a jury convicted him on five: aggravated robbery, felony murder, tampering with evidence, grand theft of a motor vehicle, and gross abuse of a corpse. He was sentenced to a total of 40 years to life in prison, which included consecutive
3
sentences of ten years for aggravated robbery and 15 years to life for felony murder. He also was ordered to pay $15,498.25 in restitution to Troutwine’s wife for economic loss.
With the assistance of counsel, Russell appealed his convictions on several grounds, including a Batson1 issue and the trial court’s failure to merge his aggravated robbery and murder convictions under Ohio Revised Code § 2941.25. The Ohio Court of Appeals found that “the trial court erred in its treatment of the Batson issue” and remanded the case for a Batson analysis, which was unsuccessful. State v. Russell, No. 23454, 2010 WL 3835645, at *4 (Ohio Ct. App. 2d Dist. 2010) (Russell I). On the issue of sentence merger, the court relied on the prevailing standard at that time, State v. Rance, 710 N.E.2d 699 (Ohio 1999), and evaluated “whether the legislature intended to permit the imposition of multiple punishments for conduct that constitutes multiple criminal offenses.” Russell I, 2010 WL 3835645, at *7. It “compar[ed] the elements of the two offenses in the abstract” and found that “commission of neither offense necessarily results in commission of the other.” Id. at *8. As a result, the court overruled this assignment of error.
Two months later, in December 2010, the Supreme Court of Ohio overruled Rance, the case upon which the Ohio Court of Appeals had relied to deny Russell’s claim that the court erred in failing to merge his sentences. State v. Johnson, 942 N.E.2d 1061, 1069 (Ohio 2010). The Johnson standard required courts to focus on the facts of each specific case rather than evaluate the crimes in the abstract, as the appeals court had done in Russell’s appeal. See id. at 1069–70.
In November 2011, Russell again appealed his conviction on Batson grounds, represented by the same appellate counsel as in 2010. State v. Russell, No. 24443, 2012 WL 368135 (Ohio Ct. App. 2d Dist. 2012) (Russell II). But this time, his attorney did not raise the merger issue. The court remanded Russell’s case for another Batson hearing, id. at *9, but once again the trial court
1 Batson v. Kentucky, 476 U.S. 79 (1986).
4
did not decide the issue in his favor. In 2012, Russell petitioned pro se to reopen his prior direct appeal, contending that his appellate counsel was ineffective for failing to raise additional challenges, including the merger issue. In his application for re-opening, Russell noted that the “Supreme Court of Ohio ha[d] recently overruled its prior judgments in this area of the law, and it articulated the proper analysis for determining whether merger is appropriate.” The Court of Appeals denied the application to re-open, finding that because the merger issue was raised and overruled, it was “barred by the law of this case.” The court did not acknowledge the change in law. Russell appealed again in 2013 and 2014, but neither petition was successful. The Supreme Court of Ohio declined all appeals.
Then, in March 2015, the Ohio Supreme Court decided State v. Ruff, which modified the Johnson approach to determining whether sentences should merge. State v. Ruff, 34 N.E.3d 892, 897 (Ohio 2015); see also State v. Earley, 49 N.E.3d 266 (Ohio 2015). In this appeal, the Warden insists that Ruff and Earley “abrogated” Johnson, nullifying any possibility of relief for Russell. Russell insists that no abrogation resulted, thereby narrowing the dispositive issue now before us.
In September 2015, Russell filed a habeas petition in federal court, raising seven grounds for relief, including ineffective assistance of appellate counsel. The district court dismissed the petition with prejudice, granting a certificate of appealability (COA) on only the Batson claim. Russell v. Marion Corr. Inst., No. 3:15–cv–331, 2016 WL 4440323, at *1 (S.D. Ohio Aug. 23, 2016) (Russell III).
On appeal to the Sixth Circuit, we originally denied Russell’s application for an expanded COA but granted his motion for the appointment of counsel. Two weeks after Russell obtained new counsel, he petitioned for rehearing, clarifying his arguments regarding his appellate counsel’s failure to reassert his merger argument on appeal and asking us to revisit our prior denial of an
5
expanded COA. In light of the clarification provided, we found that “reasonable jurists could debate whether [Russell’s] appellate counsel was ineffective for failing to re-raise his merger argument,” due to the new standard used to assess whether sentences should merge under Johnson. We therefore granted his petition and expanded Russell’s COA to cover his ineffective-assistance claim related to his appellate counsel’s failure to re-raise his merger argument. In the end, we held that “the district court erroneously dismissed Russell’s ineffective-assistance claim as procedurally barred,” and—because no federal court had yet heard this issue on the merits—we remanded the matter to the district court for a substantive ruling. Russell v. Bunting, 722 F. App’x 539, 551 (6th Cir. 2018) (Russell IV). However, we affirmed the district court’s denial of relief on the Batson claim. Id. at 546–50.
Upon remand, the district court found that there “is a reasonable probability that the [Ohio Court of Appeals] might grant relief on Russell’s claim under Ohio Revised Code § 2941.25” and granted a writ of habeas corpus conditioned on Russell being allowed to file a new appeal in state court. Russell v. Bunting, No. 3:15-cv-331, 2019 WL 120974, at *1 (S.D. Ohio, Jan. 7, 2019) (Russell V). The Warden timely filed an appeal from this judgment.
II. DISCUSSION
We review habeas corpus decisions of a district court under a mixed standard, reviewing legal conclusions de novo and factual findings for clear error. Kelly v. Lazaroff, 846 F.3d 819, 827 (6th Cir. 2017).
When a habeas claim has been presented and adjudicated properly on the merits before the state court, then a heightened standard of review is required by AEDPA, 28 U.S.C. § 2254. Kelly, 846 F.3d at 831. The Warden argues that although Russell’s ineffective-assistance claim was not adjudicated on the merits in state court, AEDPA should apply because the state court addressed
6
the underlying issue of whether Russell’s offenses should have merged. To the contrary, however, in expanding the COA on this issue, we have already determined that the Ohio Court of Appeals never addressed this claim on the merits and, therefore, that it is not subject to AEDPA deference. See Russell v. Bunting, No. 16-4022 (6th Cir. May 30, 2017) (order); see also Russell IV, 722 F. App’x at 550 (finding that Russell’s ineffective-assistance claim is not procedurally defaulted).
A. Evaluating Ineffective Assistance of Counsel: Strickland v. Washington “On the first appeal of right, a defendant is entitled to effective assistance of appellate counsel,” and we evaluate claims of ineffective assistance of appellate counsel under the Strickland standard. Goff v. Bagley, 601 F.3d 445, 462 (6th Cir. 2010) (citing Mahdi v. Bagley, 522 F.3d 631, 636 (6th Cir. 2008)); Strickland v. Washington, 466 U.S. 668 (1984). The standard “requires that the [defendant] affirmatively establish (1) that counsel’s performance was objectively deficient; and (2) prejudice, which means that there is a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different.” Id.; Goff, 601 F.3d at 462 (quoting Mahdi, 522 F.3d at 636).
1. Deficient Performance “To establish that counsel was deficient, ‘the defendant must show that counsel’s representation fell below an objective standard of reasonableness.’” Smith v. Mitchell, 567 F.3d 246, 257 (6th Cir. 2009) (quoting Strickland, 466 U.S. at 688). Under the first prong of Strickland, “[c]ounsel’s performance is objectively unreasonable only where ‘the identified acts or omissions were outside the wide range of professionally competent assistance,’ as determined by ‘prevailing professional norms.’” United States v. Munoz, 605 F.3d 359, 376 (6th Cir. 2010) (quoting Strickland, 466 U.S. at 690). When evaluating reasonable competence, we assess whether the omitted issues were “significant and obvious,” whether there was contrary authority on the issue,
7
and whether an advantage was gained by omitting the issue on appeal. Mapes v. Tate, 388 F.3d 187, 191 (6th Cir. 2004).
Because “[t]he reasonableness of counsel’s performance is to be evaluated from counsel’s perspective at the time of the alleged error and in light of all the circumstances,” we must consider the law at the time of the appeal in question to evaluate the performance of Russell’s counsel. Jacobs v. Mohr, 265 F.3d 407, 418 (6th Cir. 2001) (quoting Strickland, 466 U.S. at 689). This prong of the Strickland test, as applied here, focuses solely on whether failing to raise Johnson caused Russell’s appellate counsel’s performance to fall “below an objective standard of reasonableness.” Strickland, 466 U.S. at 688.
In granting Russell’s expanded COA, we recognized that the Ohio Court of Appeals determined that the reasoning it earlier used to deny Russell’s merger claim was abrogated by Johnson. See State v. McGail, 55 N.E.3d 513, 536 (Ohio Ct. App. 2d Dist. 2015) (“The State’s reliance on Russell [I] is unpersuasive . . . because the opinion in that case predated Johnson . . . in which the Ohio Supreme Court explicitly rejected the abstract-comparison approach used in Russell [I].”). Because Russell’s same counsel was filing another appeal, it would have been objectively reasonable to re-raise the merger of sentencing issue given the fact that the law relied on by the court in the first appeal had clearly changed.
First, the change in law was both “significant and obvious.” Mapes, 388 F.3d at 191. It was significant because—as outlined below—had the issue been raised on direct appeal, Russell’s sentences for murder and robbery likely would have been merged under the Johnson test. It was obvious because Johnson directly overruled the primary case relied on by the court in Russell’s prior appeal, Rance. Johnson, 942 N.E.2d at 1063 (“Unfortunately, the standard announced in Rance has proven difficult to apply. We take this opportunity to overrule Rance.”).
8
Further, there was no contrary law at the time of the appeal in question. Johnson was decided on December 29, 2010. The appeal that failed to re-raise the merger argument was filed in November 2011 and was decided in February 2012. See Russell II, 2012 WL 368135. No case cited by the Warden that challenges a winning argument for Russell under Johnson took place in the time period between December 2010 and November 2011.
Finally, there does not appear to have been a strategy in neglecting to re-raise the merger issue in the appeal filed in November 2011. Russell’s counsel challenged the trial court’s failure— for the second time—to find a Batson violation. Russell gained no apparent advantage by not also raising the merger of sentences under the new case law that had since developed. On the contrary, it was in fact the stronger of the two arguments.
We conclude that Russell’s appellate counsel was deficient in his representation by failing to re-raise the merger argument in the November 2011 appeal.
2. Prejudice The next question is whether the deficient performance by Russell’s appellate counsel constituted prejudice under the Strickland analysis. Establishing prejudice requires the defendant to “show that there is a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different. A reasonable probability is a probability sufficient to undermine confidence in the outcome. ” Strickland, 466 U.S. at 694. “Although this analysis necessarily involves an evaluation of the underlying claims, it does not require a decision on or a determination of these issues.” Mapes, 388 F.3d at 194. The question is whether, but for the counsel’s deficiency, there is a reasonable probability Russell would have succeeded on his appeal. Id.
9
In evaluating whether deficient performance by counsel caused prejudice, we may not consider the effect of counsel’s failure to take an action that might have been meritorious at the time of performance if it is now “wholly meritless under current governing law.” Evans v. Hudson, 575 F.3d 560, 565–66 (6th Cir. 2009) (emphasis added) (quoting Lockhart v. Fretwell, 506 U.S. 364, 374 (1993) (O’Connor, J, concurring)). In Fretwell, the Supreme Court considered whether a habeas petitioner’s trial counsel was ineffective for failing to make a Collins2 objection during the petitioner’s sentencing proceeding. 506 U.S. at 366–67. The Court found that the petitioner was not prejudiced by his counsel’s failure to raise a Collins objection because Collins, although binding law at the time of the petitioner’s sentencing, was later overruled. Fretwell, 506 U.S. at 370–71. The Court stated that “[s]heer outcome determination” is not sufficient to show prejudice under Strickland, id. at 370, and found that “[t]he result of the sentencing proceeding in the present case was neither unfair nor unreliable” because Collins was later overruled, id. at 371. It reasoned that to hold otherwise—i.e., to find prejudice even where counsel’s “error” was based on a decision that was no longer good law—“would grant criminal defendants a windfall to which they are not entitled.” Id. at 366. Thus, the Court denied the petitioner’s claim for habeas relief for lack of prejudice. Id. at 373.
In the present case, the Warden argues that Fretwell forecloses Russell’s ineffective assistance of counsel claim based on Johnson because, according to the Warden, Johnson is no longer good law in Ohio. However, contrary to the Warden’s suggestions, no Ohio court has ever expressly overruled Johnson, and, in fact, Ohio courts have continued to apply Johnson even after Ruff/Earley. See, e.g., McGail, 55 N.E.3d at 532–34. More importantly, Russell is not seeking a “windfall” based on Johnson. The core holding of Johnson—that courts should not analyze the
2 Collins v. Lockhart, 754 F.2d 258 (8th Cir. 1985), overruled by Perry v. Lockhart, 871 F.2d 1384 (8th Cir. 1989).
10
statutory elements of potential allied offenses in the abstract—remains binding law in Ohio. See Johnson, 942 N.E.2d at 1069 (overruling Rance “to the extent that it calls for a comparison of statutory elements solely in the abstract”); see also McGail, 55 N.E.3d at 533 (“[D]espite its proclaimed obsolescence, [Johnson] retains similarity to the Ruff standard, particularly with respect to the conduct and animus issues . . . .”). Accordingly, Russell would not receive a “windfall” if the state court reviews his ineffective assistance claim under either Johnson or current Ohio case law, because both require courts to analyze a defendant’s conduct in the specific factual circumstances of his case rather than in the abstract. The analysis that follows addresses Russell’s argument under Ohio’s current body of law on whether sentences should merge.
Ohio Revised Code § 2941.25 lays out when a defendant may be convicted of multiple offenses:
(A) Where the same conduct by defendant can be construed to constitute two or more allied offenses of similar import, the indictment or information may contain counts for all such offenses, but the defendant may be convicted of only one.
(B) Where the defendant’s conduct constitutes two or more offenses of dissimilar import, or where his conduct results in two or more offenses of the same or similar kind committed separately or with a separate animus as to each, the indictment or information may contain counts for all such offenses, and the defendant may be convicted of all of them.
Ohio Rev. Code Ann. § 2941.25. For eleven years, the Ohio Supreme Court relied on the test announced in Rance to determine whether crimes are “allied offenses of similar import” and should be merged under Ohio Revised Code § 2941.25. See Rance, 710 N.E.2d at 702–03. This test evaluated the “statutory elements in the abstract” and dictated that “[i]f the elements of the crimes correspond to such a degree that the commission of one crime will result in the commission of the other, the crimes are allied offenses of similar import.” Id. (internal quotations omitted).
11
In December 2010, the Ohio Supreme Court overruled Rance, eliminating the abstract analysis of the criminal statutes. Johnson, 942 N.E.2d at 1069. In doing so, the court held that Ohio Revised Code § 2941.25 “instructs courts to examine a defendant’s conduct—an inherently subjective determination” that “may result in varying results for the same set of offenses in different cases.” Id. at 1070. The approach laid out in Johnson directed courts first to decide “whether the offenses were committed by the same conduct,” which can be determined by asking “whether it is possible to commit one offense and commit the other with the same conduct, not whether it is possible to commit one without committing the other.” Id. If “the offenses correspond to such a degree that the conduct of the defendant constituting commission of one offense constitutes commission of the other, then the offenses are of similar import.” Id. But, if it is possible for the separate offenses to be committed by the same conduct, then the court must address whether they were committed in “a single act, committed with a single state of mind.” Id. (quoting State v. Brown, 895 N.E.2d 149, 158 (Ohio 2008) (Lanzinger, J., concurring in judgment)). “If the answer to both questions is yes, then the offenses are allied offenses of similar import and will be merged.” Id.
Since Johnson was decided in 2010, Earley and Ruff have further modified the standard Ohio courts apply when evaluating whether a trial court may impose cumulative sentences for two separate crimes. See Ruff, 34 N.E.3d 892; State v. Earley, 49 N.E.3d 266 (Ohio 2015). In Earley, the Ohio court described Ruff as laying out the “three-part test under R.C. 2941.25 to determine whether a defendant can be convicted of multiple offenses.” Earley, 49 N.E.3d at 269. That test requires courts to consider the “conduct, the animus, and the import” of the crimes in question. Ruff, 34 N.E.3d at 899.
12
As a practical matter, when determining whether offenses are allied offenses of similar import within the meaning of R.C. 2941.25, courts must ask three questions when defendant’s conduct supports multiple offenses: (1) Were the offenses dissimilar in import or significance? (2) Were they committed separately? and (3) Were they committed with separate animus or motivation? An affirmative answer to any of the above will permit separate convictions. The conduct, the animus, and the import must all be considered.
Id. Although the current standard requires an additional question, it retains Johnson’s framing of the conduct and animus issues. McGail, 55 N.E.3d at 533. We analyze the three questions presented under the Ruff/Earley standard separately.
a. Separate Conduct: Whether the Two Offenses Were Committed Separately One part of the analysis is focused on the conduct itself. McGail, 55 N.E.3d at 533.
Whereas Earley laid out the Ohio Supreme Court’s current standard for evaluating whether sentences should merge under Ohio Revised Code § 2941.25, McGail provides insight into how the courts have applied the standard to the merger of aggravated robbery and felony murder. In McGail, the Ohio Court of Appeals, Second District (the same intermediate court that reviewed Russell’s conviction) analyzed the evolution of the merger-of-offenses doctrine under Ohio law and applied it to a set of facts similar to those in the present case. Id. The court used the first step of Johnson (and the second step of Ruff)—whether it is possible to commit aggravated robbery and murder with the same conduct—to analyze whether the crimes were committed separately. Id. at 533–34. On this prong of the test, the court found in favor of the defendant. Id. The court held that it is possible to commit both aggravated robbery and murder with the same act because “a victim could die from the use of a deadly weapon in the course of an aggravated robbery, resulting in the victim’s murder.” Id. at 534.
Here, Russell’s facts are analogous. Russell committed murder through the use of a deadly weapon in the course of robbing his victim. As a result, there is a reasonable probability that Russell did not commit the crimes separately for purposes of § 2941.25.
13
b. Separate Animus or Motivation The next question is whether Russell committed the aggravated robbery and murder with separate animus or motivation. In analyzing this question, Ohio courts have relied on the presence of a struggle between the perpetrator and the victim as an indicator for whether “the force used to effectuate an aggravated robbery is far in excess of that required to complete the robbery, or where the circumstances suggest that a separate intent to kill existed.” McGail, 55 N.E.3d at 534 (quoting State v. Jackson, No. 24430, 2012 WL 1900373 (Ohio Ct. App. 2d Dist. 2012)); see also State v. Jefferson, 2017 WL 1193797, at *2 (Ohio Ct. App. 2d Dist. 2017).
The court in McGail, for instance, found that because the defendant testified to the gunshot being a “reflex” and because the gun discharged accidentally during the struggle, there was an “absence of any intent to kill separate and apart from the intent to rob.” McGail, 55 N.E.3d at 535. Consequently, the court determined that the defendant committed the two offenses with the same conduct and the same animus and, despite applying the Johnson standard in its analysis, came out the same way as it would have under Ruff:
We reach the same conclusion under the Ruff standard, which the Ohio Supreme Court applied in Earley. We see nothing in Ruff that alters or undermines the foregoing analysis about McGail’s commission of murder and aggravated robbery involving the same conduct committed with the same animus. For the reasons set forth above, we conclude that the two offenses were not committed separately and were not committed with a separate animus or motivation.
Id. at 536.
In the present case, the evidence provides us with minimal insight into how the robbery and murder took place, but we can infer from the record that Russell and Troutwine engaged in a struggle prior to the gunshot. In her testimony, the state’s primary witness—Candace Hargrove— testified that she left the kitchen shortly after Russell approached the victim with a gun demanding his money. From the adjacent room, she “heard tussling and then [she] heard a gunshot.” She
14
testified that Russell immediately “came running out of the kitchen saying, ‘Oh my God, I shot him. I didn’t mean to shoot him. I didn’t mean to shoot him.’” The testimony of another witness that lived in the apartment below where the robbery took place also indicated that on the morning in question, he “heard some tussling” before he heard a gunshot, “like somebody was upstairs wrestling or something.” The autopsy reveals that Troutwine was shot only once. The record thus indicates a struggle between Russell and the victim, and the state appellate court could reasonably infer that the use of the gun was not separately intended or motivated by an intent to kill.
In contrast, Ohio courts have found the use of force to establish a separate animus where the force was used after the robbery was completed, State v. Kerby, 2014 WL 3809050, at *5 (Ohio Ct. App. 2d Dist. 2014); where the perpetrator pursued, stopped, and returned with an employee who tried to escape while he was robbing the store, State v. Moore, 2018 WL 2473476, at *3 (Ohio Ct. App. 2d Dist. 2018); and where a victim was shot multiple times, thus indicating excessive force or a separate intent to kill, Jackson, 2012 WL 1900373, at *24. None of those circumstances exist in the record before us.
Because the facts established that Russell engaged in a struggle with the victim and because his reaction indicated that the shooting was accidental, there does not appear to have been separate animus or motivation involved in the robbery and the murder. Hence, this part of the Ruff test does not preclude Russell’s sentences from being merged.
c. Dissimilar in Import or Significance “[O]ffenses are not allied offenses of similar import if they are not alike in their significance and their resulting harm.” Ruff, 34 N.E.3d at 897. Although “import” is not clearly defined in Ohio law, Ruff lays out several illustrations of offenses of dissimilar import: when a “defendant’s conduct put more than one individual at risk,” id. at 898 (citing State v. Jones, 480
15
N.E.2d 408, 410 (Ohio 1985)), when neither offense is incident to the other, id. (citing State v. Moss, 433 N.E.2d 181 (Ohio 1982)), if “the defendant’s conduct constitutes offenses involving separate victims, or if the harm that results from each offense is separate and identifiable.” Id. Additional considerations in evaluating whether two offenses are of dissimilar import or significance include the language of the code and the intention of the Ohio General Assembly in criminalizing the offenses. See Earley, 49 N.E.3d at 270. The Ohio Supreme Court has emphasized that this part of the allied-offense analysis must remain fact specific and that the “evidence at trial or during a plea or sentencing hearing will reveal whether the offenses have similar import.” Ruff, 34 N.E.3d at 898.
Because only one individual, Troutwine, was at risk in this case and because the murder was incident to the aggravated robbery, we begin by looking at whether the harms that resulted from Russell’s conduct were separate and identifiable. In analyzing “whether McGail’s murder and aggravated robbery offenses [were] of dissimilar import,” the Ohio Court of Appeals evaluated the harm or harms caused to the victim. 55 N.E.3d at 536. The court found that because the defendant did not succeed in taking the intended property—marijuana—from the victim, the harm could not be the loss of property. Id. But it was not just because the theft was unsuccessful. Rather, “the real harm resulting from the aggravated robbery . . . [was] precisely the type of harm that [the Ohio aggravated robbery code] is intended to prevent—[the victim’s] loss of his life as a result of being shot when [the perpetrator] used a deadly weapon while attempting to commit a theft offense.” Id. at 536–37. The court therefore found that the murder and aggravated robbery offenses were not of dissimilar import, particularly because “the murder was part and parcel of a struggle during an on-going attempt to [commit a robbery].” Id. at 537.
16
The Warden argues that Troutwine’s loss of property—his car—is a harm separate and distinct from his loss of life. It is true that the Ohio Court of Appeals has held that “[t]he harm suffered by a victim from having his car stolen is clearly separate and identifiable from the harm … suffered from being shot in the head and killed.” State v. Wood, 2016 WL 197092, at *12 (Ohio Ct. App. 2d Dist. 2016). The question addressed in Wood, however, was specific to the merger of sentences for aggravated murder and grand theft of a motor vehicle. Id. In this case, Russell was convicted of and sentenced for grand theft of the victim’s motor vehicle, but he is not asking for merger of that sentence. Instead, Russell seeks to have his robbery and murder sentences merged. To convict Russell of aggravated robbery, the jury had to find only that Russell “had a deadly weapon on or about his person or under his control, and displayed, brandished, indicated he possessed or used the weapon” while he either committed or attempted to commit a theft offense.3 The next consideration is what the Ohio legislature intended to punish with the underlying statutes. In Russell I, the Ohio Court of Appeals addressed the societal interest and purposes underlying the felony murder and aggravated robbery statutes and analyzed them differently than did the court in McGail:
[T]he societal interests underlying the felony murder and aggravated robbery statutes are not identical . . . . [T]he societal interest underlying the felony murder statute is protection of persons from physical harm, specifically death. The societal interest underlying the aggravated robbery statute is two-fold. The purpose of the aggravated-robbery statute is to punish potential or actual harm to persons and to protect property. . . . The existence of an additional societal interest, the protection 3 At trial, the state argued that Russell stole items as part of the robbery charge, but there was no actual evidence that he stole anything from Troutwine. The jury did not have to indicate whether they found beyond a reasonable doubt that Russell committed theft, only that he attempted to. Thus, it is unclear whether—outside of the grand theft conviction—Russell deprived the victim of his property. Nonetheless, Ohio courts have found that aggravated robbery and murder counts should merge even in cases where there was stolen property. State v. Robinson, 2019 WL 495579, at *11 (Ohio Ct. App. 1st Dist. 2019); State v. Hale, 2019 WL 3855862, at *9 (Ohio Ct. App. 8th Dist. 2019). The Warden also points to the $15,498.25 in restitution given to the victim’s widow as an indication that there was property stolen other than the car. It is clear from the transcript of the sentencing hearing, however, that the trial court ordered restitution in conjunction with the murder conviction and not the grand theft charge.
17
of property, underlying the aggravated robbery statute supports finding a legislative intent to consider aggravated robbery and felony murder as having different imports.
Russell I, 2010 WL 3835645, at *12 (citations omitted). But see McGail, 55 N.E.3d at 536–37 (finding that “[the victim’s] loss of his life as a result of being shot when [defendants] used a deadly weapon while attempting to commit a theft offense” is “precisely the type of harm that [the Ohio aggravated robbery code] is intended to prevent.”). However, in reaching its conclusion, the Ohio Court of Appeals in Russell I relied on the test from Rance, which Johnson overruled and McGail explicitly rejected:
In opposition to the foregoing conclusion, the State cites [State v. Russell] for the proposition that felony murder and aggravated robbery are not allied offenses because “[c]omparing the elements of the two offenses in the abstract, commission of neither offense necessarily results in commission of the other.” . . . The State’s reliance on Russell is unpersuasive, however, because the opinion in that case predated Johnson, . . . in which the Ohio Supreme Court explicitly rejected the abstract-comparison approach used in Russell. We note too that the more recent Ruff opinion did not alter Johnson’s rejection of the abstract-comparison approach, which remains inapplicable in Ohio.
McGail, 55 N.E.3d at 535–36 (emphasis added) (citing Russell I, 2010 WL 3835645).
The allied-offense test also considers the language and structure of the statutes. Ohio courts have found murder or assault to merge with aggravated burglary or robbery because the assault or murder constituted the aggravated element of the other crime. For example, the Ohio Court of Appeals in State v. Ramey held that “the trial court erred by failing to merge the offense of aggravated burglary with the offense of murder as a proximate cause of felonious assault.” 2018 WL 3702299, at *5 (Ohio Ct. App. 2d Dist. 2018) (footnote omitted). It reasoned that because the murder was committed specifically “as a proximate cause of felonious assault,” the murder was “the aggravating element necessary to make the burglary an aggravated burglary rather than a
18
simple burglary.” Id. (“Without the physical harm caused by the felonious assault, the burglary would not have had the aggravating element of inflicting physical harm.”).
In Ruff, as another example:
The trial court imposed separate sentences for [rapes and aggravated burglaries]
because it believed that the offense of aggravated burglary was complete upon the entry into the dwelling. The court of appeals disagreed and determined that because the physical harm that constituted the aggravating factor for the burglary offense was the rape of the victim, Ruff could not be separately convicted for both the aggravated burglary and rape of each victim.
Ruff, 34 N.E.3d at 899. The Ohio Supreme Court, however, “decline[d] to create an absolute rule based upon the definition of the offenses” because it refused to “hold that every aggravated burglary and rape automatically lead to the same import.” Id. Despite the argument that rape and aggravated burglary involved two separate harms—one against a person, one against property— the court reiterated that Ohio Revised Code § 2941.25 required courts “to focus on the defendant’s conduct to determine whether one or more convictions may result, because an offense may be committed in a variety of ways and the offenses committed may have different import.” Id.
Here, to determine whether Russell was prejudiced by deficient counsel, we are tasked with determining whether there is a “reasonable probability” that the Ohio Court of Appeals would merge Russell’s aggravated robbery and murder sentences under current law. Smith, 567 F.3d at 257 (quoting Strickland, 466 U.S. at 694). In other words, we must find “a probability sufficient to undermine confidence in the outcome,” id., a probability by “less than a preponderance of the evidence.” Lyons v. Jackson, 299 F.3d 588, 599 (6th Cir. 2002) (citing Strickland, 466 U.S. at 694). In defining the current test used to evaluate sentence mergers in Ohio law, Ruff “recognize[d] that this analysis may be sometimes difficult to perform and may result in varying results for the same set of offenses in different cases. But different results are permissible, given that the statute instructs courts to examine a defendant’s conduct—an inherently subjective determination.” Ruff,
19
34 N.E.3d at 899 (quoting Johnson, 942 N.E.2d at 1070). Given the varied outcomes resulting from similar fact patterns, we conclude that there is a reasonable probability that the Ohio Court of Appeals could find, as it did in McGail, that Russell’s sentences should merge, a possibility “sufficient to undermine confidence in the outcome” of Russell’s prior appeal. Smith, 567 F.3d at 257.
In sum, although the facts of Russell’s case may vary from those in McGail, there is not a significant enough variance to predict that the court would come out differently on the two same crimes in this case. Russell’s conduct did not affect more than one person, the two crimes were incident to each other, there is no definitive finding that Russell deprived Troutwine of property (other than his vehicle)—meaning that the crimes likely resulted in only one harm. Additionally, although the Ohio Court of Appeals concluded that the legislature intended to punish Russell’s crimes separately, with that same intent the court in McGail still determined that the crimes were of similar import. We therefore conclude that the district court correctly issued the conditional writ in this case.
Finally, we address two arguments offered by the Warden as something of a last-ditch effort to defeat the remedy ordered by the district court. Both are without merit and need not detain us long.
The Warden insists that merger of the sentences would result in Russell “getting away with murder,” phrasing the issue as whether “the Ohio General Assembly intended to punish a thief the same as a thief who murders someone.” But merger of the sentences by the state courts does not mean that Russell will be punished only as “a thief” and not for the murder. “The General Assembly has made clear that it is the state that chooses which of the allied offenses to pursue at sentencing, and it may choose any of the allied offenses.” State v. Whitfield, 922 N.E.2d 182, 187
20
(Ohio 2010). If Russell is granted a new appeal and prevails, the state may pursue the greater of the offenses at sentencing and may urge the sentencing court to take into consideration all sentencing enhancements or mitigating factors that it had considered previously.
As we have emphasized, it us not our responsibility to make a determination on the merits of the sentencing merger. We need to find only that but for Russell’s appellate counsel’s failings, there is a reasonable probability that Russell would have prevailed in securing a merger of his sentences on appeal. See Mapes, 388 F.3d at 194. “Thus, as the only constitutional violation properly before us is the ineffective assistance of appellate counsel, the appropriate remedy is the one granted by the district court, that is, a writ conditioned upon Ohio courts granting a new, direct appeal.” Id. The ultimate determination on the merits needs to be made by the Ohio courts in the development of their law, in order to “avoid[] unnecessarily interfering with Ohio’s interest in correcting its own errors.” Id.
Finally, the Warden also insists that the magistrate judge ultimately recommended granting the writ due to his belief that the district court was compelled by the law-of-the-case doctrine to adhere to our statement in the expanded COA order that “there is a strong possibility that the Ohio Court of Appeals would have merged Russell’s aggravated robbery and felony murder convictions” had the issue been re-raised by appellate counsel. The magistrate judge, in his supplemental report, did state that “[a] ruling on a certificate of appealability constitutes the law of the case, binding in subsequent stages of the litigation.” However, it is clear that the statement in our prior order does not constitute the “law of the case” and should not have been relied upon to reach such a conclusion. Indeed, the record establishes that the district court independently analyzed the issue under the current state of Ohio’s law and did not rely merely on a statement from our interim order. Russell IV, 2019 WL 120974, at *1.
21
III. CONCLUSION
For the reasons set out above, we AFFIRM the decision of the district court to “issue a writ of habeas corpus conditioned on Russell’s being allowed a new appeal to the Ohio Second District Court of Appeals . . . in which the assignment of error at issue in the remand—whether [Russell]’s aggravated robbery and murder convictions should be merged under Ohio Revised Code § 2941.25—is presented for decision.”