Cortinas v. State

195 P.3d 315, 124 Nev. 1013, 124 Nev. Adv. Rep. 86, 2008 Nev. LEXIS 105

• Court: Nevada Supreme Court
• Decided: October 30, 2008
• Docket: 47905
• Status: Published

Opinion

OPINION

By the Court,

Parraguirre, J.:

The primary issue we address in this appeal is whether harmless-error review applies when a general verdict based on multiple theories of liability may rest on a legally invalid alternative theory. To resolve this issue, we must address relevant federal cases and reconcile two prior decisions by this court.

The United States Supreme Court first addressed the impact of a general verdict that may rest on a legally valid or a legally invalid alternative theory of liability in Stromberg v. California, in which the Court held that a general verdict delivered under these circumstances must be set aside unless it is possible to determine that the jury based the verdict on a legally valid ground. 1 In Keating v. Hood, the Ninth Circuit Court of Appeals reasoned that reversal is required in such cases unless the court is “absolutely certain” that the jury relied on a valid ground to reach its verdict. 2

We adopted Keating’s absolute certainty approach to Stromberg error in Bolden v. State. 3 After finding Stromberg error as the result of erroneous jury instructions on vicarious coconspirator liability for specific intent crimes, we reversed the defendant’s convictions for several specific intent offenses that were committed by his coconspirators, explaining that we could not conclude with absolute certainty that the jury did not rely on the erroneous instructions when returning those verdicts. But in a more recent case, Nay v. State 4 we reviewed an instructional error that could be characterized as Stromberg error for harmless error under the Chapman v. California 5 standard for harmless-error review. Accordingly, in Nay, after rejecting the use of an “afterthought” robbery as the predicate felony for felony murder, we reversed the defendant’s first-degree murder conviction because we could not determine beyond a reasonable doubt that the jury would have returned the same verdict had it been properly instructed.

In this appeal, we take the opportunity to reconcile Bolden’s absolute certainty approach to Stromberg error with Nay’s reliance on traditional harmless-error review. Contrary to Bolden’s implica tions, we view Stromberg error as a subcategory of trial error that is susceptible to harmless-error review under the Chapman standard as it has been applied in our instructional error cases. Thus, we conclude that harmless-error review applies when a general verdict may rest on a legally valid or a legally invalid alternative theory of liability. Accordingly, we retreat from Bolden’s absolute certainty approach and reaffirm Nay.

Conducting harmless-error review in this case, we conclude beyond a reasonable doubt that the jury would have returned the same first-degree murder verdict had it not been misled that an afterthought robbery could satisfy the felony-murder rule. Although the general verdict form obscures the theory of liability that the jury selected, based on the overwhelming evidence of premeditated and deliberate murder presented at trial, as well as the jury’s actual findings, presenting the jury with an invalid theory of felony murder was harmless error.

Separately, regarding the State’s theory of robbery in this case, we reaffirm that the general intent and the taking required for robbery may occur after a victim is deceased so long as the use of force or coercion by the defendant — for whatever purpose-occurred while the victim was alive and the defendant took advantage of the terrifying situation he created to flee with the victim’s property. Thus, we conclude that the district court did not improperly instruct the jury with regard to robbery.

FACTS

On April 20, 2003, Kathryn Kercher’s nude body was discovered in the desert south of Boulder City in an advanced stage of decomposition. Two clumps of blond hair were lying adjacent to the body, one of which appeared to have been cut from Kercher’s head. Three stab wounds appeared on Kercher’s back.

An autopsy revealed hemorrhages in various areas of Kercher’s neck and at the base of her tongue. From this, the pathologist determined that Kercher died from asphyxia due to strangulation. According to the pathologist, prolonged strangulation with a ligature could have produced a distribution of hemorrhaging consistent with Kercher’s wounds, assuming that Kercher struggled with her attacker, thus causing the ligature to move as it was held to her neck.

Shortly after Kercher’s body was discovered, officers from the Las Vegas Metropolitan Police Department responded to a call that appellant Armando Cortinas was attempting to commit suicide. Cortinas approached the responding officers briskly. He then asked to be placed in handcuffs. While restrained, Cortinas stated that he wanted to kill himself, prompting police to call an ambulance.

When police officers asked him why he wanted to commit suicide, Cortinas stated that he had done something bad that he could not live with — he had killed a prostitute. Cortinas then stated that he dumped the victim’s body in the desert near Boulder City and described the victim’s tattoos. After the officers confirmed the victim’s description with Boulder City Police, Cortinas was arrested.

Following his arrest, Cortinas consented to a search of his bedroom and volunteered that police would find the victim’s earrings in a coin bank on his dresser. During the search, police recovered the earrings and, among other things, a 10- to 12-inch steel cable PVC pipe cutter with yellow handles attached at either end tucked between Cortinas’ mattress and box spring. Cortinas later described this tool as a “garrote” that could be used for strangling.

During an interview, Cortinas’ brother told police that Cortinas had a girlfriend over to the house a week earlier. At some point, the brother heard the girl scream, thought that the two were horse-playing, and told Cortinas to keep it down. In response, Cortinas turned up his music volume. Later, when he emerged from his bedroom, Cortinas told his brother that the girl had passed out and that he would use her car to take her home, then travel back on the bus.

At the police station after his arrest, Cortinas confessed to killing Kercher. Cortinas told police officers that he used his father’s cellular phone to respond to a massage advertisement in CityLife magazine and arranged to meet with Kercher at his parents’ home. When she arrived, Cortinas paid Kercher $150 for oral sex. Afterward, Cortinas approached Kercher from behind and, before she could scream, looped a nylon lanyard keychain “in a figure eight sort of manner” around her neck. 6 In this fashion, Cortinas said that he strangled Kercher for nearly an hour, stopping at intervals to determine if she was still breathing and resuming if necessary “to finish it off.” Finally, unable to kill Kercher by strangulation, Cortinas wrapped his arm around her, fell backwards onto his bed, and broke her neck.

According to Cortinas’ confession, even this final attempt to take Kercher’s life failed, as Kercher was still gasping for air. Despite her attempts to breathe, Cortinas taped Kercher’s skirt around her head to absorb the blood that had begun to issue from her mouth. He then taped her wrists together in front of her body. With Kercher bound in this manner, Cortinas placed Kercher in the trunk of her car and drove to the Boulder City desert. Unsure that she was dead when he arrived in the desert, Cortinas stabbed Kercher three times in the back with a butterfly knife, so that she would “drown in her own blood” as it pooled in her lungs. Using the same knife, Cortinas then removed the skirt that he had taped around Kercher’s head, cutting away clumps of her hair in the process.

Returning from Boulder City, Cortinas disposed of the lanyard, knife, and other evidence in different parts of Las Vegas and Henderson and parked Kercher’s car around the corner from his parents’ house. Before discarding Kercher’s purse, Cortinas recovered his $150 as well as a bag of marijuana, which he later sold. Although he discarded Kercher’s other jewelry, he kept her diamond earrings, eventually placing them in his coin bank. 7 The next day, Cortinas moved Kercher’s car to the Stratosphere Hotel and then offered it to a friend if the friend would agree to burn the car’s contents to destroy his fingerprints.

The subsequent investigation further confirmed Cortinas’ connection to the killing. In particular, the police confirmed that Kercher’s DNA matched the DNA found on the earrings recovered from Cortinas’ coin bank, a CityLife advertisement had recently run with Kercher’s telephone number, and a call had been placed to that number from Cortinas’ father’s phone on the night that Kercher was killed.

The State charged Cortinas with first-degree murder with the use of a deadly weapon and robbery with the use of a deadly weapon. With respect to first-degree murder, the State pursued a conviction on alternative theories of willful, deliberate, and premeditated murder and felony murder. At trial, the jury instructions and the prosecutor’s closing argument indicated that the jury could convict Cortinas of first-degree murder based on a felony-murder theory even if the jury found that the robbery occurred as an afterthought to the murder. The jury returned a general verdict finding Cortinas guilty on all counts. The jury’s verdict did not reveal the theory under which the jury found Cortinas guilty of first-degree murder. Cortinas appealed.

While his appeal was pending, we issued our decision in Nay v. State , 8 in which we rejected the use of an “afterthought” robbery as the predicate felony to support a felony murder. After rejecting this theory of felony murder, we applied Chapman harmless-error review to determine if the erroneous instructions and prosecutorial argument based on afterthought robbery merited reversal. During oral argument in the present case, concerns emerged regarding whether Nay’s harmless-error analysis was correct in light of our decision in Bolden v. State. 9 Following oral argument, Cortinas filed supplemental authorities addressing this issue, which we address below.

DISCUSSION

On appeal, Cortinas argues that Nay requires reversal of his first-degree murder conviction and that his robbery conviction should be overturned because the jury was permitted to find him guilty of robbery without proof that he formed the requisite intent for that offense before the victim was killed. We disagree. Although Cortinas was entitled to an instruction precluding the use of afterthought robbery as the predicate felony for felony murder, we conclude that this instructional error was harmless beyond a reasonable doubt. We further conclude that the district court correctly instructed the jury on the elements of robbery.

Cortinas was entitled to his requested felony-murder instruction but the failure to give the instruction was harmless

We begin our analysis by considering the district court’s refusal to give Cortinas’ requested jury instruction advising that afterthought robbery may not serve as a predicate felony for felony murder. Cortinas challenges this decision, asserting that his requested instruction correctly stated the legal limitations on the use of robbery to seek a first-degree murder conviction under the felony-murder rule.

District courts have broad discretion to settle jury instructions. 10 While we normally review the decision to refuse a jury instruction for an abuse of that discretion or judicial error, we review de novo whether a particular instruction, such as the one at issue in this case, comprises a correct statement of the law. 11

While this appeal was pending, we decided Nay and joined the majority of jurisdictions that prohibit the use of an afterthought robbery as a predicate felony for felony murder. 12 Our reasons for doing so were clear: (1) the purpose of the felony-murder statute— general deterrence of dangerous felonies — is not served by punishing a killing as first-degree murder if an accused lacks a felonious intent at the time of the killing; and (2) absent an accused’s contemporaneous intent, the malice necessary to prosecute a killing as first-degree murder cannot legally be implied from the underlying felony. 13 Based on this clarification of the felony-murder rule as it relates to robbery, we concluded that failing to give the proposed instruction in Nay amounted to judicial error. 14

Similar to the instruction proposed in Nay, Cortinas’ proposed instruction on felony murder required the jury to find that he intended to commit the robbery before he mortally wounded the victim. This, we conclude, accurately corresponds to the clarification in Nay regarding the necessary timing of the intent to commit a robbery to satisfy the felony-murder rule. Applying Nay’s understanding of the felony-murder doctrine to this case, we conclude that Cortinas was entitled to his correctly framed felony-murder instruction. Having concluded that instructional error occurred in this case, we next consider that error in light of the rule set forth in Stromberg v. California 15 and determine whether such errors are subject to harmless-error review.

The rule of Stromberg v. California

In Stromberg, the United States Supreme Court held that a conviction must be reversed when the jury is presented with multiple theories of liability, one of which is unconstitutional, and it is impossible to discern from the verdict which ground the jury used to determine the defendant’s guilt. 16 Since Stromberg was decided, the Court has extended it beyond unconstitutional theories of liability to situations in which the jury returns a general verdict based on multiple theories of liability, one of which is legally invalid. 17 We recognize that the misdescription or omission in the felony-murder instruction in this case arguably gives rise to a Stromberg scenario in that the jury was presented with multiple theories and the erro neous instructions allowed for a conviction based on an invalid theory of felony murder.

Availability of harmless-error review

Our cases diverge in the way they determine whether reversal is required when an instructional error allows a jury to return a verdict based on a legally invalid theory but the jury is also presented with one or more valid alternative theories. In Bolden, we viewed an instructional error on an alternative theory as Stromberg error and adopted an absolute certainty approach that is unique to the Ninth Circuit Court of Appeals. But two years later, in Nay, we neglected to take the same approach. Although the scenarios in Nay and Bolden — & general verdict obscuring whether the jury determined guilt on a legally valid or a legally invalid alternative theory of liability — were indistinguishable as a practical matter, in Nay the idea of a Stromberg error was not raised or addressed, and we viewed the error purely as an instructional error. As a result, we applied Chapman harmless-error review in that case. We now retreat from the absolute certainty approach advanced in Bolden and reaffirm the harmless-error approach taken in Nay.

The absolute certainty approach is unsound

In determining whether the instructional error in Bolden required reversal, this court viewed the error through the prism of Stromberg. In Bolden, the defendant was charged with multiple general and specific intent offenses for his involvement with four other men in a robbery-kidnapping. Regarding each offense, the jury was instructed on three alternative theories of liability: (1) aiding and abetting, (2) direct participation, and (3) vicarious coconspirator liability based upon the natural-and-probable-consequences doctrine. 18 On appeal, we rejected the latter version of vicarious coconspirator liability for specific intent offenses and thus were required to determine whether the jury’s verdict should stand when it was silent as to which theory the jury relied on to convict Bolden. 19

As this court then observed, the United States Supreme Court “ha[d] never addressed whether harmless error analysis is available in such cases.’ ’ 20 Searching for guidance in this regard, the Bolden court adopted the Ninth Circuit Court of Appeals’ approach to Stromberg error as developed in Keating v. Hood. 21 Under that ap proach, reversal is mandatory unless the court “ ‘is absolutely certain’ that the jury relied upon the legally correct theory to convict the defendant.’ ’ 22

Applying the Keating absolute certainty approach, the Bolden court noted that the ground upon which the jury actually relied to determine the defendant’s guilt was not discernable from the verdict, and thus the court could “only speculate as to the basis for the jury’s decision.” 23 Consequently, the court determined that it could not “conclude with absolute certainty that the jury did not find Bolden guilty of the burglary and kidnapping offenses based on the erroneous instruction.” 24 Without any further examination of the record or the jury’s findings, the court reversed Bolden’s convictions of the specific intent offenses.

Soon after we decided Bolden, the Ninth Circuit, in Lara v. Ryan, 25 for the first time clarified its rationale for approaching Stromberg error in a manner that all but guarantees reversal. In Lara, a federal habeas case, the defendant was convicted of two counts of attempted murder in state court pursuant to a general jury verdict that may have rested on a legally invalid alternative theory of attempted murder based on implied malice. 26 On direct review, Lara’s convictions were affirmed under the Chapman standard of harmless-error review. On collateral review, a federal district court denied Lara habeas relief but concluded that Chapman was not the correct harmless-error standard.

On appeal, the Ninth Circuit agreed with the district court and went as far as it ever has in rationalizing its disparate treatment of Stromberg error. In particular, the court reasoned that the error was structural:

[T]he trial court did not merely omit or misstate an element of the charged offense. Rather, its error was structural, because it enabled the jury to deliver a general verdict that potentially rested on different theories of guilt, at least one of which was constitutionally invalid. As such, Lara’s claim arises from the “very limited class of cases” in which a structural error has occurred. 27

Despite concluding that the error was structural in nature, a conclusion that normally requires automatic reversal, 28 the Ninth Circuit did not reverse Lara’s convictions. Instead, considering the trial court’s numerous instructions requiring the jury to find actual malice and the jury’s special finding that the attempted murders were willful, deliberate, and premeditated, the Ninth Circuit concluded that it could be “absolutely certain that the jury convicted Lara under [a legally valid] express-malice theory.’ ’ 29

In our view, the result in Lara is at odds with its reasoning. The availability of harmless-error review for constitutional errors is governed by a basic dichotomy: “a constitutional error is either structural or it is not.” 30 If the error is structural, then it defies harmless-error review, and reversal is automatic. 31 This is true without exception, “ ‘limited’ ” or otherwise. 32 Yet the reasoning in Lara presumes that Stromberg error, which is concededly a constitutional error, is structural and at the same time not. 33 Because an error can never be both, either the Ninth Circuit’s “structural” label is improvident or its absolute certainty approach merely reprises Chapman harmless-error review, despite masquerading as a more inflexible standard. Either way, we take this opportunity to reexamine the wisdom of our decision to follow the absolute certainty approach in Bolden.

Stromberg error is instructional trial error

Although Cortinas does not specifically maintain that the Stromberg error presented in this case is structural, he nevertheless urges this court to follow Bolden’s absolute certainty approach, which depends on that assumption. We reject Cortinas’ argument and conclude that Stromberg error is not structural but instead is instructional trial error. As such, Stromberg error is amenable to harmless-error review.

Whether a particular type of error is amenable to harmless-error review depends on whether the error can be categorized as struc toral error or trial error. 34 The United States Supreme Court has explained the distinction between these two types of errors. Unlike constitutional trial errors, “which occur[ ] during the presentation of the case to the jury, and which may therefore be quantitatively assessed in the context of other evidence presented,” 35 structural errors affect the very “‘framework within which the trial proceeds.’ ” 36 Consequently, these intrinsically harmful errors “ ‘necessarily render a trial fiindamentally unfair.’ ” 37 Because they are, therefore, never harmless, structural errors require automatic reversal. 38 Correspondingly, since “ ‘most constitutional [trial] errors can be harmless,’ ” 39 structural errors arise “only in a ‘very limited class of cases.’ ” 40

Conspicuously, the Stromberg error in this case is not a member of the limited class of structural error. 41 Indeed, as we observed in Bolden, the Supreme Court has never addressed whether Stromberg error is subject to harmless-error analysis, 42 despite numerous opportunities to do so. 43 Notwithstanding this silence regarding the nature of Stromberg error in particular, the Court in Neder v. United States clearly stated that the Chapman harmless-error stan dard applies to review of instructional errors involving the omission or misdescription of an element of an offense. As our instructional error cases have interpreted Neder’s holding, unless the error ‘“vitiates all the jury’s findings,’ and produces ‘consequences that are necessarily unquantifiable and indeterminate,’ ’ ’ 44 an instructional error amounts to constitutional trial error and is, therefore, subject to harmless-error review. 45

That is not to say that instructional error can never be structural in nature. Examining the relevant federal cases on this issue, the Supreme Court has found structural error in the context of jury instructions only once. In Sullivan v. Louisiana, the jury received a defective reasonable-doubt instruction that relieved the prosecution of its burden of proof and thereby allowed the jury to render a guilty verdict based on findings supported by less than a constitutional quantum of evidence. 46 Since the instruction therefore precluded the jury from delivering an actual verdict of guilty beyond a reasonable doubt, it defied harmless-error review, as an inquiry into whether the jury would have returned the same constitutionally deficient verdict had it been properly instructed was meaningless. 47

In contrast to Sullivan’s defective burden-of-proof instruction, an erroneous instruction that makes available an invalid alternative theory of liability, as occurred here, does not vitiate the jury’s findings. Nor does it defy harmless-error review, i.e., whether the jury would have reached the same verdict had the invalid theory not been available. In other words, even in the case of Stromberg error, there is still an “object, so to speak, upon which harmless-error scrutiny can operate.” 48

Further, we do not consider it reasonable to define Stromberg error as structural. In a Stromberg scenario, more than one theory of liability is presented to the jury, at least one of which is legally invalid. In such a case, Bolden, Keating, and Lara suggest that harmless-error review is precluded. On the other hand, when only one theory of liability is available and the attendant instructions omit or misdescribe an element, Nay and Neder suggest that harmless-error review applies even though, in contrast to a Stromberg scenario, the jury has no valid alternative basis upon which to rest a verdict. 49

As the Eighth Circuit Court of Appeals has recognized, since harmless-error review traditionally extends to single-theory cases in which there are no valid alternative theories upon which to rest a verdict, it would be “anomalous to read Stromberg to preclude harmless-error review . . . because the jury also was given the option to convict based on a constitutionally valid theory.’ ’ 50 The assertion that the availability of a valid alternative basis to convict would intensify the prejudice that normally results from submitting an invalid theory of guilt to the jury “ ‘cannot possibly be right, so it is plainly wrong.’ ” 51

Thus, to the extent that Stromberg error has been distinguished from instructional error involving the omission or misdescription of an element of an offense, we believe that the distinction is unpersuasive for purposes of harmless-error review. We therefore retreat from Bolden’s absolute certainty approach 52 and conclude that harmless-error review applies when a general verdict may rest on a legally valid or a legally invalid alternative theory of liability. In this regard, we accordingly reaffirm Nay.

Application of harmless-error review

Having concluded that Stromberg error is subject to harmless-error review, the appropriate standard is that articulated in Chapman— whether it appears “beyond a reasonable doubt that the error complained of did not contribute to the verdict obtained.’ ’ 53 We therefore must consider whether the instructional error in this case is harmless beyond a reasonable doubt. Unlike in Nay, we conclude beyond a reasonable doubt that the error in this case is harmless. In Nay, the defendant was charged with one count of first-degree murder and one count of robbery, each with the use of a deadly weapon, for beating his roommate to death with a baseball bat, then taking his deceased friend’s money, marijuana, and handgun. 54 Notably, Nay confessed to the killing but claimed to have acted in self-defense. 55

At trial, the State advanced two alternative theories of first-degree murder — willful, deliberate, and premeditated murder and felony murder based on robbery. With respect to the robbery count, the jury was properly instructed that Nay could have formed the required intent after the killing and still be guilty of that offense. 56 The district court, however, rejected Nay’s proposed instruction that would have prevented the jury from basing a felony-murder finding on a robbery that was an afterthought to the killing. 57

Exploiting this void during closing argument, the prosecutor distilled the reasoning required of the jury to reach a first-degree felony-murder verdict to a simple if-then proposition: “[i]f [the defendant] committed that robbery then he is guilty of felony murder.” 58 Thereafter, the jury returned a verdict finding Nay guilty of first-degree murder. However, the verdict did not reveal whether the jury based its first-degree murder finding on a willful, deliberate, and premeditated murder or a felony murder based on an afterthought robbery. 59 Applying the Chapman harmless-error stan dard, we concluded that it was “not possible to determine beyond a reasonable doubt that the jury would have convicted Nay of first-degree murder if it had been properly instructed.” 60

A similar pattern of events occurred in this case. As in Nay, the State charged Cortinas with first-degree murder and robbery, with respective deadly weapon enhancements. For purposes of robbery, the jury was properly instructed that the taking of the victim’s property could occur as an afterthought to the use of force; the district court, however, rejected Cortinas’ proposed instruction on felony murder that would have restricted the jury from using an afterthought robbery to reach a first-degree felony-murder verdict. 61 Taking advantage of the permissive space that resulted, during closing arguments, the prosecutor simplified the jury’s felony-murder reasoning to the same if-then analysis urged by the prosecutor in Nay:

The defendant takes his money back. He takes the earrings. He takes the car. It doesn’t matter whether he said, you know, I want to kill her because I want to see what it’s like ... to plunge a knife into somebody . . . [o]r . . . you know, I want to get my money back ... I want to get those earrings, and then during the use of that force or afterwards taking advantage of that killing he takes them back, and it’s still a robbery, it’s still a felony murder, and the defendant is guilty either way of first-degree murder under the felony-murder rule.

Thus, like the prosecutor’s logic in Nay, the prosecutor’s closing remarks in this case rested on a flawed premise — that for purposes of felony murder it is irrelevant whether the defendant intended to rob the victim before using the force that led to the victim’s death.

Consistent with our harmless-error review in Nay, we reiterate “that an otherwise valid conviction should not be set aside if the reviewing court may confidently say, on the whole record, that the constitutional error was harmless beyond a reasonable doubt.” 62 More specifically, in reviewing Stromberg error for harmlessness, we are not confined to considering whether the jury actually determined guilt under a valid theory, but may look beyond what the jury actually found to what a rational jury would have found if properly instructed. 63 Thus, the evidence presented to the jury and the jury’s other findings are relevant to our harmless-error review.

The similarity between Nay and this case ends when we turn to the facts and evidence presented. Although Cortinas contended at trial that the murder was impulsive and thus not deliberate and premeditated, the evidence overwhelmingly demonstrates the contrary. Cortinas confessed to the killing — twice—and led authorities to Kercher’s body. He admitted that he strangled Kercher for over an hour, relenting at times only to determine if she had finally stopped breathing. Failing to kill Kercher after an hour, he changed course and broke her neck. Then, after binding Kercher’s head and wrists and transporting her to the desert, Cortinas further ensured her death by stabbing her in the back three times for the admitted purpose of flooding her lungs with blood. 64

Based on this evidence alone, we conclude that a rational jury would have found that the murder was willful, deliberate, and premeditated. Nevertheless, the evidence also suggests that Cortinas had long contemplated strangling a victim and killed Kercher to satisfy his own morbid curiosity. Perhaps most notably, in contrast to Nay, Cortinas did not claim self-defense, let alone attempt to minimize his responsibility for this crime. Moreover, turning to the actual verdict in this case, the jury found that Cortinas had committed this killing with a deadly weapon, a ligature, which he held to Kercher’s neck for over an hour before finally deciding to break her neck with his hands. As we have noted previously, the use of a ligature and the time required to strangle a person are legitimate circumstances from which to infer that a killing is willful, deliberate, and premeditated. 65 Based on the evidence, we can confidently say beyond a reasonable doubt that presenting the invalid felony-murder theory to the jury in this case was harmless.

Robbery conviction

Cortinas next challenges his robbery conviction, claiming that the jury was improperly permitted to convict him of that offense without proof that he formed the intent to rob Kercher before she was killed. Reviewing this issue de novo, 66 we disagree and reaffirm our long-standing principle that the intent required for robbery need not be contemporaneous with the application of force or intimidation under NRS 200.380 to complete the elements of robbery. 67 Further, we reject Cortinas’ claim that the evidence is insufficient to support his robbery conviction.

Although in Nay we rejected the use of an “afterthought” robbery as a predicate felony for felony murder, citing Leonard v. State, we reiterated nevertheless that a “robbery can occur after death.” 68 Cortinas urges us to reexamine this holding. Although we decline this invitation, we take this opportunity to explain how existing law supports our prior decisions regarding robbery of a deceased victim.

Under NRS 200.380, robbery is defined as “the unlawful taking of personal property from the person of another, or in his presence, against his will, by means of force or violence or fear of injury, immediate or future, to his person or property.’ ’ We explained in Leonard that under NRS 200.380, “ ‘it is irrelevant when the intent to steal the property is formed,’ and it is not necessary that force or violence involved in the robbery ‘be committed with the specific intent to commit robbery.’ ” 69 Accordingly, we held that “[t]he jury need not be instructed that robbery requires [the] intent to take property from a living person.’ ’ 70

Contrary to Cortinas’ arguments, Leonard does not allow a person to begin and complete a robbery on a deceased victim. 71 Because a deceased person is no longer sensitive to force or coercion, the force or coercion needed to commit a robbery under NRS 200.380 must occur while the victim is alive. Reading Leonard for the opposite proposition would be absurd since it would criminalize as robbery the actions of a stranger who scavenges articles from an abandoned body. Indeed, by challenging his robbery conviction on grounds that the State failed to prove that he intended to rob the victim before killing her, Cortinas, in effect, portrays himself on appeal as the legal equivalent of a stranger in the desert, and therefore beyond the reach of the robbery statute. In doing so, how ever, Cortinas ignores a critical aspect of Leonard and his own case — both Leonard and Cortinas used force to kill the robbery victim and then took advantage of the situation they had created through that use of force to take the victim’s property.

To the extent that any ambiguity exists, we reiterate that under Leonard the taking required for robbery may occur after the victim is dead so long as the defendant’s use of force or coercion — for whatever purpose — occurs while the victim was alive and the defendant took advantage of the terrifying situation he created to flee with the victim’s property. 72 Leonard therefore does not permit a robbery to be perpetrated from beginning to end on a deceased victim.

Consistent with our holding in Leonard, the jury in this case received an instruction stating that

although acts of violence and intimidation preceded the actual taking of the property and may have been primarily intended for another purpose, it is enough to support the charge of robbery when a person takes the property by taking advantage of the terrifying situation he created.

While this instruction rendered the timing of Cortinas’ intent irrelevant to a determination of guilt, it required the jury to find that force or coercion preceded Kercher’s death. Thus, this instruction comprises a correct statement of the law. Moreover, since the jury did not need to find that Cortinas specifically intended to rob Kercher before she was killed, 73 Cortinas’ proposed instruction that would have required the jury to make that finding was properly denied. 74

Further, we reject Cortinas’ challenge to the sufficiency of the evidence to support his robbery conviction. Viewed in the light most favorable to the prosecution, 75 there is sufficient evidence for a rational juror to infer that Cortinas took Kercher’s property by taking advantage of the situation he created through the use of force. After strangling Kercher, Cortinas used her car to drive her body to Boulder City. During the same period, Cortinas fished through Kercher’s purse and took, among other things, the $150 that he had paid her earlier that night. Cortinas also took Kercher’s diamond earrings. It was Cortinas’ use of force to kill Kercher that allowed him to then take her property. We therefore conclude that sufficient evidence supports Cortinas’ robbery conviction. 76

CONCLUSION

In this appeal, we retreat from the absolute certainty approach to Stromberg error that we adopted in Bolden and conclude that the Chapman standard of harmless-error review applies when a general verdict may rest on a legally valid or a legally invalid alternative theory of liability. Given the overwhelming evidence presented at trial and the jury’s actual findings, we conclude that the availability of an erroneous felony-murder theory in this case was harmless. Moreover, we clarify that a robbery may occur after a killing so long as some force or coercion preceded death, and we conclude that sufficient evidence supports Cortinas’ robbery conviction. Accordingly, we affirm the judgment of conviction.

Hardesty and Douglas, JJ., concur.

1 283 U.S. 359, 368 (1931).

2 191 F.3d 1053, 1063 (9th Cir. 1999), overruled, on other grounds by Payton v. Woodford, 346 F.3d 1204, 1217 n.18 (9th Cir. 2003), rev’d, 544 U.S. 133 (2005).

3 121 Nev. 908, 924, 124 P.3d 191, 201 (2005).

4 123 Nev. 326, 333-34, 167 P.3d 430, 435 (2007).

5 386 U.S. 18, 24 (1967).

6 Cortinas told police during his confession that, while he kept this keychain lanyard on his person, he did not keep any keys on it.

7 Cortinas also drew pictures of the lanyard, Kercher’s earrings, and the butterfly knife during his confession.

8 123 Nev. 326,. 332, 167 P.3d 430, 434 (2007).

9

10 Crawford, v. State, 121 Nev. 744, 748, 121 P.3d 582, 585 (2005).

11 Nay, 123 Nev. at 330, 167 P.3d at 433.

12 Id. at 332-33, 167 P.3d at 434-35. NRS 200.030(l)(b) defines felony murder as murder that is “[c]ommitted in the perpetration or attempted perpetration of . . . robbery.”

13 Id. at 333, 167 P.3d at 435.

14 Id.

15 283 U.S. 359 (1931).

16 Id. at 368. Stromberg was convicted under a general jury verdict of violating a state statute that prohibited the public display of a red flag for any of three purposes: (1) opposing organized government, (2) inviting anarchistic action, or (3) aiding seditious propaganda. Id. at 361. The Court concluded that the first purpose under the statute (opposing government) was unconstitutional because it embraced protected speech, thus requiring the Court to determine if the verdict could stand given that there were still two constitutional alternative theories of liability under the statute. Id. at 369-70.

17 See Yates v. United States, 354 U.S. 298, 312 (1957), overruled on other grounds by Burks v. United States, 437 U.S. 1, 8, 18 (1978). The Court has declined to extend Stromberg in other contexts and has expressed some disapproval of its extension in Yates. See Griffin v. United States, 502 U.S. 46, 59 (1991) (declining to extend Stromberg and Yates to general-verdict convictions that may have rested on an independent though factually inadequate theory of liability); cf. Zant v. Stephens, 462 U.S. 862, 881 (1983) (refusing to set aside a capital sentence under the Stromberg rule despite the invalidity of one of multiple statutory aggravators submitted to the jury).

18 121 Nev. 908, 923-24, 124 P.3d 191, 194 (2005).

19 Id. at 923, 124 P.3d at 201.

20 Id. at 923-24, 124 P.3d at 201.

21 191 F.3d 1053 (9th Cir. 1999), overruled on other grounds by Payton v. Woodford, 346 F.3d 1204, 1217 n.18 (9th Cir. 2003), rev'd, 544 U.S. 133 (2005).

22 Id. at 1063 (quoting Ficklin v. Hatcher, 177 F.3d 1147, 1151 (9th Cir. 1999)).

23 121 Nev. at 924, 124 P.3d at 202.

24 Id. In Bolden, as further support for confining its review under the absolute certainty approach to the verdict form itself, as opposed to the record as a whole, the court cited Neder v. United States, 527 U.S. 1, 18 (1999), the flagship case for reviewing the harmlessness of instructional errors involving omitted or misdescribed elements. 121 Nev. at 924 n.57, 124 P.3d at 202 n.57. The citation of Neder is puzzling, however, if Stromberg and Neder articulate two mutually exclusive approaches to whether reversal is required, as Cortinas suggests.

25 455 F.3d 1080 (9th Cir. 2006).

26 Id. at 1083-85 (citing People v. Lee, 738 P.2d 752, 754 (Cal. 1987) (concluding that implied malice will not support a charge of attempted murder)).

27 Id. at 1086 (internal citation omitted).

28 See Neder, 527 U.S. at 8.

29 Lara, 455 F.3d at 1087.

30 Neder, 527 U.S. at 14.

31 See id. at 8.

32 See Bolden v. State, 121 Nev. 908, 924, 124 P.3d 191, 201 (2005) (quoting Keating v. Hood, 191 F.3d 1053, 1063 (9th Cir. 1999)).

33 From one vantage point, the Ninth Circuit in Lara, in effect, reasoned that taking an absolute certainty approach to Stromberg error was justified because Stromberg error is structural, but under that approach such error was only structural under certain circumstances. See Pulido v. Chrones, 487 F.3d 669, 676 (9th Cir. 2007) (characterizing Lara as holding that the option to convict under an invalid alternative theory is structural error, but would merit reversal only if a reviewing court could not be absolutely certain that the jury did not convict under a valid theory), cert. granted, 552 U.S. 1230 (2008). This reasoning is untenable, however, because it suggests that a structural-error determination occurs on a case-by-case basis, when the distinction between structural and nonstructural error is per se and dichotomous and, therefore, easy to draw at first glance. See Neder, 527 U.S. at 14. But see id. at 37 (Scalia, J., concurring in part and dissenting in part) (suggesting that a case-by-case approach is compatible with determining structural error).

34 See Arizona v. Fulminante, 499 U.S. 279, 306-10 (1991).

35 Id. at 307-08.

36 Neder, 527 U.S. at 8 (quoting Fulminante, 499 U.S. at 310).

37 Id. (quoting Rose v. Clark, 478 U.S. 570, 577 (1986)).

38

39 Id. (quoting Fulminante, 499 U.S. at 306).

40 Id. (quoting Johnson v. United States, 520 U.S. 461, 468 (1997)).

41 In Johnson, the United States Supreme Court catalogued the errors that it has identified as structural: the complete denial of counsel, trial judge bias, racial discrimination in the jury selection process, denial of the right of self-representation, denial of a public trial, and an erroneous reasonable-doubt instruction. 520 U.S. at 468-69. This list is exhaustive.

42 121 Nev. 908, 923-24, 124 P.3d 191, 201 (2005).

43 See, e.g., Sandstrom v. Montana, 442 U.S. 510 (1979); Vites v. United States, 354 U.S. 298 (1957), overruled on other grounds by Burks v. United States, 437 U.S. 1, 8, 18 (1978); Stromberg v. California, 283 U.S. 359 (1931). We note that the United States Supreme Court recently granted a petition for certiorari in a case that appears to raise the question, albeit in a habeas context, of whether Stromberg error is subject to harmless-error review. See Pulido v. Chrones, 487 F.3d 669 (9th Cir. 2007), cert. granted, 552 U.S. 1230 (2008).

44 Neder, 527 U.S. 11 (quoting Sullivan v. Louisiana, 508 U.S. 275, 281-82 (1993) (internal citation omitted)).

45 See Coliman v. State, 116 Nev. 687, 721, 7 P.3d 426, 447 (2000); Wegner v. State, 116 Nev. 1149, 1155-56, 14 P.3d 25, 30 (2000), overruled on other grounds by Rosas v. State, 122 Nev. 1258, 1267 n.26, 1269, 147 P.3d 1101, 1108 n.26, 1109 (2006); see also Nay v. State, 123 Nev. 326, 333-34, 167 P.3d 430, 435 (2007); Sharma v. State, 118 Nev. 648, 657-58, 56 P.3d 868, 873-74 (2002).

46 508 U.S. at 278.

47 Id. at 280.

48

49 See Neder, 527 U.S. at 12-13. In Neder, for example, the Supreme Court rejected the argument that a jury instruction that omitted an element of an offense amounted to structural error since such a conclusion could not be squared with its harmless-error decisions, including California v. Roy and Pope v. Illinois. In those cases, the jury was precluded from rendering a “complete verdict” on every element of the offense because of a flawed jury instruction; nonetheless, the instructional error in each case was reviewed for harmless error. In Pope, which concerned an obscenity prosecution, the jury was misinstructed under a local obscenity statute that it could assess whether allegedly obscene material lacked serious value under “community standards” as opposed to the constitutional “reasonable person” standard. 481 U.S. 497, 499-504 (1987). Likewise, in Roy, a federal habeas case, the trial court failed to instruct the jury that it could convict the defendant of first-degree murder on an aider and abettor theory only if it found that the defendant had the “intent or purpose” of aiding the confederate’s crime. 519 U.S. 2, 3-6 (1996). What is important for our purposes is that in both Pope and Roy a single flawed theory of criminal liability was presented to the jury. Nevertheless, in both of these single-theory cases, harmless-error review was applied.

50 Becht v. U.S., 403 F.3d 541, 548 (8th Cir. 2005).

51 Id. (quoting Quigley v. Vose, 834 F.2d 14, 16 (1st Cir. 1987)); see also Pulido v. Chrones, 487 F.3d 669, 676-78 (9th Cir. 2007) (O’Scannlain, L, concurring), cert. granted, 552 U.S. 1230 (2008).

52 In retreating from Bolden, we also retreat from Phillips v. State, in which we reviewed a general guilty verdict that may have rested on a legally insufficient theory of libel under the extortion statute under Stromberg and its progeny instead of performing Chapman harmless-error review. 121 Nev. 591, 600, 119 P.3d 711, 717 (2005).

53

54 123 Nev. 326, 328-29, 167 P.3d 430, 431-32 (2007).

55 Id. at 328-29, 167 P.3d at 432.

56 Id. at 334, 167 P.3d at 435-36.

57 Id. at 330, 167 P.3d at 433.

58 Id. at 334, 167 P.3d at 435.

59

60 Id. at 334, 167 P.3d at 436.

61 Relying on Leonard v. State, 114 Nev. 1196, 1210, 969 P.2d 288, 297 (1998), the district court instructed the jury on robbery as follows:

Robbery is not confined to a fixed locus, but may spread over considerable and varying periods of time. All matters immediately antecedent to and having direct causal connection with the robbery are deemed so closely connected with it as to form in reality a part of the occurrence. Thus, although acts of violence and intimidation preceded the actual taking of the property and may have been primarily intended for another purpose, it is enough to support the charge of robbery when a person takes the property by taking advantage of the terrifying situation he created.

62 Delaware v. Van Arsdall, 475 U.S. 673, 681 (1986).

63 U.S. v. Holly, 488 F.3d 1298, 1307 (10th Cir. 2007) (citing Neder v. United States, 527 U.S. 1, 16-18 n.l (1999)).

64 Indeed, during closing arguments, Cortinas’ defense counsel was forced to concede that “this is not a case where the State ha[d] simply failed to meet [its] burden of proof.”

65 See Leonard, 117 Nev. at 76, 17 P.3d at 411; Moser v. State, 91 Nev. 809, 812, 544 P.2d 424, 426 (1975).

66 Nay v. State, 123 Nev. 326, 330, 167 P.3d 430, 433 (2007) (whether a jury instruction is a correct statement of law is reviewed de novo).

67 See Leonard, 117 Nev. at 76-77, 17 P.3d at 412.

68 123 Nev. at 334, 167 P.3d at 436.

69 117 Nev. at 76-77, 17 P.3d at 412 (quoting Chappell v. State, 114 Nev. 1403, 1408, 972 P.2d 838, 841 (1998)).

70 Id. at 76, 17 P.3d at 412.

71 Specifically, citing the Indiana Court of Appeals case Buggs v. State, Cortinas argues that beginning and completing a robbery of a dead person is a legal impossibility under NRS 200.380 because the elements of robbery cannot be satisfied if the victim is deceased. 844 N.E.2d 195, 203-04 (Ind. Ct. App. 2006). Although this is true, our prior caselaw has never permitted it.

72 Leonard, 117 Nev. at 77, 17 P.3d at 412.

73

74 Vallery v. State, 118 Nev. 357, 372, 46 P.3d 66, 77 (2002) (jury instructions that misstate applicable law or are substantially covered by other instructions may be refused).

75 See Gaxiola v. State, 121 Nev. 638, 650, 119 P.3d 1225, 1233 (2005).

76 In addition to the specific contentions addressed in this opinion, Cortinas also raises a fair-cross-section challenge based on the allegedly underrepresentative Hispanic composition of his venire, asserts that the State failed to adequately prove the corpus delicti of his convictions and deadly weapon enhancement, argues that his statements to police and the consent to search his bedroom were improperly obtained, claims that various items of evidence were irrelevant, alleges various instances of prosecutorial misconduct, and challenges the decision to permit Kercher’s parents to testify at his sentencing. Having carefully considered these contentions, we conclude that none warrant reversal.