Evans v. Michigan, 568 U.S. 313 (2013)

NOTICE: This opinion is subject to formal revision before publication in the preliminary print of the United States Reports. Readers are requested to notify the Reporter of Decisions, Supreme Court of the United States, Washington, D. C. 20543, of any typographical or other formal errors, in order that corrections may be made before the preliminary print goes to press.

SUPREME COURT OF THE UNITED STATES

_________________

No. 11–1327

_________________

LAMAR EVANS, PETITIONER v. MICHIGAN

on writ of certiorari to the supreme court of michigan

[February 20, 2013]

Justice Sotomayor delivered the opinion of the Court.

When the State of Michigan rested its case at petitioner Lamar Evans’ arson trial, the court entered a directed verdict of acquittal, based upon its view that the State had not provided sufficient evidence of a particular element of the offense. It turns out that the unproven “element” was not actually a required element at all. We must de- cide whether an erroneous acquittal such as this nevertheless constitutes an acquittal for double jeopardy purposes, which would mean that Evans could not be retried. This Court has previously held that a judicial acquittal premised upon a “misconstruction” of a criminal statute is an “acquittal on the merits . . . [that] bars retrial.” Arizona v. Rumsey, 467 U.S. 203, 211 (1984). Seeing no meaningful constitutional distinction between a trial court’s “misconstruction” of a statute and its erroneous addition of a statutory element, we hold that a midtrial acquittal in these circumstances is an acquittal for double jeopardy purposes as well.

I

The State charged Evans with burning “other real property,” a violation of Mich. Comp. Laws §750.73 (1981). The State’s evidence at trial suggested that Evans had burned down an unoccupied house. At the close of the State’s case, however, Evans moved for a directed ver- dict of acquittal. He pointed the court to the applicable Michigan Criminal Jury Instructions, which listed as the “Fourth” element of the offense “that the building was not a dwelling house.” 3 Mich. Crim. Jury Instr. §31.3, p. 31–7 (2d ed., Supp. 2006/2007). And the commentary to the Instructions emphasized, “an essential element is that the structure burned is not a dwelling house.” Id., at 31–8. Evans argued that Mich. Comp. Laws §750.72 criminal- izes common-law arson, which requires that the structure burned be a dwelling, while the provision under which he was charged, §750.73, covers all other real property.[1] Persuaded, the trial court granted the motion. 491 Mich. 1, 8, 810 N.W.2d 535, 539 (2012). The court explained that the “ ‘testimony [of the homeowner] was this was a dwelling house,’ ” so the nondwelling requirement of §750.73 was not met. Ibid.

On the State’s appeal, the Michigan Court of Appeals reversed and remanded. 288 Mich. App. 410, 794 N.W.2d 848 (2010). Evans had conceded, and the court held, that under controlling precedent, burning “other real property” is a lesser included offense under Michigan law, and disproving the greater offense is not required. Id., at 416, 794 N. W. 2d, at 852 (citing People v. Antonelli, 66 Mich. App. 138, 140, 238 N.W.2d 551, 552 (1975) (on rehearing)).[2] The court thus explained it was “undisputed that the trial court misperceived the elements of the offense with which [Evans] was charged and erred by directing a verdict.” 288 Mich. App., at 416, 794 N. W. 2d, at 852. But the court rejected Evans’ argument that the Double Jeopardy Clause barred retrial. Id., at 421–422, 794 N. W. 2d, at 856.

In a divided decision, the Supreme Court of Michigan affirmed. It held that “when a trial court grants a defendant’s motion for a directed verdict on the basis of an error of law that did not resolve any factual element of the charged offense, the trial court’s ruling does not constitute an acquittal for the purposes of double jeopardy and re- trial is therefore not barred.” 491 Mich., at 4, 810 N. W. 2d, at 536–537.

We granted certiorari to resolve the disagreement among state and federal courts on the question whether retrial is barred when a trial court grants an acquittal be- cause the prosecution had failed to prove an “element” of the offense that, in actuality, it did not have to prove.[3] 567 U. S. ___ (2012). We now reverse.

II

A

In answering this question, we do not write on a clean slate. Quite the opposite. It has been half a century since we first recognized that the Double Jeopardy Clause bars retrial following a court-decreed acquittal, even if the ac- quittal is “based upon an egregiously erroneous foundation.” Fong Foo v. United States, 369 U.S. 141, 143 (1962) (per curiam). A mistaken acquittal is an acquittal nonetheless, and we have long held that “[a] verdict of acquittal . . . could not be reviewed, on error or otherwise, without putting [a defendant] twice in jeopardy, and thereby violating the Constitution.” United States v. Ball, 163 U.S. 662, 671 (1896).

Our cases have applied Fong Foo’s principle broadly. An acquittal is unreviewable whether a judge directs a jury to return a verdict of acquittal, e.g., Fong Foo, 369 U. S., at 143, or forgoes that formality by entering a judgment of acquittal herself. See Smith v. Massachusetts, 543 U.S. 462, 467–468 (2005) (collecting cases). And an acquittal precludes retrial even if it is premised upon an erroneous decision to exclude evidence, Sanabria v. United States, 437 U.S. 54, 68–69, 78 (1978); a mistaken understanding of what evidence would suffice to sustain a conviction, Smith, 543 U. S., at 473; or a “misconstruction of the stat- ute” defining the requirements to convict, Rumsey, 467 U. S., at 203, 211; cf. Smalis v. Pennsylvania, 476 U.S. 140, 144–145, n. 7 (1986). In all these circumstances, “the fact that the acquittal may result from erroneous evidentiary rulings or erroneous interpretations of governing legal principles affects the accuracy of that determination, but it does not alter its essential character.” United States v. Scott, 437 U.S. 82, 98 (1978) (internal quotation marks and citation omitted).

Most relevant here, our cases have defined an acquittal to encompass any ruling that the prosecution’s proof is insufficient to establish criminal liability for an offense. See ibid., and n. 11; Burks v. United States, 437 U.S. 1, 10 (1978); United States v. Martin Linen Supply Co., 430 U.S. 564, 571 (1977). Thus an “acquittal” includes “a ruling by the court that the evidence is insufficient to convict,” a “factual finding [that] necessarily establish[es] the criminal defendant’s lack of criminal culpability,” and any other “rulin[g] which relate[s] to the ultimate question of guilt or innocence.” Scott, 437 U. S., at 91, 98, and n. 11 (internal quotation marks omitted). These sorts of substantive rulings stand apart from procedural rulings that may also terminate a case midtrial, which we generally refer to as dismissals or mistrials. Procedural dismissals include rulings on questions that “are unrelated to factual guilt or innocence,” but “which serve other purposes,” including “a legal judgment that a defendant, although criminally culpable, may not be punished” because of some problem like an error with the indictment. Id., at 98, and n. 11.

Both procedural dismissals and substantive rulings result in an early end to trial, but we explained in Scott that the double jeopardy consequences of each differ. “[T]he law attaches particular significance to an acquittal,” so a merits-related ruling concludes proceedings absolutely. Id., at 91. This is because “[t]o permit a second trial after an acquittal, however mistaken the acquittal may have been, would present an unacceptably high risk that the Government, with its vastly superior resources, might wear down the defendant so that ‘even though innocent he may be found guilty,’ ” ibid. (quoting Green v. United States, 355 U.S. 184, 188 (1957)). And retrial following an acquittal would upset a defendant’s ex- pectation of repose, for it would subject him to additional “embarrassment, expense and ordeal” while “compelling him to live in a continuing state of anxiety and insecurity.” Id., at 187. In contrast, a “termination of the proceedings against [a defendant] on a basis unrelated to factual guilt or innocence of the offense of which he is accused,” 437 U. S., at 98–99, i.e., some procedural ground, does not pose the same concerns, because no expectation of finality attaches to a properly granted mistrial.

Here, “it is plain that the [trial court] . . . evaluated the [State’s] evidence and determined that it was legally insufficient to sustain a conviction.” Martin Linen, 430 U. S., at 572. The trial court granted Evans’ motion under a rule that requires the court to “direct a verdict of acquittal on any charged offense as to which the evidence is insufficient to support conviction.” Mich. Rule Crim. Proc. 6.419(A) (2012). And the court’s oral ruling leaves no doubt that it made its determination on the basis of “ ‘[t]he testimony’ ” that the State had presented. 491 Mich., at 8, 810 N. W. 2d, at 539. This ruling was not a dismissal on a procedural ground “unrelated to factual guilt or innocence,” like the question of “preindictment delay” in Scott, but rather a determination that the State had failed to prove its case. 437 U. S., at 98, 99. Under our precedents, then, Evans was acquitted.

There is no question the trial court’s ruling was wrong; it was predicated upon a clear misunderstanding of what facts the State needed to prove under State law. But that is of no moment. Martin Linen, Sanabria, Rumsey, Smalis, and Smith all instruct that an acquittal due to insufficient evidence precludes retrial, whether the court’s evaluation of the evidence was “correct or not,” Martin Linen, 430 U. S., at 571, and regardless of whether the court’s decision flowed from an incorrect antecedent ruling of law. Here Evans’ acquittal was the product of an “erroneous interpretatio[n] of governing legal principles,” but as in our other cases, that error affects only “the accuracy of [the] determination” to acquit, not “its essential character.” Scott, 437 U. S., at 98 (internal quotation marks omitted).

B

The court below saw things differently. It identified a “constitutionally meaningful difference” between this case and our previous decisions. Those cases, the court found, “involve[d] evidentiary errors regarding the proof needed to establish a factual element of the . . . crimes at issue,” but still ultimately involved “a resolution regarding the sufficiency of the factual elements of the charged offense.” 491 Mich., at 14–15, 810 N. W. 2d, at 542–543. When a court mistakenly “identifie[s] an extraneous element and dismisse[s] the case solely on that basis,” however, it has “not resolve[d] or even address[ed] any factual element necessary to establish” the offense. Id., at 15, 20, 810 N. W. 2d, at 543, 546. As a result, the court below reasoned, the case terminates “based on an error of law unrelated to [the] defendant’s guilt or innocence on the ele-ments of the charged offense,” and thus falls outside the definition of an acquittal. Id., at 21, 810 N. W. 2d, at 546.

We fail to perceive the difference. This case, like our previous ones, involves an antecedent legal error that led to an acquittal because the State failed to prove some fact it was not actually required to prove. Consider Rumsey. There the trial court, sitting as sentencer in a capital case involving a murder committed during a robbery, mistakenly held that Arizona’s statutory aggravating factor describing killings for pecuniary gain was limited to murders for hire. Accordingly, it found the State had failed to prove the killing was for pecuniary gain and sentenced the defendant to life imprisonment. After the State successfully appealed and obtained a death sentence on remand, we held that retrial on the penalty phase question was a double jeopardy violation.[4]

The only relevant difference between that situation and this one is that in Rumsey the trial court’s error was called a “misinterpretation” and a “misconstruction of the statute,” 467 U. S., at 207, 211, whereas here the error has been designated the “erroneous addition of [an] extraneous element to the charged offense.” 491 Mich., at 3–4, 810 N. W. 2d, at 536. But we have emphasized that labels do not control our analysis in this context; rather, the substance of a court’s decision does. See Smalis, 476 U. S., at 144, n. 5; Scott, 437 U. S., at 96–97; Martin Linen, 430 U. S., at 571. The error in Rumsey could just as easily have been characterized as the erroneous addition of an element of the statutory aggravating circumstance: that the homicide be a murder-for-hire. Conversely, the error here could be viewed as a misinterpretation of the statute’s phrase “building or other real property” to exclude dwellings.[5] This is far too fine a distinction to be meaningful, and we reject the notion that a defendant’s constitutional rights would turn on the happenstance of how an appellate court chooses to describe a trial court’s error.

Echoing the Michigan Supreme Court, the State and the United States, as well as the dissent, emphasize Martin Linen’s description of an acquittal as the “resolution, correct or not, of some or all of the factual elements of the offense charged.” 430 U. S., at 571 (emphasis added); see Brief for Respondent 11–17; see Brief for United States as Amicus Curiae 11–15 (hereinafter U. S. Brief); see post, at 6–8. They observe that the Double Jeopardy Clause protects against being twice placed in jeopardy for the same “offence,” U. S. Const., Amdt. 5, cl. 2, and they note that an offense comprises constituent parts called elements, which are facts that must be proved to sustain a conviction. See, e.g., United States v. Dixon, 509 U.S. 688, 696–697 (1993). Consequently, they argue, only if an actual element of the offense is resolved can it be said that there has been an acquittal of the offense, because “ ‘innocence of the charged offense’ cannot turn on something that is concededly not an element of the offense.” U. S. Brief 15. Because Evans’ trial ended without resolution of even one actual element, they conclude, there was no acquittal.

This argument reads Martin Linen too narrowly, and it is inconsistent with our decisions since then. Our focus in Martin Linen was on the significance of a judicial acquittal under Fed. Rule Crim. Proc. 29. The District Court in that case had “evaluated the Government’s evidence and determined that it was legally insufficient to sustain a con-viction.” 430 U. S., at 572. That determination of nonculpability was enough to make the acquittal akin to a jury verdict; our holding did not depend upon defining the “elements” of the offense. As we have explained, supra, at 5–6, Scott confirms that the relevant distinction is between judicial determinations that go to “the criminal defendant’s lack of criminal culpability,” and those that hold “that a defendant, although criminally culpable, may not be punished because of a supposed” procedural error. 437 U. S., at 98. Culpability (i.e., the “ultimate question of guilt or innocence”) is the touchstone, not whether any particular elements were resolved or whether the determination of nonculpability was legally correct. Id., at 98, n. 11 (internal quotation marks omitted).

Perhaps most inconsistent with the State’s and United States’ argument is Burks. There we held that when a defendant raises insanity as a defense, and a court decides the “Government ha[s] failed to come forward with sufficient proof of [the defendant’s] capacity to be responsible for criminal acts,” the defendant has been acquitted because the court decided that “criminal culpability ha[s] not been established.” 437 U. S., at 10. Lack of insanity was not an “element” of Burks’ offense, bank robbery by use of a dangerous weapon. See 18 U. S. C. §2113(d) (1976 ed.). Rather, insanity was an affirmative defense to criminal liability. Our conclusion thus depended upon equating a judicial acquittal with an order finding insufficient evidence of culpability, not insufficient evidence of any particular element of the offense.[6]

In the end, this case follows those that have come before it. The trial court’s judgment of acquittal resolved the question of Evans’ guilt or innocence as a matter of the sufficiency of the evidence, not on unrelated procedural grounds. That judgment, “however erroneous” it was, precludes reprosecution on this charge, and so should have barred the State’s appeal as well. Sanabria, 437 U. S., at 69.

III

A

The State, supported by the United States, offers three other reasons why the distinction drawn by the court be- low should be maintained. None persuades us.

To start, the State argues that unless an actual element of the offense is resolved by the trial court, the only way to know whether the court’s ruling was an “acquittal” is to rely upon the label used by the court, which would wrongly allow the form of the trial court’s action to control. Brief for Respondent 17–18, 21–22. We disagree. Our decision turns not on the form of the trial court’s action, but rather whether it “serve[s]” substantive “purposes” or procedural ones. Scott, 437 U. S., at 98, n. 11. If a trial court were to announce, midtrial, “The defendant shall be acquitted because he was prejudiced by preindictment delay,” the Double Jeopardy Clause would pose no barrier to reprosecution, notwithstanding the “acquittal” label. Cf. Scott, 437 U.S. 82. Here we know the trial court acquitted Evans, not because it incanted the word “acquit” (which it did not), but because it acted on its view that the prosecution had failed to prove its case.

Next, the State and the United States fear that if the grounds for an acquittal are untethered from the actual elements of the offense, a trial court could issue an unreviewable order finding insufficient evidence to convict for any reason at all, such as that the prosecution failed to prove “that the structure burned [was] blue.” Brief for Respondent 16–17; U. S. Brief 15. If the concern is that there is no limit to the magnitude of the error that could yield an acquittal, the response is that we have long held as much. See supra, at 4. If the concern is instead that our holding will make it easier for courts to insulate from review acquittals that are granted as a form of nullifi- cation, see Brief for Respondent 30, n. 58, we reject the premise. We presume here, as in other contexts, that courts exercise their duties in good faith. Cf. Harrington v. Richter, 562 U. S. ___, ___ (2011) (slip op., at 13).

Finally, the State suggests that because Evans induced the trial court’s error, he should not be heard to complain when that error is corrected and the State wishes to retry him. Brief for Respondent 32–33; cf. id., at 5–9. But we have recognized that “most [judgments of acquittal] re- sult from defense motions,” so “[t]o hold that a defendant waives his double jeopardy protection whenever a trial court error in his favor on a midtrial motion leads to an acquittal would undercut the adversary assumption on which our system of criminal justice rests, and would vitiate one of the fundamental rights established by the Fifth Amendment.” Sanabria, 437 U. S., at 78 (citation omitted).[7] It is true that when a defendant persuades the court to declare a mistrial, jeopardy continues and retrial is generally allowed. See United States v. Dinitz, 424 U.S. 600 (1976). But in such circumstances the defendant consents to a disposition that contemplates reprosecution, whereas when a defendant moves for acquittal he does not. See Sanabria, 437 U. S., at 75.

The United States makes a related argument. It contends that Evans could have asked the court to resolve whether nondwelling status is an element of the offense before jeopardy attached, so having elected to wait until trial was underway to raise the point, he cannot now claim a double jeopardy violation. U. S. Brief 22–25. The Government relies upon Lee v. United States, 432 U.S. 23 (1977), in which the District Court dismissed an indictment midtrial because it had failed to allege the required intent element of the offense. We held that retrial on a corrected indictment was not barred, because the dismissal was akin to a mistrial, not an acquittal. This was clear because the District Court had separately denied the defendant’s motion for judgment of acquittal, explaining that the defendant “ ‘has been proven [guilty] beyond any reasonable doubt in the world,’ ” while acknowledging that the error in the indictment required dismissal. Id., at 26–27. Because the defendant “invited the court to interrupt the proceedings before formalizing a finding on the merits” by raising the indictment issue so late, we held the principles governing a defendant’s consent to mistrial should apply. Id., at 28 (citing Dinitz, 424 U. S. 600).

The Government suggests the situation here is “functionally similar,” because “identifying the elements of an offense is a necessary step in determining the sufficiency of a charging document.” U. S. Brief 23. But we can- not ignore the fact that what the trial court actually did here was rule on the sufficiency of the State’s proof, not the sufficiency of the information filed against him. Lee demonstrates that the two need not rise or fall together. And even if the Government is correct that Evans could have challenged the charging document on the same legal theory he used to challenge the sufficiency of the evidence, it matters that he made only the latter motion, a motion that necessarily may not be made until trial is underway. Evans cannot be penalized for requesting from the court a ruling on the merits of the State’s case, as the Michigan Rules entitled him to do; whether he could have also brought a distinct procedural objection earlier on is beside the point.

B

In the alternative, the State and the United States ask us to reconsider our past decisions. Brief for Respondent 34–56 (suggesting overruling our cases since at least Fong Foo); U. S. Brief 27–32 (suggesting overruling Smith, Rumsey, and Smalis).[8] We declined to revisit our cases when the United States made a similar request in Smalis. 476 U. S., at 144; see Brief for United States as Amicus Curiae in Smalis v. Pennsylvania, O. T. 1985, No. 85–227, pp. 19–25. And we decline to do so here.

First, we have no reason to believe the existing rules have become so “unworkable” as to justify overruling precedent. Payne v. Tennessee, 501 U.S. 808, 827 (1991). The distinction drawn in Scott has stood the test of time, and we expect courts will continue to have little “difficulty in distinguishing between those rulings which relate to the ultimate question of guilt or innocence and those which serve other purposes.” 437 U. S., at 98, n. 11 (internal quotation marks omitted). See, e.g., United States v. Dionisio, 503 F.3d 78, 83–88 (CA2 2007) (collecting cases); 6 W. LaFave, J. Israel, N. King, & O. Kerr, Criminal Procedure §25.3(a), p. 629 (3d ed. 2007) (same).

Second, the logic of these cases still holds. There is no question that a jury verdict of acquittal precludes retrial, and thus bars appeal of any legal error that may have led to that acquittal. See Ball, 163 U. S., at 671. So, had the trial court here instructed the jury that it must find the burned structure was not a dwelling in order to convict, the jury would have acquitted Evans accordingly; “ ‘[a] jury is presumed to follow its instructions.’ ” Blueford v. Arkansas, 566 U. S. ___, ___ (2012) (slip op., at 6) (quoting Weeks v. Angelone, 528 U.S. 225, 234 (2000)). And that would have been the end of the matter. From that premise, Fong Foo’s holding follows: If a trial court instead exercises its discretion to direct a jury to return a verdict of acquittal, jeopardy also terminates notwithstanding any legal error, because there too it is the jury that returns an acquittal. And from there, Martin Linen’s conclusion is unavoidable: It should make no difference whether the court employs the formality of directing the jury to return an acquittal or whether the court enters an acquittal itself. Sanabria, Rumsey, Smalis, and Smith merely apply Fong Foo and Martin Linen in tandem: If a trial court makes an antecedent legal error (as in Fong Foo), and then grants a judgment of acquittal rather than directing the jury to acquit (as in Martin Linen), the result is an acquittal all the same.

In other words, there is no way for antecedent legal errors to be reviewable in the context of judicial acquittals unless those errors are also reviewable when they give rise to jury acquittals (contrary to the settled understanding that a jury verdict of acquittal is unreviewable), or unless we distinguish between juries that acquit pursuant to their instructions and judicial acquittals (notwithstand- ing that this is a purely formal distinction). Neither option has become more attractive with time. We therefore reiterate: “any contention that the Double Jeopardy Clause must itself . . . leave open a way of correcting legal errors is at odds with the well-established rule that the bar will attach to a preverdict acquittal that is patently wrong in law.” Smith, 543 U. S., at 473.

Finally, the State and the United States object that this rule denies the prosecution a full and fair opportunity to present its evidence to the jury, while the defendant reaps a “windfall” from the trial court’s unreviewable error. Brief for Respondent 6; U. S. Brief 31–32. But sovereigns are hardly powerless to prevent this sort of situation, as we observed in Smith, 543 U. S., at 474. Nothing obligates a jurisdiction to afford its trial courts the power to grant a midtrial acquittal, and at least two States disallow the practice. See Nev. Rev. Stat. §175.381(1) (2011); State v. Parfait, 96, 1814 (La. App. 1 Cir. 05/09/97), 693 So. 2d 1232, 1242. Many jurisdictions, including the federal system, allow or encourage their courts to defer consideration of a motion to acquit until after the jury returns a verdict, which mitigates double jeopardy concerns.[9] See Fed. Rule Crim. Proc. 29(b). And for cases such as this, in which a trial court’s interpretation of the relevant criminal statute is likely to prove dispositive, we see no reason why jurisdictions could not provide for mandatory continuances or expedited interlocutory appeals if they wished to prevent misguided acquittals from being entered.[10] But having chosen to vest its courts with the power to grant midtrial acquittals, the State must bear the corresponding risk that some acquittals will be granted in error.

*  *  *

We hold that Evans’ trial ended in an acquittal when the trial court ruled the State had failed to produce sufficient evidence of his guilt. The Double Jeopardy Clause thus bars retrial for his offense and should have barred the State’s appeal. The judgment of the Supreme Court of Michigan is

Reversed.

SUPREME COURT OF THE UNITED STATES

_________________

No. 11–1327

_________________

LAMAR EVANS, PETITIONER v. MICHIGAN

on writ of certiorari to the supreme court of michigan

[February 20, 2013]

Justice Alito, dissenting.

The Court holds that the Double Jeopardy Clause bars petitioner’s retrial for arson because his attorney managed to convince a judge to terminate petitioner’s first trial prior to verdict on the specious ground that the offense with which he was charged contains an imaginary “element” that the prosecution could not prove. The Court’s decision makes no sense. It is not consistent with the original meaning of the Double Jeopardy Clause; it does not serve the purposes of the prohibition against double jeopardy; and contrary to the Court’s reasoning, the trial judge’s ruling was not an “acquittal,” which our cases have “consistently” defined as a decision that “ ‘actually represents a resolution, correct or not, of some or all of the factual elements of the offense charged.’ ” E.g., Smith v. Massachusetts, 543 U.S. 462, 468 (2005) (quoting United States v. Martin Linen Supply Co., 430 U.S. 564, 571 (1977); emphasis added). For no good reason, the Court deprives the State of Michigan of its right to have one fair opportunity to convict petitioner, and I therefore respectfully dissent.

I

After Detroit police officers heard an explosion at a burning house, they observed petitioner running away from the building with a gasoline can. The officers pursued and ultimately apprehended petitioner, who admitted that he had burned down the house. No one was living in the house at the time of the fire.

If the house in question had been a “dwelling house,” petitioner could have been charged under Mich. Comp. Laws §750.72 (1981) for burning a dwelling, an offense punishable by imprisonment for up to 20 years. But petitioner was instead charged with “[b]urning other real property” in violation of Mich. Comp. Laws §750.73. This offense, which carries a maximum penalty of 10 years’ imprisonment, applies to “[a]ny person who wilfully or maliciously burns any building or other real property . . . other than those specified in [§750.72].” This crime is a lesser included offense of the crime of burning a dwelling house. The “necessary elements to prove either offense are the same, except to prove the greater [offense] it must be shown that the building is a dwelling.” 491 Mich. 1, 19–20, 810 N.W.2d 535, 545–546 (2012) (internal quotation marks omitted). To prove the lesser offense, however, “ ‘it is not necessary to prove that the building is not a dwelling.’ ” Id., at 20, 810 N. W. 2d, at 546 (emphasis added).

At the close of the prosecution’s case, petitioner’s attorney moved for a directed verdict on the ground that (1) the prosecution was required to prove, as an “element” of the charged offense, that “the building was not a dwelling” and (2) “the prosecution had failed to prove that the burned building was not a dwelling house.” Id., at 5, 810 N. W. 2d, at 537. The prosecutor responded by arguing that nothing in the charged offense requires proof that the building was not a dwelling, and the prosecutor requested “a moment” to “pull the statute” and “consult with [her] supervisors.” Id., at 5–7, 810 N. W. 2d, at 537–539. The trial judge denied the prosecutor’s requests and errone- ously concluded that the prosecution was required to prove that the burned building was not a dwelling. After determining that the State had not proved this nonexistent “element,” the trial judge granted petitioner’s motion for a directed verdict and entered an order that it labeled an “[a]cquittal.” App. to Pet. for Cert. 72.

The trial judge’s ruling was plainly wrong, and on appeal, defense counsel did not even attempt to defend its correctness, conceding that the judge had “wrongly added an extraneous element to the statute” under which his client was charged. 491 Mich., at 3, 810 N. W. 2d, at 536; see also 288 Mich. App. 410, 416, and n. 2, 794 N.W.2d 848, 852, and n. 2 (2010). The Michigan Court of Appeals agreed with this concession and went on to hold that the trial judge’s ruling did not constitute an “acquittal” for double jeopardy purposes because the ruling did not represent “a resolution in the defendant’s favor . . . of a fac- tual element necessary for a criminal conviction.” Id., at 421–422, 794 N. W. 2d, at 856 (internal quotation marks omitted). The Michigan Supreme Court affirmed, holding that when, as here, a trial judge erroneously adds an ex- tra “element” to a charged offense and subsequently determines that the prosecution did not prove that extra “element,” the trial judge’s decision is not based on the defendant’s guilt or innocence of the elements of the charged offense. 491 Mich., at 3–4, 19–21, 810 N. W. 2d, at 536–537, 545–546. Accordingly, the Michigan Supreme Court concluded that the judge’s ruling in this case “does not constitute an acquittal for the purposes of double jeopardy and retrial is . . . not barred.” Id., at 4, 810 N. W. 2d, at 537.

II

This Court now reverses the decision of the State Supreme Court, but the Court’s holding is supported by neither the original understanding of the prohibition against double jeopardy nor any of the reasons for that prohibition.

A

The prohibition against double jeopardy “had its origin in the three common-law pleas of autrefois acquit, autrefois convict, and pardon,” which “prevented the retrial of a person who had previously been acquitted, convicted, or pardoned for the same offense.” United States v. Scott, 437 U.S. 82, 87 (1978); see Crist v. Bretz, 437 U.S. 28, 33 (1978). As the Court has previously explained, “the common-law protection against double jeopardy historically applied only to charges on which a jury had rendered a verdict.” Smith, 543 U. S., at 466 (emphasis added).[1] As a result, the original understanding of the Clause, which is “hardly a matter of dispute,” Scott, supra, at 87, does not compel the Court’s conclusion that a defendant is acquitted for double jeopardy purposes whenever a judge issues a preverdict ruling that the prosecution has failed to prove a nonexistent “element” of the charged offense.

Although our decisions have expanded double jeopardy protection beyond its common-law origins, see, e.g., Smith, supra, at 466–467 (acknowledging the Court’s expansion of “the common-law protection against double jeopardy”); Crist, supra, at 33–34, I nonetheless count it significant that the result the Court reaches today finds no support in the relevant common-law analogues that “lie at the core of the area protected by the Double Jeopardy Clause,” see Scott, 437 U. S., at 96. And given how far we have departed from the common-law principles that applied at the time of the founding, we should at least ensure that our decisions in this area serve the underlying purposes of the constitutional prohibition against double jeopardy. See id., at 95–96, 100–101. Yet today’s decision fails to advance the purposes of the Double Jeopardy Clause.

B

The Double Jeopardy Clause is largely based on “the deeply ingrained principle that the State with all its resources and power should not be allowed to make repeated attempts to convict an individual for an alleged offense, thereby subjecting him to embarrassment, expense and or- deal and compelling him to live in a continuing state of anxiety and insecurity, as well as enhancing the possibility that even though innocent he may be found guilty.” Yeager v. United States, 557 U.S. 110, 117–118 (2009) (internal quotation marks omitted); see also Blueford v. Arkansas, 566 U. S. ___, ___ (2012) (slip op., at 5); Martin Linen, 430 U. S., at 569. Allowing retrial in the circumstances of the present case would not result in any such abuse. The prosecution would not be afforded a second opportunity to persuade the factfinder that its evidence satisfies the actual elements of the offense. Instead, because the trial judge’s ruling in the first trial was not based on an actual element of the charged offense, retrial would simply give the prosecution one fair opportunity to prove its case.

Allowing retrial in this case would not permit prosecutors “to make repeated attempts to convict an individual for an alleged offense,” Yeager, supra, at 117. It was petitioner, not the prosecutor, who sought to terminate the trial prior to verdict. Thus, contrary to the Court’s unexplained suggestion, see ante, at 5–6, “[t]his case hardly presents the specter of ‘an all-powerful state relentlessly pursuing a defendant who had either been found not guilty or who had at least insisted on having the issue of guilt submitted to the first trier of fact.’ ” Sattazahn v. Pennsylvania, 537 U.S. 101, 114–115 (2003) (quoting Scott, supra, at 96). On the contrary, this is a case in which defense counsel fooled the judge into committing an error that provided his client with an undeserved benefit, the termination of a trial that the defense obviously did not want to run to completion. The Double Jeopardy Clause does not require that the defense receive an even greater benefit, the protection provided by an acquittal. As this Court has repeatedly emphasized in double jeopardy cases, a State has an interest in receiving “one complete opportunity to convict those who have violated its laws,” Sattazahn, supra, at 115 (internal quotation marks omitted); Scott, supra, at 100, but today’s decision deprives the State of Michigan of this valuable right.

C

The Court’s decision also flies in the face of our established understanding of the meaning of an acquittal for double jeopardy purposes. The Double Jeopardy Clause provides that no person shall “be subject for the same offence to be twice put in jeopardy of life or limb.” U. S. Const., Amdt. 5 (emphasis added). Thus, “[d]ouble-jeopardy analysis focuses on the individual ‘offence’ charged.” Smith, 543 U. S., at 469, n. 3. And to determine what constitutes “the individual ‘offence’ charged,” ibid., the Court homes in on the elements of the offense. See United States v. Dixon, 509 U.S. 688, 696 (1993) (“In both the multiple punishment and multiple prosecution contexts, this Court has concluded that where the two of- fenses for which the defendant is punished or tried cannot survive the ‘same-elements’ test, the double jeopardy bar applies”). Consistent with the constitutional text’s focus on the “offence”—and thus the elements—with which a defendant is charged, the Court’s “double-jeopardy cases have consistently” defined an acquittal as a decision that “ ‘actually represents a resolution, correct or not, of some or all of the factual elements of the offense charged.’ ” Smith, supra, at 468 (quoting Martin Linen, supra, at 571); see also Scott, supra, at 97 (“[A] defendant is acquitted only when the ruling of the judge, whatever its label, actually represents a resolution in the defendant’s favor, correct or not, of some or all of the factual elements of the offense charged” (internal quotation marks and brackets omitted)).

Today, the Court effectively abandons the well-established definition of an acquittal. Indeed, in the face of our repeated holdings that an acquittal for double jeopardy purposes requires a “resolution, correct or not, of some or all of the factual elements of the offense charged,” Smith, supra, at 468; Martin Linen, supra, at 571; see also Scott, supra, at 97, the Court now declares that “the touchstone [is] not whether any particular elements were resolved,” ante, at 10 (emphasis added). Instead, the Court proclaims that the dispositive question is whether a midtrial termination represented a “procedural dismissa[l]” or a “substantive rulin[g],” ante, at 5. This reformulation of double jeopardy law is not faithful to our prece- dents—or to the Double Jeopardy Clause itself. The key question is not whether a ruling is “procedural” or “substantive” (whatever those terms mean in this context), but whether a ruling relates to the defendant’s factual guilt or innocence with respect to the “offence,” see U. S. Const., Amdt. 5—and thus the elements—with which he is charged. See Scott, supra, at 87, 97–99, and n. 11.

When a judge evaluates the evidence and determines that the prosecution has not proved facts that are legally sufficient to satisfy the actual elements of the charged offense, the ruling, however labeled, represents an acquittal because it is founded on the defendant’s factual innocence. See Martin Linen, 430 U. S., at 572. But when a judge manufactures an additional “element” of an offense and then holds that there is insufficient evidence to prove that extra “element,” the judge has not resolved the defendant’s “factual guilt or innocence” as to any of the actual elements of the offense.[2] Thus, the ruling, no matter what the judge calls it, does not acquit the defendant of the offense with which he is charged. No acquittal occurs when a criminal trial is terminated “on a basis unrelated to factual guilt or innocence of the offense of which [a defendant] is accused.” Scott, 437 U. S., at 87, 94–95, 98–99. “[I]n a case such as this the defendant, by deliberately choosing to seek termination of the proceedings against him on a basis unrelated to factual guilt or innocence of the offense of which he is accused, suffers no injury cognizable under the Double Jeopardy Clause if the Government is permitted to appeal from such a ruling of the trial court in favor of the defendant.” Id., at 98–99 (reasoning that, in such a case, the defendant was “neither acquitted nor convicted, because he himself successfully undertook to persuade the trial court not to submit the issue of guilt or innocence to the jury which had been empaneled to try him”).

III

Contrary to the Court’s opinion, its decision in this case is not supported by prior precedent. In all three of the principal cases on which the Court relies—Smalis v. Pennsylvania, 476 U.S. 140 (1986); Smith, 543 U.S. 462; and Arizona v. Rumsey, 467 U.S. 203 (1984)—trial judges ruled that the prosecution had failed to introduce sufficient evidence to prove one or more of the actual elements of the offenses in question. In none of these cases (and in none of our other double jeopardy cases) did a trial judge terminate a prosecution before verdict based on an element of the judge’s own creation.

The first two cases, Smalis and Smith, involved gar- den variety preverdict acquittals, i.e., rulings based on the ground that the prosecution had failed to introduce sufficient evidence to prove one or more of the actual elements of an offense. (Using conventional modern terminology, Rule 29(a) of the Federal Rules of Criminal Procedure explicitly labels such rulings “acquittal[s].”)

In Smalis, the judge, at the close of the prosecution’s case in chief, granted a demurrer with respect to certain charges on the ground that the evidence regarding those charges was “legally insufficient to support a conviction.” 476 U. S., at 141. The State Supreme Court held that this ruling was not an acquittal for double jeopardy purposes because it was based on a legal determination (i.e., that the evidence was not sufficient) rather than a factual finding, but we rejected that distinction. Id., at 143–144. See also Sanabria v. United States, 437 U.S. 54, 71–72 (1978).

Smith involved a similar situation. There, one of the elements of a firearms offense with which the defendant was charged required proof that the gun “had a barrel ‘less than 16 inches’ in length,” 543 U. S., at 464, and the trial judge dismissed this charge before verdict on the ground that the prosecution had not introduced sufficient evidence to establish this undisputed element, id., at 464–465. Before the remaining charges were submitted to the jury, however, the judge reversed this ruling and allowed the charge to go to the jury. Id., at 465. We held, how- ever, that the judge’s prior ruling constituted an acquittal and therefore barred the defendant’s conviction for this offense. Id., at 467–469. Thus, both Smalis and Smith involved rulings that were very different from the one at issue here. In both of those earlier cases, the trial judges held that the evidence was insufficient to prove undis- puted elements of the offenses in question. In neither case did the judge invent a new element.

The final case, Rumsey, differs from Smalis and Smith in only one particular. Like Smalis and Smith, Rumsey involved a ruling that the prosecution’s evidence was in- sufficient to prove an element, but in Rumsey the ruling was predicated on a misconstruction of an element. In that case, after the defendant was found guilty of first-degree murder, the “trial judge, with no jury, . . . con- ducted a separate sentencing hearing” at which he determined that no aggravating circumstances were present. 467 U. S., at 205. In particular, the judge found that the prosecution had not proved that the murder had been committed “ ‘as consideration for the receipt, or in expectation of the receipt, of anything of pecuniary value.’ ” Id., at 205–206 (quoting Ariz. Rev. Stat. Ann. §13–703(F)(5) (Supp. 1983–1984)). The judge reached this conclusion because, in his (incorrect) view, that aggravating circumstance was limited to contract killings. 467 U. S., at 205–206. Holding that the judge’s ruling constituted an acquittal on the merits of the question whether a death sentence was appropriate, we noted that the ruling rested on “a misconstruction of the statute defining the pecuniary gain aggravating circumstance.” Id., at 211. Accordingly, the ruling was based on a determination that there was insufficient evidence to prove a real element; it was not based on the judicial invention of an extra “element.” And for that reason, it does not support the nonsensical result that the Court reaches today.

The Court may feel compelled to reach that result because it thinks that it would be unworkable to draw a distinction between a preverdict termination based on the trial judge’s misconstruction of an element of an offense and a preverdict termination based on the judge’s perception that a statute contains an “element” that is actually nonexistent. This practical concern is overblown. There may be cases in which this determination presents problems, but surely there are many cases in which the de- termination is quite easy. The present case is a perfect example, for here there is no real dispute that the trial judge’s ruling was based on a nonexistent statutory “element.” As noted, defense counsel conceded on appeal that the judge had “wrongly added an extraneous element to the statute” under which his client was charged. 491 Mich., at 3, 810 N. W. 2d, at 536.

Another good example is provided by State v. Korsen, 138 Idaho 706, 69 P.3d 126 (2003), where a Magistrate erroneously concluded that the offense of criminal trespass under Idaho law requires a showing that the defendant did something to justify the property owner’s request for the defendant to leave the premises. Id., at 710, 716–717, 69 P. 3d, at 130, 136–137. There is no question that the Magistrate in Korsen “effectively created an additional statutory element” before concluding that the prosecution had presented insufficient evidence as to this purported “element.” See ibid. (holding that double jeopardy did not bar a retrial because the Magistrate’s “finding did not actually determine in [defendant’s] favor any of the essential elements of the crime of trespass”).

Cases in which it can be said that a trial judge did not simply misinterpret a real element of an offense but instead invented an entirely new and nonexistent “element” are cases in which the judge’s error is particularly egregious. Permitting retrial in these egregious cases is especially appropriate.

*  *  *

I would hold that double jeopardy protection is not triggered by a judge’s erroneous preverdict ruling that creates an “element” out of thin air and then holds that the element is not satisfied. I therefore respectfully dissent.