Chief Justice Rehnquist

delivered the opinion of the Court.

The Bail Reform Act of 1984 (Act) allows a federal court to detain an arrestee pending trial if the Government demonstrates by clear and convincing evidence after an adversary hearing that no release conditions “will reasonably assure . . . the safety of any other person and the community.” The United States Court of Appeals for the Second Circuit struck down this provision of the Act as facially unconstitutional, because, in that court’s words, this type of pretrial detention violates “substantive due process.” We granted certiorari because of a conflict among the Courts of Appeals regarding the validity of the Act.1 479 U. S. 929 (1986). We hold that, as against the facial attack mounted by these respondents, the Act fully comports with constitutional requirements. We therefore reverse.

*742r — 4

Responding to “the alarming problem of crimes committed by persons on release,” S. Rep. No. 98-225, p. 3 (1983), Congress formulated the Bail Reform Act of 1984, 18 U. S. C. §3141 et seq. (1982 ed., Supp. III), as the solution to a bail crisis in the federal courts. The Act represents the National Legislature’s considered response to numerous perceived deficiencies in the federal bail process. By providing for sweeping changes in both the way federal courts consider bail applications and the circumstances under which bail is granted, Congress hoped to “give the courts adequate authority to make release decisions that give appropriate recognition to the danger a person may pose to others if released.” S. Rep. No. 98-225, at 3.

To this end, § 3141(a) of the Act requires a judicial officer to determine whether an arrestee shall be detained. Section 3142(e) provides that “[i]f, after a hearing pursuant to the provisions of subsection (f), the judicial officer finds that no condition or combination of conditions will reasonably assure the appearance of the person as required and the safety of any other person and the community, he shall order the detention of the person prior to trial.” Section 3142(f) provides the arrestee with a number of procedural safeguards. He may request the presence of counsel at the detention hearing, he may testify and present witnesses in his behalf, as well as proffer evidence, and he may cross-examine other witnesses appearing at the hearing. If the judicial officer finds that no conditions of pretrial release can reasonably assure the safety of other persons and the community, he must state his findings of fact in writing, § 3142(i), and support his conclusion with “clear and convincing evidence,” § 3142(f).

The judicial officer is not given unbridled discretion in making the detention determination. Congress has specified the considerations relevant to that decision. These factors include the nature and seriousness of the charges, the substan-tiality of the Government’s evidence against the arrestee, the *743arrestee’s background and characteristics, and the nature and seriousness of the danger posed by the suspect’s release. § 3142(g). Should a judicial officer order detention, the detainee is entitled to expedited appellate review of the detention order. §§ 3145(b), (c).

Respondents Anthony Salerno and Vincent Cafaro were arrested on March 21,1986, after being charged in a 29-count indictment alleging various Racketeer Influenced and Corrupt Organizations Act (RICO) violations, mail and wire fraud offenses, extortion, and various criminal gambling violations. The RICO counts alleged 35 acts of racketeering activity, including fraud, extortion, gambling, and conspiracy to commit murder. At respondents’ arraignment, the Government moved to have Salerno and Cafaro detained pursuant to § 3142(e), on the ground that no condition of release would assure the safety of the community or any person. The District Court held a hearing at which the Government made a detailed proffer of evidence. The Government’s case showed that Salerno was the “boss” of the Genovese crime family of La Cosa Nostra and that Cafaro was a “captain” in the Genovese family. According to the Government’s proffer, based in large part on conversations intercepted by a court-ordered wiretap, the two respondents had participated in wide-ranging conspiracies to aid their illegitimate enterprises through violent means. The Government also offered the testimony of two of its trial witnesses, who would assert that Salerno personally participated in two murder conspiracies. Salerno opposed the motion for detention, challenging the credibility of the Government’s witnesses. He offered the testimony of several character witnesses as well as a letter from his doctor stating that he was suffering from a serious medical condition. Cafaro presented no evidence at the hearing, but instead characterized the wiretap conversations as merely “tough talk.”

The District Court granted the Government’s detention motion, concluding that the Government had established by *744clear and convincing evidence that no condition or combination of conditions of release would ensure the safety of the community or any person:

“The activities of a criminal organization such as the Genovese Family do not cease with the arrest of its principals and their release on even the most stringent of bail conditions. The illegal businesses, in place for many years, require constant attention and protection, or they will fail. Under these circumstances, this court recognizes a strong incentive on the part of its leadership to continue business as usual. When business as usual involves threats, beatings, and murder, the present danger such people pose in the community is self-evident.” 631 F. Supp. 1364, 1375 (SDNY 1986).2

Respondents appealed, contending that to the extent that the Bail Reform Act permits pretrial detention on the ground that the arrestee is likely to commit future crimes, it is unconstitutional on its face. Over a dissent, the United States Court of Appeals for the Second Circuit agreed. 794 F. 2d 64 (1986). Although the court agreed that pretrial detention could be imposed if the defendants were likely to intimidate witnesses or otherwise jeopardize the trial process, it found “§ 3142(e)’s authorization of pretrial detention [on the ground of future dangerousness] repugnant to the concept of substantive due process, which we believe prohibits the total deprivation of liberty simply as a means of preventing future crimes.” Id., at 71-72. The court concluded that the Government could not, consistent with due process, detain persons who had not been accused of any crime merely because they were thought to present a danger to the community. Id., at 72, quoting United States v. Melendez-Carrion, 790 F. *7452d 984, 1000-1001 (CA2 1986) (opinion of Newman, J.). It reasoned that our criminal law system holds persons accountable for past actions, not anticipated future actions. Although a court could detain an arrestee who threatened to flee before trial, such detention would be permissible because it would serve the basic objective of a criminal system— bringing the accused to trial. The court distinguished our decision in Gerstein v. Pugh, 420 U. S. 103 (1975), in which we upheld police detention pursuant to arrest. The court construed Gerstein as limiting such detention to the “ ‘administrative steps incident to arrest.’ ” 794 F. 2d, at 74, quoting Gerstein, supra, at 114. The Court of Appeals also found our decision in Schall v. Martin, 467 U. S. 253 (1984), upholding postarrest, pretrial detention of juveniles, inapposite because juveniles have a lesser interest in liberty than do adults. The dissenting judge concluded that on its face, the Bail Reform Act adequately balanced the Federal Government’s compelling interests in public safety against the detainee’s liberty interests.

II

A facial challenge to a legislative Act is, of course, the most difficult challenge to mount successfully, since the challenger must establish that no set of circumstances exists under which the Act would be valid. The fact that the Bail Reform Act might operate unconstitutionally under some conceivable set of circumstances is insufficient to render it wholly invalid, since we have not recognized an “overbreadth” doctrine outside the limited context of the First Amendment. Schall v. Martin, supra, at 269, n. 18. We think respondents have failed to shoulder their heavy burden to demonstrate that the Act is “facially” unconstitutional.3

*746Respondents present two grounds for invalidating the Bail Reform Act’s provisions permitting pretrial detention on the basis of future dangerousness. First, they rely upon the Court of Appeals’ conclusion that the Act exceeds the limitations placed upon the Federal Government by the Due Process Clause of the Fifth Amendment. Second, they contend that the Act contravenes the Eighth Amendment’s proscription, against excessive bail. We treat these contentions in turn.

A

The Due Process Clause of the Fifth Amendment provides that “No person shall... be deprived of life, liberty, or property, without due process of law . . . .” This Court has held that the Due Process Clause protects individuals against two types of government action. So-called “substantive due process” prevents the government from engaging in conduct that “shocks the conscience,” Rochin v. California, 342 U. S. 165, 172 (1952), or interferes with rights “implicit in the concept of ordered liberty,” Palko v. Connecticut, 302 U. S. 319, 325-326 (1937). When government action depriving a person of life, liberty, or property survives substantive due process scrutiny, it must still be implemented in a fair manner. Mathews v. Eldridge, 424 U. S. 319, 335 (1976). This requirement has traditionally been referred to as “procedural” due process.

Respondents first argue that the Act violates substantive due process because the pretrial detention it authorizes constitutes impermissible punishment before trial. See Bell v. Wolfish, 441 U. S. 520, 535, and n. 16 (1979). The Government, however, has never argued that pretrial detention could be upheld if it were “punishment.” The Court of Appeals assumed that pretrial detention under the Bail Reform Act is regulatory, not penal, and we agree that it is.

As an initial matter, the mere fact that a person is detained does not inexorably lead to the conclusion that the government has imposed punishment. Bell v. Wolfish, supra, at *747537. To determine whether a restriction on liberty constitutes impermissible punishment or permissible regulation, we first look to legislative intent. Schall v. Martin, 467 U. S., at 269. Unless Congress expressly intended to impose punitive restrictions, the punitive/regulatory distinction turns on “‘whether an alternative purpose to which [the restriction] may rationally be connected is assignable for it, and whether it appears excessive in relation to the alternative purpose assigned [to it].’” Ibid., quoting Kennedy v. Mendoza-Martinez, 372 U. S. 144, 168-169 (1963).

We conclude that the detention imposed by the Act falls on the regulatory side of the dichotomy. The legislative history of the Bail Reform Act clearly indicates that Congress did not formulate the pretrial detention provisions as punishment for dangerous individuals. See S. Rep. No. 98-225, at 8. Congress instead perceived pretrial detention as a potential solution to a pressing societal problem. Id., at 4-7. There is no doubt that preventing danger to the community is a legitimate regulatory goal. Schall v. Martin, supra.

Nor are the incidents of pretrial detention excessive in relation to the regulatory goal Congress sought to achieve. The Bail Reform Act carefully limits the circumstances under which detention may be sought to the most serious of crimes. See 18 U. S. C. § 3142(f) (detention hearings available if case involves crimes of violence, offenses for which the sentence is life imprisonment or death, serious drug offenses, or certain repeat offenders). The arrestee is entitled to a prompt detention hearing, ibid., and the maximum length of pretrial detention is limited by the stringent time limitations of the Speedy Trial Act.4 See 18 U. S. C. §3161 et seq. (1982 ed. and Supp. III). Moreover, as in Schall v. Martin, the conditions of confinement envisioned by the Act “appear to reflect the regulatory purposes relied upon by the” Government. *748467 U. S., at 270. As in Schall, the statute at issue here requires that detainees be housed in a “facility separate, to the extent practicable, from persons awaiting or serving sentences or being held in custody pending appeal.” 18 U. S. C. § 3142(i)(2). We conclude, therefore, that the pretrial detention contemplated by the Bail Reform Act is regulatory in nature, and does not constitute punishment before trial in violation of the Due Process Clause.

The Court of Appeals nevertheless concluded that “the Due Process Clause prohibits pretrial detention on the ground of danger to the community as a regulatory measure, without regard to the duration of the detention.” 794 F. 2d, at 71. Respondents characterize the Due Process Clause as erecting an impenetrable “wall” in this area that “no governmental interest — rational, important, compelling or otherwise — may surmount.” Brief for Respondents 16.

We do not think the Clause lays down any such categorical imperative. We have repeatedly held that the Government’s regulatory interest in community safety can, in appropriate circumstances, outweigh an individual’s liberty interest. For example, in times of war or insurrection, when society’s interest is at its peak, the Government may detain individuals whom the Government believes to be dangerous. See Ludecke v. Watkins, 335 U. S. 160 (1948) (approving un-reviewable executive power to detain enemy aliens in time of war); Moyer v. Peabody, 212 U. S. 78, 84-85 (1909) (rejecting due process claim of individual jailed without probable cause by Governor in time of insurrection). Even outside the exigencies of war, we have found that sufficiently compelling governmental interests can justify detention of dangerous persons. Thus, we have found no absolute constitutional barrier to detention of potentially dangerous resident aliens pending deportation proceedings. Carlson v. Landon, 342 U. S. 524, 537-542 (1952); Wong Wing v. United States, 163 U. S. 228 (1896). We have also held that the government may detain mentally unstable individuals who present a dan*749ger to the public, Addington v. Texas, 441 U. S. 418 (1979), and dangerous defendants who become incompetent to stand trial, Jackson v. Indiana, 406 U. S. 715, 731-739 (1972); Greenwood v. United States, 350 U. S. 366 (1956). We have approved of postarrest regulatory detention of juveniles when they present a continuing danger to the community. Schall v. Martin, supra. Even competent adults may face substantial liberty restrictions as a result of the operation of our criminal justice system. If the police suspect an individual of a crime, they may arrest and hold him until a neutral magistrate determines whether probable cause exists. Gerstein v. Pugh, 420 U. S. 103 (1975). Finally, respondents concede and the Court of Appeals noted that an arrestee may be incarcerated until trial if he presents a risk of flight, see Bell v. Wolfish, 441 U. S., at 534, or a danger to witnesses.

Respondents characterize all of these cases as exceptions to the “general rule” of substantive due process that the government may not detain a person prior to a judgment of guilt in a criminal trial. Such a “general rule” may freely be conceded, but we think that these cases show a sufficient number of exceptions to the rule that the congressional action challenged here can hardly be characterized as totally novel. Given the well-established authority of the government, in special circumstances, to restrain individuals’ liberty prior to or even without criminal trial and conviction, we think that the present statute providing for pretrial detention on the basis of dangerousness must be evaluated in precisely the same manner that we evaluated the laws in the cases discussed above.

The government’s interest in preventing crime by arrest-ees is both legitimate and compelling. De Veau v. Braisted, 363 U. S. 144, 155 (1960). In Schall, supra, we recognized the strength of the State’s interest in preventing juvenile crime. This general concern with crime prevention is no less compelling when the suspects are adults. Indeed, “[t]he *750harm suffered by the victim of a crime is not dependent upon the age of the perpetrator.” Schall v. Martin, supra, at 264-265. The Bail Reform Act of 1984 responds to an even more particularized governmental interest than the interest we sustained in Schall. The statute we upheld in Schall permitted pretrial detention of any juvenile arrested on any charge after a showing that the individual might commit some undefined further crimes. The Bail Reform Act, in contrast, narrowly focuses on a particularly acute problem in which the Government interests are overwhelming. The Act operates only on individuals who have been arrested for a specific category of extremely serious offenses. 18 U. S. C. § 3142(f). Congress specifically found that these individuals are far more likely to be responsible for dangerous acts in the community after arrest. See S. Rep. No. 98-225, at 6-7. Nor is the Act by any means a scattershot attempt to incapacitate those who are merely suspected of these serious crimes. The Government must first of all demonstrate probable cause to believe that the charged crime has been committed by the arrestee, but that is not enough. In a full-blown adversary hearing, the Government must convince a neutral decisionmaker by clear and convincing evidence that no conditions of release can reasonably assure the safety of the community or any person. 18 U. S. C. § 3142(f). While the Government’s general interest in preventing crime is compelling, even this interest is heightened when the Government musters convincing proof that the arrestee, already indicted or held to answer for a serious crime, presents a demonstrable danger to the community. Under these narrow circumstances, society’s interest in crime prevention is at its greatest.

On the other side of the scale, of course, is the individual’s strong interest in liberty. We do not minimize the importance and fundamental nature of this right. But, as our cases hold, this right may, in circumstances where the government’s interest is sufficiently weighty, be subordinated *751to the greater needs of society. We think that Congress’ careful delineation of the circumstances under which detention will be permitted satisfies this standard. When the Government proves by clear and convincing evidence that an arrestee presents an identified and articulable threat to an individual or the community, we believe that, consistent with the Due Process Clause, a court may disable the arrestee from executing that threat. Under these circumstances, we cannot categorically state that pretrial detention “offends some principle of justice so rooted in the traditions and conscience of our people as to be ranked as fundamental.” Snyder v. Massachusetts, 291 U. S. 97, 105 (1934).

Finally, we may dispose briefly of respondents’ facial challenge to the procedures of the Bail Reform Act. To sustain them against such a challenge, we need only find them “adequate to authorize the pretrial detention of at least some [persons] charged with crimes,” Schall, supra, at 264, whether or not they might be insufficient in some particular circumstances. We think they pass that test. As we stated in Schall, “there is nothing inherently unattainable about a prediction of future criminal conduct.” 467 U. S., at 278; see Jurek v. Texas, 428 U. S. 262, 274 (1976) (joint opinion of Stewart, Powell, and Stevens, JJ.); id., at 279 (White, J., concurring in judgment).

Under the Bail Reform Act, the procedures by which a judicial officer evaluates the likelihood of future dangerousness are specifically designed to further the accuracy of that determination. Detainees have a right to counsel at the detention hearing. 18 U. S. C. § 3142(f). They may testify in their own behalf, present information by proffer or otherwise, and cross-examine witnesses who appear at the hearing. Ibid. The judicial officer charged with the responsibility of determining the appropriateness of detention is guided by statutorily enumerated factors, which include the nature and the circumstances of the charges, the weight of the evidence, the history and characteristics of the putative of*752fender, and the danger to the community. § 3142(g). The Government must prove its case by clear and convincing evidence. § 3142(f). Finally, the judicial officer must include written findings of fact and a written statement of reasons for a decision to detain. § 3142(i). The Act’s review provisions, § 3145(c), provide for immediate appellate review of the detention decision.

We think these extensive safeguards suffice to repel a facial challenge. The protections are more exacting than those we found sufficient in the juvenile context, see Schall, supra, at 275-281, and they far exceed what we found necessary to effect limited postarrest detention in Gerstein v. Pugh, 420 U. S. 103 (1975). Given the legitimate and compelling regulatory purpose of the Act and the procedural protections it offers, we conclude that the Act is not facially invalid under the Due Process Clause of the Fifth Amendment.

B

Respondents also contend that the Bail Reform Act violates the Excessive Bail Clause of the Eighth Amendment. The Court of Appeals did not address this issue because it found that the Act violates the Due Process Clause. We think that the Act survives a challenge founded upon the Eighth Amendment.

The Eighth Amendment addresses pretrial release by providing merely that “[ejxcessive bail shall not be required.” This Clause, of course, says nothing about whether bail shall be available at all. Respondents nevertheless contend that this Clause grants them a right to bail calculated solely upon considerations of flight. They rely on Stack v. Boyle, 342 U. S. 1, 5 (1951), in which the Court stated that “[bjail set at a figure higher than an amount reasonably calculated [to ensure the defendant’s presence at trial] is ‘excessive’ under the Eighth Amendment.” In respondents’ view, since the Bail Reform Act allows a court essentially to set bail at an infinite amount for reasons not related to the risk of flight, it *753violates the Excessive Bail Clause. Respondents concede that the right to bail they have discovered in the Eighth Amendment is not absolute. A court may, for example, refuse bail in capital cases. And, as the Court of Appeals noted and respondents admit, a court may refuse bail when the defendant presents a threat to the judicial process by intimidating witnesses. Brief for Respondents 21-22. Respondents characterize these exceptions as consistent with what they claim to be the sole purpose of bail — to ensure the integrity of the judicial process.

While we agree that a primary function of bail is to safeguard the courts’ role in adjudicating the guilt or innocence of defendants, we reject the proposition that the Eighth Amendment categorically prohibits the government from pursuing other admittedly compelling interests through regulation of pretrial release. The above-quoted dictum in Stack v. Boyle is far too slender a reed on which to rest this argument. The Court in Stack had no occasion to consider whether the Excessive Bail Clause requires courts to admit all defendants to bail, because the statute before the Court in that case in fact allowed the defendants to be bailed. Thus, the Court had to determine only whether bail, admittedly available in that case, was excessive if set at a sum greater than that necessary to ensure the arrestees’ presence at trial.

The holding of Stack is illuminated by the Court’s holding just four months later in Carlson v. Landon, 342 U. S. 524 (1952). In that case, remarkably similar to the present action, the detainees had been arrested and held without bail pending a determination of deportability. The Attorney General refused to release the individuals, “on the ground that there was reasonable cause to believe that [their] release would be prejudicial to the public interest and would endanger the welfare and safety of the United States.” Id., at 529 (emphasis added). The detainees brought the same challenge that respondents bring to us today: the Eighth Amend*754ment required them to be admitted to bail. The Court squarely rejected this proposition:

“The bail clause was lifted with slight changes from the English Bill of Rights Act. In England that clause has never been thought to accord a right to bail in all cases, but merely to provide that bail shall not be excessive in those cases where it is proper to grant bail. When this clause was carried over into our Bill of Rights, nothing was said that indicated any different concept. The Eighth Amendment has not prevented Congress from defining the classes of cases in which bail shall be allowed in this country. Thus, in criminal cases bail is not compulsory where the punishment may be death. Indeed, the very language of the Amendment fails to say all arrests must be bailable.” Id., at 545-546 (footnotes omitted).

Carlson v. Landon was a civil case, and we need not decide today whether the Excessive Bail Clause speaks at all to Congress’ power to define the classes of criminal arrestees who shall be admitted to bail. For even if we were to conclude that the Eighth Amendment imposes some substantive limitations on the National Legislature’s powers in this area, we would still hold that the Bail Reform Act is valid. Nothing in the text of the Bail Clause limits permissible Government considerations solely to questions of flight. The only arguable substantive limitation of the Bail Clause is that the Government’s proposed conditions of release or detention not be “excessive” in light of the perceived evil. Of course, to determine whether the Government’s response is excessive, we must compare that response against the interest the Government seeks to protect by means of that response. Thus, when the Government has admitted that its only interest is in preventing flight, bail must be set by a court at a sum designed to ensure that goal, and no more. Stack v. Boyle, supra. We believe that when Congress has mandated detention on the basis of a compelling interest other than pre*755vention of flight, as it has here, the Eighth Amendment does not require release on bail.

I — I I — I HH

In our society liberty is the norm, and detention prior to trial or without trial is the carefully limited exception. We hold that the provisions for pretrial detention in the Bail Reform Act of 1984 fall within that carefully limited exception. The Act authorizes the detention prior to trial of arrestees charged with serious felonies who are found after an adversary hearing to pose a threat to the safety of individuals or to the community which no condition of release can dispel. The numerous procedural safeguards detailed above must attend this adversary hearing. We are unwilling to say that this congressional determination, based as it is upon that primary concern of every government — a concern for the safety and indeed the lives of its citizens — on its face violates either the Due Process Clause of the Fifth Amendment or the Excessive Bail Clause of the Eighth Amendment.

The judgment of the Court of Appeals is therefore

Reversed.

Justice Marshall,

with whom Justice Brennan joins, dissenting.

This case brings before the Court for the first time a statute in which Congress declares that a person innocent of any crime may be jailed indefinitely, pending the trial of allegations which are legally presumed to be untrue, if the Government shows to the satisfaction of a judge that the accused is likely to commit crimes, unrelated to the pending charges, at any time in the future. Such statutes, consistent with the usages of tyranny and the excesses of what bitter experience teaches us to call the police state, have long been thought incompatible with the fundamental human rights protected by our Constitution. Today a majority of this Court holds otherwise. Its decision disregards basic principles of justice *756established centuries ago and enshrined beyond the reach of governmental interference in the Bill of Rights.

A few preliminary words are necessary with respect to the majority’s treatment of the facts in this case. The two paragraphs which the majority devotes to the procedural posture are essentially correct, but they omit certain matters which are of substantial legal relevance.

The Solicitor General’s petition for certiorari was filed on July 21, 1986. On October 9, 1986, respondent Salerno filed a response to the petition. No response or appearance of counsel was filed on behalf of respondent Cafaro. The petition for certiorari was granted on November 3, 1986.

On November 19, 1986, respondent Salerno was convicted after a jury trial on charges unrelated to those alleged in the indictment in this case. On January 13, 1987, Salerno was sentenced on those charges to 100 years’ imprisonment. As of that date, the Government no longer required a pretrial detention order for the purpose of keeping Salerno incarcerated; it could simply take him into custody on the judgment and commitment order. The present case thus became moot as to respondent Salerno.1

*757The situation with respect to respondent Cafaro is still more disturbing. In early October 1986, before the Solicitor General’s petition for certiorari was granted, respondent Cafaro became a cooperating witness, assisting the Government’s investigation “by working in a covert capacity.”2 The information that Cafaro was cooperating with the Government was not revealed to his codefendants, including respondent Salerno. On October 9, 1986, respondent Cafaro was released, ostensibly “temporarily for medical care and treatment,” with the Government’s consent. Docket, SS 86 Cr. 245-2, p. 6 (MJL) (SDNY) (Lowe, J.).3 This release was conditioned upon execution of a personal recognizance bond in the sum of $1 million, under the general pretrial *758release provisions of 18 U. S. C. §3141 (1982 ed., Supp. III). In short, respondent Cafaro became an informant and the Government agreed to his release on bail in order that he might better serve the Government’s purposes. As to Cafaro, this case was no longer justiciable even before cer-tiorari was granted, but the information bearing upon the essential issue of the Court’s jurisdiction was not made available to us.

The Government thus invites the Court to address the facial constitutionality of the pretrial detention statute in a case involving two respondents, one of whom has been sentenced to a century of jail time in another case and released pending appeal with the Government’s consent, while the other was released on bail in this case, with the Government’s consent, because he had become an informant. These facts raise, at the very least, a substantial question as to the Court’s jurisdiction, for it is far from clear that there is now an actual controversy between these parties. As we have recently said, “Article III of the Constitution requires that there be a live case or controversy at the time that a federal court decides the case; it is not enough that there may have been a live case or controversy when the case was decided by the court whose judgment we are reviewing.” Burke v. Barnes, 479 U. S. 361, 363 (1987); see Sosna v. Iowa, 419 U. S. 393, 402 (1975); Golden v. Zwickler, 394 U. S. 103, 108 (1969). Only by flatly ignoring these matters is the majority able to maintain the pretense that it has jurisdiction to decide the question which it is in such a hurry to reach.

h-1 1 — 1

The majority approaches respondents’ challenge to the Act by dividing the discussion into two sections, one concerned with the substantive guarantees implicit in the Due Process Clause, and the other concerned with the protection afforded by the Excessive Bail Clause of the Eighth Amendment. This is a sterile formalism, which divides a unitary argument *759into two independent parts and then professes to demonstrate that the parts are individually inadequate.

On the due process side of this false dichotomy appears an argument concerning the distinction between regulatory and punitive legislation. The majority concludes that the Act is a regulatory rather than a punitive measure. The ease with which the conclusion is reached suggests the worthlessness of the achievement. The major premise is that “[ujnless Congress expressly intended to impose punitive restrictions, the punitive/regulatory distinction turns on ‘“whether an alternative purpose to which [the restriction] may rationally be connected is assignable for it, and whether it appears excessive in relation to the alternative purpose assigned [to it].”’” Ante, at 747 (citations omitted). The majority finds that “Congress did not formulate the pretrial detention provisions as punishment for dangerous individuals,” but instead was pursuing the “legitimate regulatory goal” of “preventing danger to the community.” Ibid.4 Concluding that pretrial detention is not an excessive solution to the problem of preventing danger to the community, the majority thus finds that no substantive element of the guarantee of due process invalidates the statute.

*760This argument does not demonstrate the conclusion it purports to justify. Let us apply the majority’s reasoning to a similar, hypothetical case. After investigation, Congress determines (not unrealistically) that a large proportion of violent crime is perpetrated by persons who are unemployed. It also determines, equally reasonably, that much violent crime is committed at night. From amongst the panoply of “potential solutions,” Congress chooses a statute which permits, after judicial proceedings, the imposition of a dusk-to-dawn curfew on anyone who is unemployed. Since this is not a measure enacted for the purpose of punishing the unemployed, and since the majority finds that preventing danger to the community is a legitimate regulatory goal, the curfew statute would, according to the majority’s analysis, be a mere “regulatory” detention statute, entirely compatible with the substantive components of the Due Process Clause.

The absurdity of this conclusion arises, of course, from the majority’s cramped concept of substantive due process. The majority proceeds as though the only substantive right protected by the Due Process Clause is a right to be free from punishment before conviction. The majority’s technique for infringing this right is simple: merely redefine any measure which is claimed to be punishment as “regulation,” and, magically, the Constitution no longer prohibits its imposition. Because, as I discuss in Part III, infra, the Due Process Clause protects other substantive rights which are infringed by this legislation, the majority’s argument is merely an exercise in obfuscation.

The logic of the majority’s Eighth Amendment analysis is equally unsatisfactory. The Eighth Amendment, as the majority notes, states that “[ejxcessive bail shall not be required.” The majority then declares, as if it were undeniable, that: “[t]his Clause, of course, says nothing about whether bail shall be available at all.” Ante, at 752. If excessive bail is imposed the defendant stays in jail. The same result is achieved if bail is denied altogether. Whether the *761magistrate sets bail at $1 billion or refuses to set bail at all, the consequences are indistinguishable. It would be mere sophistry to suggest that the Eighth Amendment protects against the former decision, and not the latter. Indeed, such a result would lead to the conclusion that there was no need for Congress to pass a preventive detention measure of any kind; every federal magistrate and district judge could simply refuse, despite the absence of any evidence of risk of flight or danger to the community, to set bail. This would be entirely constitutional, since, according to the majority, the Eighth Amendment “says nothing about whether bail shall be available at all.”

But perhaps, the majority says, this manifest absurdity can be avoided. Perhaps the Bail Clause is addressed only to the Judiciary. “[W]e need not decide today,” the majority says, “whether the Excessive Bail Clause speaks at all to Congress’ power to define the classes of criminal arrestees who shall be admitted to bail.” Ante, at 754. The majority is correct that this question need not be decided today; it was decided long ago. Federal and state statutes which purport to accomplish what the Eighth Amendment forbids, such as imposing cruel and unusual punishments, may not stand. See, e. g., Trop v. Dulles, 356 U. S. 86 (1958); Furman v. Georgia, 408 U. S. 238 (1972). The text of the Amendment, which provides simply that “[ejxcessive bail shall not be required, nor excessive fines imposed, nor cruel and unusual punishments inflicted,” provides absolutely no support for the majority’s speculation that both courts and Congress are forbidden to inflict cruel and unusual punishments, while only the courts are forbidden to require excessive bail.5

*762The majority’s attempts to deny the relevance of the Bail Clause to this case are unavailing, but the majority is nonetheless correct that the prohibition of excessive bail means that in order “to determine whether the Government’s response is excessive, we must compare that response against the interest the Government seeks to protect by means of that response.” Ante, at 754. The majority concedes, as it must, that “when the Government has admitted that its only interest is in preventing flight, bail must be set by a court at a sum designed to ensure that goal, and no more.” Ibid. But, the majority says, “when Congress has mandated detention on the basis of a compelling interest other than prevention of flight, as it has here, the Eighth Amendment does not require release on bail.” Ante, at 754-755. This conclusion follows only if the “compelling” interest upon which Congress acted is an interest which the Constitution permits Congress to further through the denial of bail. The majority does not ask, as a result of its disingenuous division of the analysis, if there are any substantive limits contained in both the Eighth Amendment and the Due Process Clause which render this system of preventive detention unconstitutional. The majority does not ask because the answer is apparent and, to the majority, inconvenient.

Ill

The essence of this case may be found, ironically enough, in a provision of the Act to which the majority does not refer. Title 18 U. S. C. §3142(j) (1982 ed., Supp. Ill) provides that “[njothing in this section shall be construed as modifying or limiting the presumption of innocence.” But the very pith *763and purpose of this statute is an abhorrent limitation of the presumption of innocence. The majority’s untenable conclusion that the present Act is constitutional arises from a specious denial of the role of the Bail Clause and the Due Process Clause in protecting the invaluable guarantee afforded by the presumption of innocence.

“The principle that there is a presumption of innocence in favor of the accused is the undoubted law, axiomatic and elementary, and its enforcement lies at the foundation of the administration of our criminal law.” Coffin v. United States, 156 U. S. 432, 453 (1895). Our society’s belief, reinforced over the centuries, that all are innocent until the state has proved them to be guilty, like the companion principle that guilt must be proved beyond a reasonable doubt, is “implicit in the concept of ordered liberty,” Palko v. Connecticut, 302 U. S. 319, 325 (1937), and is established beyond legislative contravention in the Due Process Clause. See Estelle v. Williams, 425 U. S. 501, 503 (1976); In re Winship, 397 U. S. 358, 364 (1970). See also Taylor v. Kentucky, 436 U. S. 478, 483 (1978); Kentucky v. Whorton, 441 U. S. 786, 790 (1979) (Stewart, J., dissenting).

The statute now before us declares that persons who have been indicted may be detained if a judicial officer finds clear and convincing evidence that they pose a danger to individuals or to the community. The statute does not authorize the Government to imprison anyone it has evidence is dangerous; indictment is necessary. But let us suppose that a defendant is indicted and the Government shows by clear and convincing evidence that he is dangerous and should be detained pending a trial, at which trial the defendant is acquitted. May the Government continue to hold the defendant in detention based upon its showing that he is dangerous? The answer cannot be yes, for that would allow the Government to imprison someone for uncommitted crimes based upon “proof” not beyond a reasonable doubt. The result must therefore be that once the indictment has failed, detention *764cannot continue. But our fundamental principles of justice declare that the defendant is as innocent on the day before his trial as he is on the morning after his acquittal. Under this statute an untried indictment somehow acts to permit a detention, based on other charges, which after an acquittal would be unconstitutional. The conclusion is inescapable that the indictment has been turned into evidence, if not that the defendant is guilty of the crime charged, then that left to his own devices he will soon be guilty of something else. “ ‘If it suffices to accuse, what will become of the innocent?’” Coffin v. United States, supra, at 455 (quoting Ammianus Marcellinus, Rerum Gestarum Libri Qui Supersunt, L. XVIII, c. 1, A. D. 359).

To be sure, an indictment is not without legal consequences. It establishes that there is probable cause to believe that an offense was committed, and that the defendant committed it. Upon probable cause a warrant for the defendant’s arrest may issue; a period of administrative detention may occur before the evidence of probable cause is presented to a neutral magistrate. See Gerstein v. Pugh, 420 U. S. 103 (1975). Once a defendant has been committed for trial he may be detained in custody if the magistrate finds that no conditions of release will prevent him from becoming a fugitive. But in this connection the charging instrument is evidence of nothing more than the fact that there will be a trial, and

“release before trial is conditioned upon the accused’s giving adequate assurance that he will stand trial and submit to sentence if found guilty. Like the ancient practice of securing the oaths of responsible persons to stand as sureties for the accused, the modern practice of requiring a bail bond or the deposit of a sum of money subject to forfeiture serves as additional assurance of the *765presence of an accused.” Stack v. Boyle, 342 U. S. 1, 4-5 (1951) (citation omitted).6

The finding of probable cause conveys power to try, and the power to try imports of necessity the power to assure that the processes of justice will not be evaded or obstructed.7 “Pretrial detention to prevent future crimes against society at large, however, is not justified by any concern for holding a trial on the charges for which a defendant has been arrested.” 794 F. 2d 64, 73 (CA2 1986) (quoting United States v. Melendez-Carrion, 790 F. 2d 984, 1002 (CA2 1986) (opinion of Newman, J.)). The detention purportedly authorized by this statute bears no relation to the Government’s power to try charges supported by a finding of probable cause, and thus the interests it serves are outside the scope of interests which may be considered in weighing the excessiveness of bail under the Eighth Amendment.

*766It is not a novel proposition that the Bail Clause plays a vital role in protecting the presumption of innocence. Reviewing the application for bail pending appeal by members of the American Communist Party convicted under the Smith Act, 18 U. S. C. § 2385, Justice Jackson wrote:

“Grave public danger is said to result from what [the defendants] may be expected to do, in addition to what they have done since their conviction. If I assume that defendants are disposed to commit every opportune disloyal act helpful to Communist countries, it is still difficult to reconcile with traditional American law the jailing of persons by the courts because of anticipated but as yet uncommitted crimes. Imprisonment to protect society from predicted but unconsummated offenses is . . . unprecedented in this country and . . . fraught with danger of excesses and injustice....” Williamson v. United States, 95 L. Ed. 1379, 1382 (1950) (opinion in chambers) (footnote omitted).

As Chief Justice Vinson wrote for the Court in Stack v. Boyle, supra: “Unless th[e] right to bail before trial is preserved, the presumption of innocence, secured only after centuries of struggle, would lose its meaning.” 342 U. S., at 4.

> I — i

There is a connection between the peculiar facts of this case and the evident constitutional defects in the statute which the Court upholds today. Respondent Cafaro was originally incarcerated for an indeterminate period at the request of the Government, which believed (or professed to believe) that his release imminently threatened the safety of the community. That threat apparently vanished, from the Government’s point of view, when Cafaro agreed to act as a covert agent of the Government. There could be no more eloquent demonstration of the coercive power of authority to imprison upon prediction, or of the dangers which the almost *767inevitable abuses pose to the cherished liberties of a free society.

“It is a fair summary of history to say that the safeguards of liberty have frequently been forged in controversies involving not very nice people.” United States v. Rabinowitz, 339 U. S. 56, 69 (1950) (Frankfurter, J., dissenting). Honoring the presumption of innocence is often difficult; sometimes we must pay substantial social costs as a result of our commitment to the values we espouse. But at the end of the day the presumption of innocence protects the innocent; the shortcuts we take with those whom we believe to be guilty injure only those wrongfully accused and, ultimately, ourselves.

Throughout the world today there are men, women, and children interned indefinitely, awaiting trials which may never come or which may be a mockery of the word, because their governments believe them to be “dangerous.” Our Constitution, whose construction began two centuries ago, can shelter us forever from the evils of such unchecked power. Over 200 years it has slowly, through our efforts, grown more durable, more expansive, and more just. But it cannot protect us if we lack the courage, and the self-restraint, to protect ourselves. Today a majority of the Court applies itself to an ominous exercise in demolition. Theirs is truly a decision which will go forth without authority, and come back without respect.

I dissent.

Justice Stevens,

dissenting.

There may be times when the Government’s interest in protecting the safety of the community will justify the brief detention of a person who has not committed any crime, see ante, at 748-749, see also United States v. Greene, 497 F. 2d 1068, 1088-1089 (CA7 1974) (Stevens, J., dissenting).1 To *768use Judge Feinberg’s example, it is indeed difficult to accept the proposition that the Government is without power to detain a person when it is a virtual certainty that he or she would otherwise kill a group of innocent people in the immediate future. United States v. Salerno, 794 F. 2d 64, 77 (CA2 1986) (dissenting opinion). Similarly, I am unwilling to decide today that the police may never impose a limited curfew during a time of crisis. These questions are obviously not presented in this case, but they lurk in the background and preclude me from answering the question that is presented in as broad a manner as Justice. Marshall has. Nonetheless, I firmly agree with Justice Marshall that the provision of the Bail Reform Act allowing pretrial detention on the basis of future dangerousness is unconstitutional. Whatever the answers are to the questions I have mentioned, it is clear to me that a pending indictment may not be given any weight in evaluating an individual’s risk to the community or the need for immediate detention.

If the evidence of imminent danger is strong enough to warrant emergency detention, it should support that preventive measure regardless of whether the person has been charged, convicted, or acquitted of some other offense. In this case, for example, it is unrealistic to assume that the danger to the community that was present when respondents were at large did not justify their detention before they were indicted, but did require that measure the moment that the grand jury found probable cause to believe they had committed crimes in the past.2 It is equally unrealistic to assume that the danger will vanish if a jury happens to acquit them. *769Justice Marshall has demonstrated that the fact of indictment cannot, consistent with the presumption of innocence and the Eighth Amendment’s Excessive Bail Clause, be used to create a special class, the members of which are, alone, eligible for detention because of future dangerousness.

Several factors combine to give me an uneasy feeling about the case the Court decides today. The facts set forth in Part I of Justice Marshall’s opinion strongly support the possibility that the Government is much more interested in litigating a “test case” than in resolving an actual controversy concerning respondents’ threat to the safety of the community. Since Salerno has been convicted and sentenced on other crimes, there is no need to employ novel pretrial detention procedures against him. Cafaro’s case is even more curious because he is apparently at large and was content to have his case argued by Salerno’s lawyer even though his interests would appear to conflict with Salerno’s. But if the merits must be reached, there is no answer to the arguments made in Parts II and III of Justice Marshall’s dissent. His conclusion, and not the Court’s, is faithful to the “fundamental principles as they have been understood by the traditions of our people and our law.” Lochner v. New York, 198 U. S. 45, 76 (1905) (Holmes, J., dissenting). Accordingly, I respectfully dissent.

Fried, J.

In these cases we consider the constitutionality of G. L. c. 276, § 58A (1994 ed.),1 which allows a judge to detain, prior to trial, a person accused of certain crimes on the ground of that person’s dangerousness (preventive detention). In Aime v. Commonwealth, 414 Mass. 667, 668 (1992), we concluded that G. L. c. 276, § 58, as amended through St. 1992, c. 201, § 3, a predecessor to § 58A, violated the due process clause of the Fourteenth Amendment to the United States Constitution. That earlier provision differed in several respects from the Bail Reform Act of 1984, 18 U.S.C. § 3142, id. at 676 & n.12, which the Supreme Court of the United States upheld against a facial constitutional challenge in *773 United States v. Salerno, 481 U.S. 739 (1987). In Aime, supra at 681 n,18, we expressly declined to consider the constitutionality of preventive detention under our Declaration of Rights. Section 58A is similar in most respects to the Federal Bail Reform Act.

Kevin Mendonza now challenges its constitutionality under arts. 1, 10, 12, and 26 of the Massachusetts Declaration of Rights. He challenges as well those aspects of § 58A that differ from the Federal legislation on both State and Federal constitutional grounds. Vincent Callender challenges the constitutionality under both Constitutions of the provision in § 58A (4) which the Commonwealth and the District Court read as requiring the pretrial detention of a defendant for three days if the Commonwealth requests a continuance prior to a hearing on dangerousness. Although the Bad Reform Act contains a similar provision,2 the Supreme Court did not address it in Salerno.

We conclude that in the respects challenged by Mendonza, § 58A is valid under the Fourteenth Amendment and the Declaration of Rights, both on its face and as applied to him. A provision that would require detention during a three-day continuance requested by the Commonwealth, without a hearing or any determination of need could not stand under either Constitution, but it is not necessary to read § 58A (4) in this way. The Commonwealth must show good cause for such a continuance, and the judge must in each case make a specific finding indicating what such cause is.

I

A

General Laws c. 276, § 58A, allows a District or Superior Court judge, on motion of the Commonwealth, to order the pretrial detention of a person charged with certain felonies and other ofienses involving the use, or threatened use, of violence or abuse, or the violation of protective orders, including protective orders obtained under G. L. c. 209A (1994 *774ed.). Section 58A sets out a comprehensive scheme of measures available with respect to arrested persons charged with crime. The preferred pretrial disposition is release on personal recognizance. G. L. c. 276, § 58 (1994 ed.). Only if such release will not reasonably assure the presence of the arrested person at trial or the safety of any other persons may the judge impose the least restrictive conditions proposed to him necessary to provide such assurance, including bail in a reasonably necessary amount. G. L. c. 276, § 58A (2). Release is conditioned on the accused’s not committing a Federal, State, or local crime during the period of release. G. L. c. 276, § 58A (2) (A). Only if the judge determines after a hearing that no conditions of release will reasonably assure the safety of any other persons may he order the pretrial detention of a person arrested for one of the specified offenses. G. L. c. 276, § 58A (3). A person so detained shall be brought to trial as soon as reasonably possible. In the absence of good cause, he may not be detained longer than ninety days. G. L. c. 276, § 58A (3). The judge is precluded from imposing a financial condition that results in pretrial detention in order to assure the safety of other persons, although financial conditions having that effect are not precluded for the purpose of assuring his appearance before the court. G. L. c. 276, § 58A (2) and (3). Accordingly, the provision should end any tendency to require high bail as a device for effecting preventive detention because it directs that all decisions based on dangerousness be made under the procedures set forth for that specific purpose.

The hearing at which the detention order is sought must take place at the arrested person’s first appearance, absent a request for a continuance by one of the parties. The hearing may be continued for three business days on motion of the Commonwealth. The prisoner shall be detained during such continuance on a showing that there was probable cause to arrest him. § 58A (4). The judge at the hearing must find the requisite dangerousness by clear and convincing evidence, G. L. c. 276, § 58A (3) and (4), and shall consider, among other factors, the nature and seriousness of the danger posed by the prisoner if released, the nature and circumstances of the offense charged, his family ties, his employment record, his record of convictions, and whether he is on bail pending adjudication of other prior charges. G. L. c. 276, § 58A (5). The prisoner has a right to counsel at the hearing, including, *775if appropriate, appointed counsel. He shall have the right to testify, present witnesses and information, and to cross-examine witnesses who appear against him. The rules of evidence applicable in a criminal trial do not apply at the hearing. The prisoner may be detained pending completion of the hearing. If the judge orders detention, he must issue a written opinion stating his findings of fact as to why such detention is necessary under the terms of the statute. G. L. c. 276, § 58A (4). The judge may reopen the order at any time to consider material new information, G. L. c. 276, § 58A (4), and the prisoner has the right to petition the Superior Court for review of a decision in the District Court. G. L. c. 276, § 58A (7). There is no provision for review of a Superior Court detention order although such review may be had by application to a single justice of this court. G. L. c. 211, § 3 (1994 ed.). The detention must take place, to the extent practicable, in a facility separate from that in which are held persons awaiting or serving sentence or awaiting the outcome of an appeal. G. L. c. 276, § 58A (4).

B

Four police officers came to Mendonza’s house to serve him with a protective order obtained by his wife, requiring him to move out of the family home. Mendonza asked permission to gather his belongings from his bedroom. As Mendonza and three of the officers approached the bedroom he ran into the room and barricaded the door with furniture. When the officers got inside, they found Mendonza standing in a comer pouring gasoline from a canister over his head with one hand and holding a book of matches in the other. Mendonza said, “I have matches and I’ll set the whole place on fire.” One of the officers subdued Mendonza with “caps-tun” (a disabling substance), arrested him, and charged him with three counts of armed assault with intent to murder, assault and battery on a police officer, and violation of a protective order for his failure to leave the house when served with the order.

At his arraignment on these charges, the Commonwealth sought a detention order under G. L. c. 276, § 58A, but the judge ordered Mendonza committed to Bridgewater State Hospital for an evaluation of his competency to stand trial and of his criminal responsibility. During this time a grand *776jury indicted Mendonza on the charges enumerated above. Almost four months after the initial arraignment and after a series of additional commitments, Mendonza was returned to the Superior Court. On the same day, a hearing was held, and a § 58A detention order issued. The evidence at the hearing included the testimony of a police officer as to statements made to the officer by Mendonza’s wife about threats he had made to kill her and an incident in which he had awakened her as he stood over her with a knife. The judge also noted evidence that Mendonza was, or had been, addicted to drugs, that he had a prior history of psychiatric hospitalizations, and that psychiatrists who had observed him during his commitment had thought him an “acute suicide risk.” Mendonza filed a petition before a single justice of this court, challenging, among other things, the constitutionality of § 58A. The single justice denied relief and this appeal followed. Mendonza pleaded guilty to the charges of assault and battery, violation of the restraining order, and entered an Alford plea, see North Carolina v. Alford, 400 U.S. 25 (1970), to the charge of attempted arson. The armed assault charges were placed on file. He is no longer in custody.

Callender was arrested for banging on the door of an apartment which a G. L. c. 209A protective order forbade him from visiting. At the time, Callender was on probation for three violations of G. L. c. 209A protective orders. At his arraignment, the Commonwealth moved for a § 58A detention order and requested a continuance for three days. The District Court judge declined to grant the three-day continuance on the ground that the § 58 (4) requirement that a continuance be granted merely at the request of the Commonwealth was unconstitutional. Callender was, however, held for that period on the probation violation, at the end of which time he admitted to sufficient facts on the charges, and was found guilty. Accordingly, no detention order ever issued. Nevertheless the judge reported to the Appeals Court the following question:

“Does the provision of M. G. L. c. 276, s. 58A(4), which allows for the pre-trial detention of a criminal defendant for three days upon the filing of a motion for an order of pre-trial detention and a request for a three-day continuance, without any other preliminary showing of danger*777ousness or any other type of hearing, violate a criminal defendant’s rights under the Fourteenth Amendment to the United States Constitution or Article 12 of the Massachusetts Declaration of Rights?”

We granted Callender’s application for direct appellate review.

II

A

We note at the outset that, although both cases are moot since both Callender and Mendonza. have admitted guilt and are no longer subject to § 5 8A, it is entirely appropriate that we proceed to adjudicate both claims. The issues are certainly of recurring importance to the administration of justice in the Commonwealth, yet owing to the limited time during which detention is authorized and the very short time during which a continuance on the motion of the Commonwealth may be granted, they would almost certainly evade review in this court. In such cases we have often proceeded to consider the important recurring question. See Aime v. Commonwealth, 414 Mass. 667, 670 (1993) (in case raising constitutional concerns nearly identical to those here, “we do not hesitate to reach the merits of cases that no longer involve a live dispute so as to further the public interest”); Lockhart v. Attorney Gen., 390 Mass. 780, 783-785 (1984) (“in our discretion . . . we should not offer our views on constitutional issues that will not necessarily arise again in substantially the same form,” but “[i]f ... the issues . . . reappear, they need not evade review before they become moot”); Boston Herald, Inc. v. Superior Court Dep’t of the Trial Court, 421 Mass. 502, 504 (1995) (taking the opportunity to “reiterate the principles which govern” in a constitutional case even though the dispute before the court was moot, and it was “highly speculative” that it would repeat itself). Cf. Roe v. Wade, 410 U.S. 113 (1973). Nor is it indispensable that the case be capable of repetition in respect only to the particular claimant. This doctrine is designed to assist in the clarification of the law generally, and not simply to assist the situation of a particular party. See Wolf v. Commissioner of Pub. Welfare, 367 Mass. 293, 298 (1975); Norwood Hosp. v. Munoz, 409 Mass. 116, 121 (1991).

*778B

Mendonza designates his first and most general challenge to the scheme of § 58A under the rubric of substantive due process. Our decision in Aime, supra, also considered at length the distinction between substantive and procedural due process and pointed to the long and strongly entrenched tradition that the former concept enjoys under both our own and the Federal Constitutions. Id. at 673-675. The serious challenge Mendonza makes speaks at a level more basic than the particular provisions assuring fair procedures in both Constitutions, which is why in considering the validity under the Fourteenth Amendment of the predecessor statute we were obliged to recur to that linguistically awkward concept of substantive due process. The objection to the procedures in the previous statute were not simply to the way in which the issue of dangerousness was adjudicated but more fundamentally to the use of a prediction of dangerousness as a predicate for a deprivation of liberty at all. This challenge has a firmer textual footing in arts. 1 and 10 than in the Fourteenth Amendment. Both articles recognize explicitly a substantive right to liberty,3 but the Aime court surely was correct to find such a substantive right in the Federal Constitution as well. The firm and detailed safeguards of the criminal process set out in the Fourth, Fifth, Sixth, and Eighth Amendments, the substantive prohibitions against bills of attainder and ex post facto laws in §§ 9 and 10 of art. I of the United States Constitution, and the implicit constitutional requirement that the government has the burden of proving guilt beyond a reasonable doubt, In re Winship, 397 U.S. 358, 361-364 (1970), would have little meaning if the same deprivations could be worked free of their constraints, though under a different *779rubric than that of the criminal law. Thus it is an underlying premise of both Constitutions that certain deprivations of life, liberty, or property — however they are designated — must be the work of the criminal law and must comply with the safeguards required there.

The Supreme Court in United States v. Salerno, 481 U.S. 739 (1987), noted that there have been several contexts in which deprivations of liberty have been approved without the full panoply of rights and procedures attendant on the criminal process. It gave as examples the detention of dangerous persons in time of war or insurrection, of potentially dangerous resident aliens, of mentally unstable persons who present a danger to .the public, of dangerous defendants who become incompetent to stand trial, and of juveniles awaiting trial who present a continuing danger to the community. Id. at 748-749. The looming danger in moving directly from this list to the conclusion that therefore any and all deprivations of liberty need only be tested by the test of due process in general — that is, are the procedural guarantees commensurate with the severity of the imposition, Mathews v. Eldridge, 424 U.S. 319 (1976) — is that the distinctive nature of the procedural guarantees of the criminal process enshrined with such emphasis and specificity in both Constitutions would be swallowed up in the undifferentiated weighing invited by a generalized due process analysis. Unfortunately, the Salerno decision did little to acknowledge or head off this danger beyond noting that the special safeguards of the Bill of Rights are only applicable in the case of punishment and not where the deprivation of liberty is regulatory. Salerno, supra at 746-747. But until we have a firm grasp of what is punishment and what is regulation, this is an example of explaining the obscure by the more obscure. The Justices of this court recently had occasion to address that very question in considering the constitutionality of a bill, since passed into law, permitting notification by law enforcement agencies to police, public and private agencies, and the public of the presence in a community of certain sex offenders who had completed their sentences.

“[Our criminal justice system] is built on the premise of personal responsibility: Individuals are taken to exercise a choice whether or not to obey the law. Punish*780ment is the consequence visited on willing disobedience. . . . Our system of criminal justice is not predictive in the sense that it would seek systematically to identify those who may present a danger to society and to incapacitate them before that danger may be realized.

“[T]he Justices are governed by the overriding concern that the distinctiveness of the criminal justice system not be elided, lest we move in the direction of a regime where persons, and not just particular activities and occupations, are seen as regulated by government, rather than a regime where persons are seen as personally responsible for conforming their conduct to the clearly promulgated standards of the criminal law. . . .” Opinion of the Justices, post 1201, 1218-1223 (1996).

The preventive regime of § 58A has a particularly heavy burden to overcome because it is explicitly “predictive” and “seek[s] systematically to identify those who may present a danger to society and to incapacitate them before that danger may be realized.” Id. at 1219. We are not inclined to believe that the Supreme Court in Salerno wished to allow any and all loss of liberty to be justified by a prediction of dangerousness with only generalized due process safeguards. See Foucha v. Louisiana, 504 U.S. 71, 75-83 (1992). In any event, if we are wrong about that, the text and structure of our own more emphatic guarantees in arts. 1, 10, and 14 do not permit that baleful conclusion. Each of the examples the Supreme Court cited included some additional feature that justified taking it out of the regime of the criminal law: the wartime cases involve the government’s enhanced powers to protect the citizenry from the grave dangers of war and insurrection, see Moyer v. Peabody, 212 U.S. 78 (1909); the detention of certain aliens invokes the government’s more plenary power over persons who may in certain circumstances be removed from the country, see Carlson v. Landon, 342 U.S. 524, 537-542 (1952); juveniles and the insane both are thought to act with diminished responsibility and thus are susceptible to regimes at once both more intrusive and more paternalistic, see Jackson v. Indiana, 406 U.S. 715, 731-739 (1972).

In the case of § 5 8A, like the Bail Reform Act, the factor additional to predicted dangerousness is the pendency of a *781charge of serious crime, a charge which we may take to have been tested by the standard of probable cause. It requires explanation why that factor should suffice to allow at the very beginning a loss of liberty, which is a consequence that may be imposed only at the end of the criminal process under a much higher standard. Indeed it was argued in Salerno and is argued to us now that the excessive bail clauses of the Eighth Amendment and of art. 26 state the only constitutionally permissible means to the only permissible end of pretrial control of an accused, that is assuring the accused’s presence at trial and his noninterference with witnesses and other aspects of the pretrial process. The Federal statute followed extensive legislative factfinding that tended to show that a surprising percentage of crimes are committed by persons awaiting trial. There is no reason to doubt that these findings are generally applicable. Senate Rep. No. 225, 98th Cong., 1st Sess., 4-9 (1983). Goldkamp, Gottfredson, Jones & Weiland, Personal Liberty and Community Safety 89-91 (1995) (in cases before Boston Municipal Court, even after detaining dangerous defendants under § 5 8A, thirteen per cent of released defendants were rearrested and twenty per cent failed to appear). Cf. Note, Preventive Detention: An Empirical Analysis, 6 Harv. C.R.-C.L.L. Rev. 289 (1971). This fact, together with the delays between arrest and the conclusion of the trial and the denser population and greater anonymity of our communities since the time of the framings, argues for some form of preliminary relief for the government in incapacitating persons who pose a particular danger to the public. The necessary predicate that there be probable cause to believe that the person deemed dangerous has committed a serious crime offers a more substantial hurdle to the loss of liberty than some of the instances cited by the Supreme Court, and the conclusion of the trial itself provides an inevitable end point to the State’s preventive authority. Finally, this very preventive authority may sometimes have been exercised sub rosa by the imposition of very high bail, which cannot be explained simply by the need to assure the accused’s presence at trial and his noninterference with the pretrial process. See, e.g., Snow v. Commonwealth, 404 Mass. 1007 (1989) (upheld bail increase from $100,000 [$10,000 cash] to $500,000 [$50,000 cash] based on threat by defendant that he would kill three people if he were released on bail). Moreover, certain *782quite traditional techniques, such as revocation of bail for misconduct, G. L. c. 276, § 58; Delaney v. Commonwealth, 415 Mass. 490 (1993), necessarily imply some authority to control the behavior of offenders before their guilt has been definitively adjudicated. It would be odd indeed if bail could be revoked for misconduct which is not tried to conviction but that the very misconduct which led to the original charge might not allow preliminary detention. It is one of the virtues of § 58A that the need for preliminary relief on account of dangerousness is addressed explicitly and both the Commonwealth and the accused may direct their arguments and proofs openly to that question.

In rejecting Mendonza’s substantive due process claim, we reject the claim that a person accused of crime may never be detained on grounds of dangerousness prior to an adjudication of guilt in a criminal trial. In the balance of this opinion we consider the objections Mendonza makes to specific features of § 58A.4

C

1. Mendonza argues that § 58A violates art. 12 in that the Commonwealth need only prove dangerousness by clear and convincing evidence instead of beyond a reasonable doubt. This argument is supported by our decisions that the civil commitment of dangerous persons as well as the commitment of sexually dangerous persons require the showing of *783dangerousness to be made beyond a reasonable doubt, while the Supreme Court has ruled in analogous circumstances that clear and convincing proof is sufficient to satisfy due process under the Fourteenth Amendment. Addington v. Texas, 441 U.S. 418, 423-433 (1979). In holding that proof beyond a reasonable doubt was required to commit sexually dangerous persons, we said, “we do not think that the Commonwealth can be permitted to label a person sexually dangerous and deprive him of his liberty for an indeterminate period which may extend over the rest of his natural life merely on proof that it is more likely than not that the person is sexually dangerous.” Andrews, petitioner, 368 Mass. 468, 489 (1975). See Superintendent of Worcester State Hosp. v. Hagberg, 374 Mass. 271, 275-277 (1978) (requiring beyond a reasonable doubt standard to indefinite civil commitments). There is, however, a significant difference between the cases in which we have ruled that clear and convincing evidence is not sufficient to justify civil confinement and preventive detention under § 58A. Although the civil confinement of the dangerously mentally ill and sexually dangerous persons requires annual review under the same high standard of proof, the term of confinement is potentially indefinite and even lifelong. Indeed, the confinement under G. L. c. 123A (1994 ed.) is “for an indefinite period of a minimum of one day and a maximum of such person’s natural life.” G. L. c. 123A, § 5 (1994 ed.). Under § 58A, by contrast, the detention is limited and preliminary. The trial itself is the “main event,” Wainwright v. Sykes, 433 U.S. 72, 90 (1977), at which the accused’s liberty is definitively adjudicated. The better analogy, thus, is not to confinement of the dangerously mentally ill and sexually dangerous persons, but to the various temporary confinements authorized by G. L. c. 123 (1994 ed.) for observation or for emergency restraint. We have never suggested that proof beyond a reasonable doubt would be required for these temporary confinements.5

If there is sufficient constitutional warrant for the Legisla*784turc to institute pretrial detention, to require proof beyond a reasonable doubt at the very early stage of the proceedings when the Commonwealth is preparing its case on the merits, which of course must eventually be proved beyond a reasonable doubt, risks making this form of preliminary relief unavailable in practice.6

*7852. Mendonza also argues that the proceedings authorized by § 58A violate art. 12 because the rules of evidence do not apply in those proceedings. G. L. c. 276, § 58A (4). Of particular concern is the admissibility of hearsay evidence in such proceedings. More particularly still Mendonza insists that he should not suffer a deprivation of liberty on the basis of unreliable hearsay evidence. The assertion begs the question whether all hearsay evidence is so inherently unreliable that it may never form part of the Commonwealth’s case in a § 58A hearing. The principle that governs this issue is stated in Mullane v. Central Hanover Bank & Trust Co., 339 U.S. 306, 313 (1950), and adopted by us in Myers v. Commonwealth, 363 Mass. 843, 854 (1973): “At the very least, due process of the law requires that deprivation of life, liberty or property by adjudication be preceded by notice and opportunity for a hearing appropriate to the nature of the case” (emphasis supplied).

In Myers, supra at 855-856, we did say that in a probable cause hearing “the defendant must be given the opportunity to cross-examine his accusers and present testimony in his own behalf in order to insure that the hearing’s vital screening function will be effectuated.” We held these procedural guarantees were “appropriate to the nature of the case,” id. at 855, quoting Goldberg v. Kelly, 397 U.S. 254, 268 (1970). We also said “the rules of evidence at the preliminary hearing should in general be the same rules that are applicable at the criminal trial.” Myers, supra at 849 n.6. However, the rules of evidence and more specifically the rules against the admission of hearsay evidence are not required in all proceedings. Article 12, like the Sixth Amendment, only grants the right to a defendant in a criminal proceeding to confront the witness offering the declaration. Thus even in a criminal trial, where the right of confrontation is at its zenith, there are numerous rules allowing the introduction of hearsay evidence. Furthermore, this court has sanctioned certain proceedings where the rules of evidence do not apply, even where deprivation of liberty is at stake as is the case here. Commonwealth v. Durling, *786407 Mass. 108, 114-119 (1990) (rules of evidence need not apply in a hearing for revocation of probation). The probable cause hearing serves to test the strength of the Commonwealth’s case against an accused, while the pretrial detention hearing seeks only a flexible and reliable preliminary determination of dangerousness, and thus it is more “appropriate to the nature of [the probable cause] case,” Myers, supra at 855, that the evidence in the probable cause hearings be judged by the same rules of admissibility as will govern the trial of the case. In Myers itself, the vice was not the introduction of hearsay testimony but the hearing judge’s improper curtailment of the defendant’s cross-examination of the complaining witness and presentation of his own evidence. Id. at 845. The proceeding under § 58A suffers from no such defects and allows counsel for the accused the fullest opportunity not only to cross-examine those witnesses produced by the Commonwealth but also to offer witnesses and evidence — including hearsay evidence — on the accused’s behalf. That is how the Supreme Court interpreted the analogous Federal procedure, Salerno, supra at 751-752, and we are sure that § 58A allows at least as ample an opportunity for testing of, and response to, the Commonwealth’s showing of dangerousness.

3. Mendonza argues that several differences between the Federal statute and § 58A rise to the level of constitutional defects.

a. The Federal Bail Reform Act may be invoked where the accused is charged with a crime of violence, a capital crime or a crime allowing life imprisonment, certain drug offenses for which the maximum sentence is ten or more years, or any listed felony if the accused has already been convicted of two or more of the listed offenses. 18 U.S.C. § 3142(f)(1). The Supreme Court in Salerno mentioned this limitation to serious crimes as one of the factors supporting the constitutionality of the Act. United States v. Salerno, supra at 747. The predicate offenses listed in § 58A (1) are defined somewhat differently but are also limited to serious crimes. Included are felonies that involve the use of force or attempted or threatened use of force. Subsection (1) goes on to include “any other felony that by its nature involves a substantial risk that physical force . . . may result . . . whether or not a person has been placed at risk thereof,” and specifically lists *787burglary and arson. This may exceed the scope of the Federal list in some cases, but the menace of dangerousness is plainly enough present in the excess that the statute does not raise constitutional doubts. Mendonza emphasizes the inclusion of what he designates as lesser offenses: violations of restraining orders to vacate the family house, to cease abuse, to protect the personal liberty of the other spouse and children, and to award child custody on an emergency basis. See G. L. c. 208, §§ 18, 34B, 34C; c. 209, § 32; c. 209A, §§ 3, 4, 5 (1994 ed.). That the list is more extensive than that in the Federal Bail Reform Act reflects the different roles of State and Federal law enforcement. Federal law enforcement is directed to specified crimes often having a significant national impact, while it is the States which have the primary and basic responsibility for assuring the ordinary conditions of peace, security, and good order in our communities. Offenses involving violence or threats of violence within the family are peculiarly within the province of State concern, so that it is no surprise that they were not included among the listed offenses in 18 U.S.C. § 3142(f)(1). Indeed, a distressing proportion of violent crimes, including homicides, , are committed in domestic settings by persons well known to each other. See P. Claus & M. Ranel, United States Department of Justice, Bureau of Justice Statistics, Special Report: Family Violence (1984); Zorza, The Criminal Law of Misdemeanor Domestic Violence, 1970-1990, 83 J. Crim. L. & Criminology 46 (1992) (for example, thirty-one per cent of women murdered aré killed by husbands, ex-husbands, or lovers). That these potentially lethal domestic offenses may sometimes be characterized as misdemeanors is of no significance.

b. Mendonza complains that § 5 8A, unlike the Federal statute, does not require a finding that there is probable cause to believe that the person to be detained has committed one of the predicate offenses. It is true that the Supreme Court in the Salerno case gave as one of the features of the Act that assured its constitutionality the requirement that “[t]he Government . . . first of all demonstrate probable cause to believe that the charged crime has been committed by the arrestee. . . .” Salerno at 750. But the provision to which Mendonza points as imposing that requirement, 18 U.S.C. § 3142(e), does not in fact explicitly set out any such general predicate finding. The Supreme Court ruled as it did because *788the requirement is implicit in the scheme of the statute and in the system of pretrial procedures in which it is embedded. The same is true of § 5 8A: persons arrested without a warrant must be brought before a judicial magistrate without unreasonable delay (usually within twenty-four hours) and may not be further held unless the Commonwealth demonstrates probable cause for their arrest or a hearing is waived. Jenkins v. Chief Justice of the Dist. Court Dep’t, 416 Mass. 221, 232, 237 (1993). Indeed, the provision for continuances in subsection (4), which we consider below, states explicitly that the accused may only be detained during the delay there granted if probable cause for his arrest has been demonstrated. In this case Mendonza had been indicted by a grand jury for the predicate offenses, and thus probable cause had been determined by the most constitutionally solemn means available.

c. The Federal statute instructs judges to consider the “nature and circumstances of the offense charged” and “the weight of the evidence against that person” in determining whether the accused should be detained. 18 U.S.C. § 3142(g)(1) and (2). Section 58A (5) directs the judge to consider, among other factors, “the nature and seriousness of the danger” posed by the accused and “the nature and circumstances of the offense charged,” but does not mention the weight of the evidence regarding the predicate offense. This omission is without significance. As we have said, it must already have been established that there is probable cause to believe that the defendant has committed the crime, and we see no reason to ask the court to reconsider that issue on some other standard. The focus of the detention hearing quite properly is on the further question of dangerousness, which must be made out by clear and convincing evidence. The statute directs the courts to the factors that bear on the rational determination of that issue.

4. The judge detained Mendonza on a finding that there was clear and convincing evidence that “no conditions of release will reasonably assure the safety of others, namely the defendant’s spouse and children.” Mendonza complains that there was, however, no evidence that he had ever harmed his wife or children. We do not take him to be arguing that clear and convincing evidence of danger to his family was lacking, and if he is, we reject that argument. Rather it seems that Mendonza is arguing that the Commonwealth must show *789some nexus between the dangerousness of the charged offenses and the danger the accused might present if left at large. The statute does not require such a nexus, although it does direct attention, as we have said, to the nature and circumstances of the offense charged as relevant to the issue of dangerousness. Nor do we see any constitutional reason to insist on such a nexus. A person charged with burglary of an inhabited dwelling at nighttime, for example, may be detained if it is shown that he may harm others in quite different settings. The statistics on crimes committed by persons awaiting trial do not show any close correlation between the crimes of which they stand accused and the further crimes they may commit. E.g., Note, Preventive Detention: An Empirical Analysis, 6 Harv. C.R.-C.L.L. Rev. 289, 294 (1971); Bureau of Justice Statistics, Sourcebook of Criminal Justice Statistics 445 (1995). In any event, there was a close connection between the offenses with which Mendonza was charged and a potential threat to his family. One of those offenses was violation of a protective order protecting Mendonza’s family and ordering Mendonza to vacate the family home. The alleged assault on the police officers occurred in the family home when the officers ordered Mendonza to vacate the house to protect the family.

Ill

Callender makes a different and narrower claim. Conceding arguendo the constitutionality of § 58A in all other respects, Callender challenges the provision in § 58A (4) which appears to give the Commonwealth an extension of three business days more or less for the asking, during which time an accused is detained simply on the showing that there was probable cause to arrest him. The analogous provision in 18 U.S.C. § 3142(f)(2) is similarly worded: it states that “[e]xcept for good cause ... a continuance on motion of the attorney for the Government may not exceed three days . . . [djuring . . . [which] the person shall be detained . . . .” The Supreme Court in Salerno did not address that provision, nor has it passed on it explicitly since.7 In the absence of any express intimation on the meaning and constitutionality of *790this provision by the Supreme Court we must make our own determination. If we should be wrong on either score, we make clear that our decision is based on our construction of § 58A and the demands of art. 12 as well. Michigan v. Long, 463 U.S. 1032 (1983).

The Commonwealth’s expansive reading of the continuance provision raises serious constitutional concerns. We have just considered Mendonza’s arguments that the Declaration of Rights and our constitutional traditions do not allow the State to deprive a person of his liberty except on conviction or other determination of dangerousness made beyond a reasonable doubt. We did not reject those arguments because we thought them frivolous or far-fetched but because, in spite of their considerable force, they are too absolute. We concluded that on balance the Declaration of Rights, which may be more but is certainly not less protective than the Federal Constitution, allows preventive detention in carefully circumscribed circumstances and subject to quite demanding procedures. The Commonwealth’s burden to prove dangerousness by clear and convincing evidence is an important part of those procedures. One of the reasons we did not require an even higher standard of proof is that detention under § 58A is temporary and provisional.

Having engaged in such a careful balancing in respect to those claims we cannot, without more, admit that any detention of a person after his required first appearance, at which probable cause is determined, can be permitted on so casual and untested a basis.8 Even a brief street stop in the most difficult circumstances requires a justification by reasonable *791suspicion. Terry v. Ohio, 392 U.S. 1 (1968). The judge who reported this question to us stated:

“While no empirical analysis has been performed, this Court is well aware of numerous instances where the Commonwealth has filed the original petition pursuant to s. 58A(4), as well as the request for the three-day continuance pursuant to s. 58A(4), and after the three-day detention and at the time of the hearing, has withdrawn its petition. The criminal defendant has been detained for a period of three days without any preliminary showing of dangerousness or without any hearing simply upon the filing of the petition.”

The Commonwealth did not address this allegation, nor have we been pointed to any other discussion of it. However general such a practice may or may not be, it is a vice of the Commonwealth’s contention that like any other form of unilateral and untested (either before or after the fact) exercise of power it invites abuse. We need not belabor the point: the Commonwealth’s contention is so extreme that we shall not construe the statute to underwrite it.

There is, to be sure, another side to the issue. Callender himself goes too far in suggesting that, if § 58A is generally valid, then any detention, even that during which the Commonwealth is gathering its evidence to make a clear and convincing showing of dangerousness, must also be justified by clear and convincing evidence of dangerousness. In many situations where the threat of danger looms, particularly where the police make an arrest shortly after learning of the commission of a serious and violent crime, the brief period of time allowed before first appearance will not suffice for this further purpose. The police may know enough to establish probable cause that such a crime has been committed and *792that the arrested person committed it. They may also have grounds for concern that this person presents a continuing danger, but not nearly enough to meet the high burden of § 58A. The statute allows detention during a continuance to meet just such cases. To require that the arrested person be released pending that further inquiry may invite a sub rosa relaxation of the clear and convincing standard.

The sensible reading of the continuance provision is to allow such a three-day continuance at the request of the Commonwealth only if the Commonwealth can show good cause for it. Such a reading would meet the serious constitutional doubts raised by the automatic stay for which the Commonwealth argues. The judge should then make a specific finding that such cause has been shown and what such cause is.

IV

We affirm the judgment of the single justice as to Mendonza’s claim. As to Callender, the answer to the question reported to us is that properly construed to require good cause § 58A (4) is not unconstitutional.

So ordered.