In People v LaValle (3 NY3d 88 [2004]), the death sentence was vacated because the trial judge gave a coercive deadlock instruction to the jury. But no such coercive instruction was given in this case. Here, defendant John Taylor’s death sentence is vacated on the basis of a non-coercive deadlock instruction. Fair-minded citizens might well be forgiven for wondering whether the Court of Appeals is simply unwilling ever to uphold a death sentence, no matter how the law is written (or may be rewritten), no matter how carefully the trial judge and the jury carry out their responsibilities. I respectfully dissent.
I.
The Court’s Decision in LaValle
The deadlock instruction is set forth in CPL 400.27 (10), which deals with summations and the charge to the jury in the separate sentencing proceeding that follows a defendant’s conviction of capital murder. This provision states in its entirety as follows:
“At the conclusion of all the evidence, the people and the defendant may present argument in summation for or against the sentence sought by the people. The people may deliver the first summation and the defendant may then deliver the last sum*161mation. Thereafter, the court shall deliver a charge to the jury on any matters appropriate in the circumstances. In its charge, the court must instruct the jury that with respect to each count of murder in the first degree the jury should consider whether or not a sentence of death should be imposed and whether or not a sentence of life imprisonment without parole should be imposed, and that the jury must be unanimous with respect to either sentence. The court must also instruct the jury that in the event the jury fails to reach unanimous agreement with respect to the sentence, the court will sentence the defendant to a term of imprisonment with a minimum term of between twenty and twenty-five years and a maximum term of life. Following the court’s charge, the jury shall retire to consider the sentence to be imposed. Unless inconsistent with the provisions of this section, the provisions of sections 310.10 [‘Jury deliberation; requirement of; where conducted’], 310.20 [‘Jury deliberation; use of exhibits and other material’] and 310.30 [‘Jury deliberation; request for information’] shall govern the deliberations of the jury” (emphasis added).
In LaValle, the jury found the defendant guilty of first-degree murder in the course of and in furtherance of first-degree rape (Penal Law § 125.27 [1] [a] [vii]). The trial judge delivered a short, unembellished instruction to the jury on the subject of potential deadlock, advising simply that “in the event that you fail to reach unanimous agreement [on death or life without parole], then I will sentence the defendant to life imprisonment with a minimum term of between 20 and 25 years for Murder, 1st Degree.” On appeal, we “address[ed] the constitutionality of [this] ‘deadlock instruction’ ” (3 NY3d at 116).
We started our analysis by surveying several empirical studies on juror behavior regarding capital sentencing, which concluded that “jurors tend to grossly underestimate how long capital murderers not sentenced to death usually stay in prison”; and that “the sooner jurors think a defendant will be released from prison, the more likely they are to vote for death and the more likely they are to see the defendant as dangerous” (id. at 117 [internal quotation marks and citations omitted]). As a result, we reasoned that the deadlock instruction was objectionable on two related grounds. First, the instruction suggested that the defendant might be paroled in as few as 20 years if the jurors *162proved unable to achieve unanimity, thus introducing into their deliberations a consideration that was not a statutory aggravator—the defendant’s future dangerousness. Second, “[b]y interjecting future dangerousness, the deadlock instruction [gave] rise to an unconstitutionally palpable risk that one or more jurors who [could not] bear the thought that a defendant [might] walk the streets again after serving 20 to 25 years [would] join jurors favoring death in order to avoid the deadlock sentence” (id. at 118). At the conclusion of this discussion, we stated that “[w]e hold today that the deadlock instruction required by CPL 400.27 (10) is unconstitutional under the State Constitution because of the unacceptable risk that it may result in a coercive, and thus arbitrary and unreliable, sentence” (id. at 120 [emphasis added]).
We buttressed this holding with discussions of the legislative debate on the deadlock instruction; federal precedent, most prominently the United States Supreme Court’s decision in Jones v United States (527 US 373 [1999]); and New York precedent on coerced verdicts. We stated that we regarded Jones “as unfaithful to the often repeated principle that death is qualitatively different and thus subject to a heightened standard of reliability” (id. at 127), which our State Constitution mandated. Accordingly, we concluded that it was
“necessarily our responsibility to strike down the deadlock instruction in CPL 400.27 (10) because it creates the substantial risk of coercing jurors into sentencing a defendant to death in violation of our Due Process Clause. The deadlock instruction is invalid under our own case law condemning coercive instructions, and the State Constitution’s Due Process Clause, providing greater protection than its federal counterpart. Consequently, defendant’s death sentence must be set aside” (id. at 128 [emphasis added]).
Next, we “conclude[d] that the absence of any instruction [was] no better than the current instruction under our constitutional analysis,” and so again “decline[d] to adopt Jones” (id.). We reasoned that without an instruction as to the consequences of a deadlock, jurors might speculate, “as the Legislature apparently feared when it decided to prescribe the [deadlock] instruction” (id.). We again pointed to empirical studies to support the view that “jurors might fear that the failure to reach a unanimous verdict would lead to a defendant’s release, retrial or *163sentence to an even lesser term than the one currently prescribed in the deadlock scenario [a minimum of 20 to 25 years to life]”; and that “[findeed, a key motivation for jurors to vote for the death penalty is undoubtedly their fear that a defendant will otherwise pose a danger on the streets” (id.).
In this section of the opinion, we articulated our holding in two related ways. First, we stated that “[w]e hold that in this case the Due Process Clause of the New York Constitution requires a higher standard of fairness than the Federal Constitution as interpreted by the Jones majority” (id. at 129).1 Second, “we [held] that providing no deadlock instruction in the course of capital sentencing violates our Due Process Clause” (id. at 130). As further support for these holdings, we called attention to “the clear legislative intent that there be a jury instruction on the consequences of a deadlock,” and court rules and legislative enactments to this effect in eight states (id.).
At the end of the discussion of the deadlock instruction, we determined that “[w]e cannot. . . ourselves craft a new instruction, because to do so would usurp legislative prerogative” (id. at 131). We then made the following pronouncements, which lie at the heart of our quarrel with the majority on this appeal: “We thus conclude that under the present statute, the death penalty may not be imposed. Cases in which death notices have been filed may go forward as noncapital first degree murder prosecutions” (id.).
In sum, LdValle held that the deadlock instruction delivered by the trial judge in that case violated the Due Process Clause of our State Constitution by creating a substantial risk that a juror favoring life without parole would be coerced into voting for the death sentence so as to prevent the defendant’s eventual release on parole. In addition, LdValle held that the Due Process Clause of our State Constitution mandates that our death penalty statute include a deadlock instruction, which only the Legislature may devise.
I continue to believe that LdValle was wrongly decided for the reasons articulated by Judge R.S. Smith in his dissent, which I joined. Nonetheless, I accept the foregoing holdings as binding *164precedent for reasons of stare decisis. The majority has now, however, chosen to convert the spare closing comment in LaValle that “under the present statute, the death penalty may not be imposed” into our “holding” in that case that “the death penalty sentencing statute is unconstitutional on its face” (plurality op at 155).
There was no discussion in LaValle of the deadlock instruction’s facial constitutionality; the opinion is devoid of any legal reasoning to support a holding that the deadlock instruction in CPL 400.27 (10) is facially unconstitutional. While “both the legitimacy and the ability of the judiciary to function dictate that legal issues that have been addressed by a jurisdiction should not be revisited every time they arise” (plurality op at 148), this maxim presupposes that the legal issue in question has, in fact, been analyzed and decided by a court. To the extent that any judicial utterance in LaValle may be read as purporting to hold the deadlock instruction facially unconstitutional, it does not meet this test and is dictum. (“A judge’s power to bind is limited to the issue that is before him; he cannot transmute dictum into decision by waving a wand and uttering the word ‘hold’ ” [United States v Rubin, 609 F2d 51, 69 n 2 (2d Cir 1979, Friendly, J., concurring)].)
In the New York University School of Law’s annual James Madison Lecture in 2005, Judge Pierre N. Leval of the United States Court of Appeals for the Second Circuit discussed the hazards inherent in the failure of courts to distinguish between dictum and holding (see Leval, Madison Lecture: Judging Under the Constitution: Dicta about Dicta, 81 NYU L Rev 1249 [2006]). Two of his observations are particularly telling in the context of this appeal.
First, Judge Leval set out a handy test to separate dictum from holding:
“To identify dictum, it is useful to turn the questioned proposition around to assert its opposite, or to assert whatever alternative proposition the court rejected in its favor. If the insertion of the rejected proposition into the court’s reasoning, in place of the one adopted, would not require a change in either the court’s judgment or the reasoning that supports it, then the proposition is dictum. It is superfluous. It had no functional role in compelling the judgment” (id. at 1257).
*165Here, the “questioned proposition” is that the deadlock instruction is facially unconstitutional, and “its opposite” is that the deadlock instruction is, in fact, facially constitutional. If we applied this analysis to LaValle, our judgment in the defendant’s favor would not have changed because the deadlock instruction delivered by the trial judge created a substantial risk of a coerced verdict of death. Any assertion in LaValle about whether the deadlock instruction was facially unconstitutional was “superfluous” to this reasoning and “had no functional role in compelling” the judgment.
Second, Judge Leval observed that “[h]owever grievous the errors a court commits when it writes dictum disguised as holding, those errors would be neutralized if the next court would recognize the prior dictum as nonbinding and go on to grapple with and decide the issue” (id. at 1268-1269 [emphasis added]). The way the majority now treats the LaValle dictum as precedent is a cautionary tale in this regard. Wrapping itself in a false mantle of stare decisis, the majority ignores the legal issues presented by this case. If ever so politely, the majority even chides the People for having the temerity to prosecute an appeal in a case where “we are ultimately left exactly where we were three years ago” (plurality op at 155).2
Rather than shrinking defensively from the least suggestion of error or inadvertence, the plurality should have taken the path of self-correction advised by Judge Leval. The plurality should have squarely confronted and decided the legal issues raised by this appeal: whether the deadlock instruction was constitutional as applied to the defendant in this case, and whether, even if it was, his conviction must nonetheless be vacated because the deadlock instruction in CPL 400.27 (10) is facially unconstitutional. I now turn to those issues.
II.
The Deadlock Instruction in This Case
In keeping with CPL 400.27 (10), the trial judge informed the jury that in the event of deadlock he would sentence defendant *166himself; that the law required him to “sentence [defendant] to life imprisonment,” but also to “fix a point at which the defendant [would] become eligible for parole”; and that the law further required him to “fix that point between twenty and twenty-five years for each count.” He then provided further clarification to the jury, advising that “[i]n other words, on each count I would sentence the defendant to life imprisonment and order that he not become eligible for parole until he had served the minimum term that I fix, a term of between twenty and twenty-five years for each count.”
The judge next placed what he had just said in the context of the case before the jury. He informed the jurors that he thought it “fair to tell” them
“that the six [count] s of first degree murder, and the two counts of first degree attempted murder on which you have convicted the defendant, are precisely the type of crimes that almost always induce a judge to give the maximum sentence permissible.
“In this case I would have the authority to sentence the defendant, not only to the maximum on each count, but also to make those sentences run consecutively. So, the maximum sentence I could give and would almost certainly impose in this case, would be a sentence of 175 years to life, which means that the defendant would become eligible for parole, but only after he had served 175 years in jail.”
No rational juror listening to this charge could have harbored any reasonable fear that a deadlock might lead to defendant’s eventual release back into the community. Thus, the deadlock instruction delivered in this case simply did not pose the risk of a coerced and unreliable verdict of death that caused us to vacate the death sentence in LaValle.
Moreover, the trial judge was clearly empowered to give the charge that he gave. Section 400.27 (10) mandates that “the court shall deliver a charge to the jury on any matters appropriate in the circumstances” (emphasis added). Here, the trial judge thought it “fair” to communicate truthful and accurate information to the jury as to the “almost certain[ ]” practical consequences of a deadlock in the circumstances of this case. There was nothing wrong with this; in fact, it is exactly what the trial judge should have done to protect the constitutional rights of a defendant on trial for his life (see Gregg v Georgia, 428 US 153, *167192-193 [1976]; see also California v Ramos, 463 US 992, 1009 [1983]).
Defendant contends, however, that the charge amounted to “judicial rewriting” of the deadlock instruction in contravention of CPL 400.27 (10) and our decision in LaValle. This is incorrect. The trial judge gave the deadlock instruction required by the statute, and then explained its implications to the jury. Nothing in section 400.27 (10) forbids this, and, in this preLaValle case, it is commendable that the judge took the extra step to extinguish any possibility of a coercive instruction. If the Legislature had wanted to prevent a trial judge from expanding on the deadlock instruction, it surely could have and would have done so {compare CPL 400.27 [10], with CPL 300.10 [3] [setting out within quotation marks the exact words that a trial judge must charge “without elaboration” where a defendant has raised the affirmative defense of lack of criminal responsibility by reason of mental disease or defect]). As for our decision in LaValle, we merely stated there that the Court of Appeals could not craft a new generic deadlock instruction to replace the existing one in CPL 400.27 (10). It perhaps bears repeating that the trial judge here did, in fact, deliver the statutorily required deadlock instruction. I therefore conclude that the deadlock instruction given in this case was not unconstitutionally coercive under the analysis in LaValle. The deadlock instruction was constitutionally applied to defendant.
III.
Facial Constitutionality of the Deadlock Instruction
Our precedent is well-established: “A party mounting a facial constitutional challenge bears the substantial burden of demonstrating that in any degree and in every conceivable application, the law suffers wholesale constitutional impairment. In other words, the challenger must establish that no set of circumstances exists under which the Act would be valid” (Matter of Moran Towing Corp. v Urbach, 99 NY2d 443, 448 [2003, Ciparick, J.] [internal quotation marks and citations omitted], quoting Cohen v State of New York, 94 NY2d 1, 8 [1999] and United States v Salerno, 481 US 739, 745 [1987]; see also People v Stuart, 100 NY2d 412, 421 [2003] [“A successful facial challenge means that the law is invalid in toto—and therefore incapable of any valid application” (internal quotation marks and citations omitted)]). The general approach taken by the courts is to “limit *168the solution to the problem” and “enjoin only the unconstitutional applications of a statute while leaving other applications in force, or to sever its problematic portions while leaving the remainder intact” (Ayotte v Planned Parenthood of N. New England, 546 US 320, 328, 329 [2006] [citations omitted]).3
We decided in LaValle that the deadlock instruction is not severable from the other statutory provisions authorizing the death penalty. The question therefore becomes whether the constitutional and unconstitutional applications of the deadlock instruction are severable. If they are not, defendant’s death sentence must be vacated even though the deadlock instruction was constitutionally applied to him. Defendant, while conceding that in certain circumstances the statute may be constitutionally applied, offers three reasons why the Court should take this extraordinary step, which is highly disfavored by long-standing state and federal precedent: (1) the remaining constitutional applications are too few and create a “freakish” regime contrary to legislative intent; (2) relatedly, the death penalty could not be constitutionally applied to the elderly or the terminally ill; and (3) there are administrative problems.
(1) The Remaining Applications
Both the federal and state constitutions require that the states “genuinely narrow the class of persons eligible for the death penalty” and “reasonably justify the imposition of a more severe sentence” on those made death-eligible “compared to others found guilty of murder” (Zant v Stephens, 462 US 862, 877 [1983]; see also People v Harris, 98 NY2d 452, 476-477 [2002]; Matter of Hynes v Tomei, 92 NY2d 613, 628 [1998]). Accordingly, the Legislature in Penal Law § 125.27 limited death-eligible, first-degree murder to those who kill intentionally, and whose conduct includes at least one of 13 separately listed aggravating factors (Penal Law § 125.27 [1] [a] [i]-[xiii]).
As a result of our decision in LaValle, however, only a subcategory of death-eligible defendants—those who will never be released from prison even if the jury deadlocks on the sentence of death or life without parole—would be at risk of the *169death penalty. This is because only these first-degree murderers would be immune to the coercive effect of the deadlock instruction. This subcategory, however, embraces the “worst of the worst”; specifically, first-degree murderers who have killed multiple victims and/or have committed other separate and distinct crimes such that they are eligible for consecutive sentences so lengthy in total as to be the functional equivalent of life without parole (the case here); and first-degree murderers whose separate and distinct additional crimes are so severe that the Legislature has prescribed an automatic sentence of life without parole.
For example, a person serving a life sentence who kills a correction officer will have no possibility of parole, regardless of whether the jury deadlocks, for two reasons: he is already serving life without parole on the prior conviction, and the killing of a correction officer constitutes not only first-degree murder but also “aggravated murder” (Penal Law § 125.26 [1] [a] [iii]), which automatically results in a sentence of life without parole (Penal Law § 70.00 [3] [a] [i]; §§ 60.06, 70.00 [5]).4 The killing of a police officer or a peace officer is also aggravated murder (Penal Law § 125.26 [1] [a] [i], [ii]). Additionally, the Legislature has mandated a sentence of life without parole for someone, 18 years of age or older, who intentionally murders a child under the age of 14 during the course of certain sex crimes (see L 2004, ch 459 [“Joan’s Law”]; see also Penal Law § 125.25 [5]; §§ 60.06, 70.00 [5]); and the crimes of terrorism where the underlying offense is a class A-I felony, or when there is possession of a chemical or biological weapon in the first degree, or use of a chemical or biological weapon in the. first degree (see L 2004, ch 1 [creating the State Office of Homeland Security and enacting various anti-terrorism measures]; see also Penal Law § 490.25 [1], [2] [c], [d]; §§ 490.45, 490.55, 60.06, 70.00 [5]).
In the case of a mandatory life-without-parole crime, the trial judge would inform the jury that he was required to sentence the defendant to life without parole notwithstanding any deadlock on a count of first-degree murder. We specifically stated in LaValle that “[i]f the deadlock sentence had been life without parole, then jurors would have no reason to fear that a deadlock would result in the eventual release of the defendant. In that *170instance jurors committed to life without parole would not be coerced into giving up their conscientious belief in order to reach a verdict” (3 NY3d at 126 n 19).
Further, the Legislature first amended sections 60.06 and 70.00 (5) of the Penal Law in relation to sentencing for mandatory life-without-parole crimes in July 2004, about a month after our decision in LaValle. The Legislature included within these provisions the admonition that “nothing in this section” (Penal Law § 60.06) or “subdivision” (Penal Law § 70.00 [5]) respectively “shall preclude or prevent a sentence of death when the defendant is also convicted of murder in the first degree as defined in section 125.27 of this chapter.” Sections 60.06 and 70.00 (5) have been amended twice since to add new crimes.
Defendant argues that the deadlock instruction is nonetheless facially unconstitutional because its constitutional applications post -LaValle are too few. This is, of course, a curious position to take since the federal and state constitutions affirmatively require limiting death eligibility. As a result, the key consideration is not how many constitutional applications of the death penalty remain after our decision in LaValle, but whether those remaining applications are rational. A death penalty applicable to first-degree murderers who kill multiple victims and/or commit multiple crimes; or whose crimes include killing a police officer, peace officer or correction officer; or killing a child during a sex crime; or killing in the course of terrorist-related activities surely meets the test of rationality. Certainly, if the Legislature had purposely confined the death penalty to these applications when it adopted the statute in 1995, its decision would not have been amenable to judicial second-guessing on the basis that too few crimes or murderers were covered (see Harris, 98 NY2d at 476-477).
Defendant also contends, though, that the deadlock instruction is facially unconstitutional precisely because it allows for far fewer constitutional applications than originally envisioned by the Legislature. Defendant surmises that the Legislature would have preferred no death penalty at all to the “freakish” regime left after LaValle. Similarly, the concurrence opines that “any attempt to save a remnant of the death penalty statute through an exercise in ‘application severability’ would be a mistake” (concurring op at 157).
The death penalty statute was adopted by the Legislature in 1995 after almost two decades of public debate and political re*171action (see Sack, Political Budget Gap is Widening, New York Times, Mar. 9, 1995; Dao, New York Leaders Offer Limited Bill on Death Penalty, New York Times, Mar. 4, 1995). The death penalty was then, as now, controversial. Every aspect of this new legislation was clearly going to be litigated, with unpredictable results. The deadlock instruction itself is a perfect example. As the legislative debate set out in LaValle illustrates, at least one apparent motivation for including an anticipatory instruction on the consequences of deadlock was to avoid a potential “constitutional problem” (LaValle, 3 NY3d at 121). The United States Supreme Court did not decide Jones until four years later, in 1999; we did not decide LaValle until 2004. And, as previously discussed, the judicial outcomes differed.
In light of its past difficulties in enacting a death penalty and the uncertainties going forward, the Legislature included a severability provision in the statute (see L 1995, ch 1, § 37). By so doing, the Legislature told us in the clearest way possible that it preferred a judicially redesigned or “rewritten” statute with fewer applications to a nonexistent one. Allowing the death penalty to be applied to a core group of defendants charged with the worst crimes does not undermine legislative intent; it preserves the legislative will by restricting the death penalty’s application in a rational and constitutional manner.
The concurrence protests, however, that “a sensible Legislature” anticipating the LaValle decision would not have enacted a statute this limited, which does not “bear[ ] a reasonable resemblance to the statute the Legislature did enact” (concurring op at 158). This amounts to pure judicial guesswork, as it finds no support whatsoever in the statute’s text, structure, purpose or history—and the concurrence does not bother to claim otherwise. Professions of deference to the Legislature and judicial modesty ring hollow if a reviewing court is unwilling to uphold as much of a statute as possible, especially where the conventional sources for statutory interpretation (here, text and legislative history) clearly signal in this direction.5 Of course, the Legislature may always amend the death penalty statute in *172any number of ways, as the concurrence observes. The whole point of severability doctrine, however, is to salvage as much of the Legislature’s handiwork as possible, and to free the Legislature from the burden of reenacting that which has already once successfully run the lawmaking gauntlet. There can be no doubt that the Legislature intended to make the death penalty enforceable against someone like defendant, who has been convicted of multiple counts of first-degree murder.
(2) The Elderly and the Terminally 111
Defendant hypothesizes a 70-year-old or terminally ill defendant who is convicted of only one count of capital murder. A minimum sentence of 20 to 25 years to life would exceed this defendant’s life expectancy, but state and federal proscriptions against cruel and unusual punishment would bar imposition of the death penalty. Accordingly, defendant argues that the deadlock instruction, even if not unconstitutionally coercive in some applications, is still facially unconstitutional because it violates the Eighth Amendment.
Of course, senior citizens do not swell the ranks of first-degree murderers in New York. According to defendant, out of 560 first-degree murder prosecutions in New York from September 5, 1995 through November 14, 2005, only one involved a 70 year old. There are no comparable figures for the terminally ill, but it is probably safe to surmise that few of these 560 prosecutions involved defendants known to be near death. Assuming a prosecutor in the future is foolhardy enough to seek the death penalty against an elderly or terminally ill defendant on one count of capital murder, and assuming that a jury imposes the death sentence, after hearing mitigating evidence that would presumably dwell on the defendant’s age and health, the death *173sentence would never survive Eighth Amendment review, or examination for proportionality under CPL 470.30 (3) (b) (Court of Appeals mandated to determine “whether the sentence of death is excessive or disproportionate to the penalty imposed in similar cases considering both the crime and the defendant”). In short, the notion that the deadlock instruction infringes the constitutional rights of the elderly and the terminally ill is beyond farfetched; it is a chimera. Even in First Amendment overbreadth review, where the party mounting a facial challenge need not demonstrate “wholesale constitutional impairment,” the courts look at whether the hypothesized constitutional infirmity is something more than insubstantial (see Barton, 8 NY3d at 75-76 [“The test for determining overbreadth is whether the law on its face prohibits a real and substantial amount of constitutionally protected conduct. (T)he mere fact that one can conceive of some impermissible applications of a statute is not sufficient to render it susceptible to an overbreadth challenge” (internal quotation marks and citations omitted)]; see also People v Broadie, 37 NY2d 100, 119 [1975] [upholding Rockefeller Drug Laws, which were challenged on Eighth Amendment grounds, as facially constitutional although “in some rare case on its particular facts” it may be found “that the statutes have been unconstitutionally applied”]).
(3) Administrative Problems
Defendant also catalogs a “host of difficult questions of administration” that would supposedly cripple enforcement of the death penalty in those cases where the deadlock instruction could be constitutionally applied. These range from the fanciful (the prospect of genetic testing to assess whether a defendant suffered, or was at risk, for a terminal illness) to mundane questions of timing. In its sole comment on the merits in this appeal, the plurality echoes this complaint, worrying that the trial court would be compelled to act without the benefit of a presentencing report in violation of CPL 380.30 and 390.20; and that the “new framework” would be “repugnant to the Legislature’s intent as the court would have to conduct some sort of actuarial analysis and health assessment to determine defendant’s expected longevity as well as a thorough examination of the existence of convictions that may require consecutive sentencing” (plurality op at 154-155).
As the People point out, these concerns do not pose insurmountable practical problems; they are largely matters of scheduling. For example, a presentencing report may be ordered *174immediately after the jury hands down a guilty verdict and before the penalty-phase trial even begins; a trial judge should thoroughly examine the existence of convictions that may require consecutive sentencing in every case (see Penal Law § 70.25).
Further, a defendant-specific “actuarial analysis” is not required—the trial court did not need to conduct one in this case. The issue boils down to whether a defendant is subject to consecutive sentences so lengthy in relation to life expectancy in the United States that no rational juror could harbor a reasonable belief that the defendant would ever be released from prison even if the jury deadlocked on the sentence of death or life without parole;6 or an actual sentence of life without parole is otherwise mandated.
The former standard was easily met in three of the six capital appeals considered by the Court, where the defendant was eligible for consecutive sentencing totaling in excess of 100 years.7 It eliminates the theoretical risk that an aged or frail defendant could be executed while “a 20 year old who committed the exact same crimes could not be” (concurring op at 159). The concurrence evidently does not consider such an approach to be a satisfactory or permissible limiting principle for constraining the statute. I disagree for reasons already explained. But even if I were to agree, the concurrence nowhere explains why the deadlock instruction is facially unconstitutional even though it *175may be constitutionally applied to a first-degree murderer who also stands convicted of a mandatory life-without-parole crime. Neither does the plurality. This aspect of today’s decision contradicts and frustrates the Legislature’s manifest intention when it amended sections 60.06 and 70.00 (5) of the Penal Law post -LaValle.
W.
Conclusion
The crime in this case was horrific: apparently to eliminate the witnesses to a robbery, seven human beings were shot in the head at point-blank range; five of them died. After a textbook trial, defendant was convicted of six counts of capital murder and sentenced to death by the jury. Yet, a majority of the Court vacates the sentence of death. Why? Not because the deadlock instruction delivered by the trial judge was coercive—it was not, as the concurrence concedes and the plurality does not dispute. Instead, John Taylor’s death sentence is vacated because the deadlock instruction delivered in the capital trial of Stephen LaValle was coercive.
It is, to say the least, highly unusual for us to declare a statute facially unconstitutional so that its constitutional applications fall by the wayside along with its unconstitutional ones. It is unheard of for us to do such a thing without explanation. The plurality points to our “holding” in LaValle that “the death penalty sentencing statute is unconstitutional on its face” (plurality op at 155). The careful reader, however, will not discover those words among the LaValle opinion’s several self-styled “holdings”; the careful reader will not encounter any discussion of facial constitutionality or explanation why the deadlock instruction’s constitutional and unconstitutional applications are not severable. Certainly, “death is qualitatively different and thus subject to a heightened standard of reliability” (LaValle, 3 NY3d at 127). But this does not mean that “because of the unique severity and finality of a sentence of death, a defendant must be given every possible opportunity to escape from it” {id. at 148-149 [R.S. Smith, J., dissenting]). It does not release an appellate court from the fundamental obligation to furnish the parties and the public with a reasoned explanation for its decision. I respectfully dissent.
*176Chief Judge Kaye and Judge Jones concur with Judge Ciparick; Judge Smith concurs in result in a separate opinion; Judge Read dissents in another opinion in which Judges Graffeo and Pigott concur.
Judgment modified, etc.
. As the LdValle dissent noted, although Jones was a 5-4 decision, not one of the nine Justices endorsed the idea that the Federal Constitution mandates an anticipatory instruction in a capital case on the consequences of deadlock (see 3 NY3d at 146-147).
. Unaccountably, the plurality also spills a great deal of ink defending our decision in LaValle that the deadlock instruction is not severable from an attack that the People do not mount (plurality op at 153-154). The People took the position on this appeal that, while they agreed on the merits with the LaValle dissent, they “operate[d] under the LaValle Court’s reasoning, accepting ... as stare decisis” its rulings that the deadlock instruction was unconstitutionally coercive as applied in that case, and was not severable from the remainder of the statute.
. There is a recognized exception to the general rule where the facial challenge is based on constitutional free-speech grounds. The “overbreadth” doctrine may render a statute invalid in all of its applications (i.e., facially invalid) if invalid in any of them so as not to forestall or “chill” constitutionally protected expression (see Broadrick v Oklahoma, 413 US 601, 610-613 [1973]; see also People v Barton, 8 NY3d 70, 75-76 [2006] [discussing test for First Amendment overbreadth review]). This is not an overbreadth case.
. The “Crimes Against Police Act” established the new crime of aggravated murder (see L 2005, ch 765; see also Senate Mem in Support, 2005 McKinney’s Session Laws of NY, at 2575).
. Our decision in Association of Surrogates & Supreme Ct. Reporters Within City of N.Y. v State of New York (79 NY2d 39 [1992]) does not support the proposition that application severability is unavailable in the circumstances presented in this case (plurality op at 155; concurring op at 157). In Surrogates, the Legislature adopted a statute intended to offset anticipated state budgetary shortfalls for fiscal year 1991-1992 by imposing a five-day lag payroll upon both represented and unrepresented employees of the Unified Court System. We concluded that this legislation unconstitutionally impaired the *172represented employees’ employment contracts. We noted that application severance would preserve the statute’s purpose—which was to generate a specified level of budgetary savings—“[i]n a general way” because unrepresented employees, who made up 10% of the employees covered by the statute, would remain subject to it (79 NY2d at 48). We decided, however, that application severability was not available for two reasons. First, we stated that “[significantly” there was no severability clause in the statute (id.). Then we turned to legislative history, noting that the Legislature had “enacted lag payrolls on two prior occasions in the recent past and in neither of them did it apply the legislation to only a limited segment of the employee pool, or provide for severability if the statute was invalid as to some of those affected” (id.). In sum, in Surrogates we looked first to see whether there was any suggestion in the text that the Legislature preferred severability; finding none, we examined other indicia of legislative intent. The difference between this case and Surrogates is that, applying the same rules, the facts point in opposite directions.
. Life expectancy tables, sourced from the National Center for Health Statistics, are so commonly used in the courts that they are included in the Pattern Jury Instructions (see 1B NY PJI3d Appendix A [2007]). The most recent edition of the PJI includes life tables for the United States based on age-specific death rates in 1997, which were published in 1999. These tables project life expectancy for males born in 1997 at 73.6 years, and life expectancy for females born in 1997 at 79.4 years (see id. at 1632, 1635). A defendant would necessarily be at least 18 years old and born earlier than whatever birth year is specified in the life expectancy tables in the PJI, because they are updated periodically. Accordingly, the deadlock charge is unquestionably (and comfortably) noncoercive whenever a defendant, regardless of youth or age or health status, is eligible for consecutive sentences exceeding these widely accepted life expectancy numbers for males and females.
. The three are the defendants in this case, People v Mateo (2 NY3d 383 [2004]) and Harris. In addition, because the defendant in People v Shulman (6 NY3d 1 [2005]) was a serial killer, the People might have charged his crimes differently to run up the potential consecutive sentencing had they known this to be necessary to insure a noncoercive deadlock instruction. The defendants in LaValle and People v Cahill (2 NY3d 14 [2003]) were single-victim first-degree murderers. A noncoercive deadlock charge was not possible in these two cases in light of the LaValle analysis.