announced the judgment of the Court and an opinion in which Mr. Justice Stewart, Mr. Justice White, and Mr. Justice Blackmun join.

This case presents the narrow but important question of whether the constitutional privilege against compulsory self-incrimination is infringed by California’s so-called “hit and run” statute which requires the driver of a motor vehicle involved in an accident to stop at the scene and give his name and address. Similar “hit and run” or “stop and report” statutes are in effect in all 50 States and the District of Columbia.

On August 22, 1966, respondent Byers was charged in a two-count criminal complaint with two misdemeanor violations of the California Vehicle Code. Count 1 charged that on August 20 Byers passed another vehicle without maintaining the “safe distance” required by *426§21750 (Supp. 1971). The second count charged that Byers had been involved in an accident but had failed to stop and identify himself as required by § 20002 (a)(1) (Supp. 1971).

This statute provides:1

“The driver of any vehicle involved in an accident resulting in damage to any property including vehicles shall immediately stop the vehicle at the scene of the accident and shall then and there . . . [l]ocate and notify the owner or person in charge of such property of the name and address of the driver and owner of the vehicle involved . . . .”

It is stipulated that both charges arose out of the same accident.

Byers demurred to Count 2 on the ground that it violated his privilege against compulsory self-incrimination. His position was ultimately sustained by the California Supreme Court.2 That court held that the privilege protected a driver who “reasonably believes that compliance with the statute will result in self-incrimination.” 71 *427Cal. 2d 1039, 1047, 458 P. 2d 465, 471 (1969). Here the eourt found that Byers’ apprehensions were reasonable because compliance with §20002 (a)(1) confronted him with “substantial hazards of self-incrimination.” Nevertheless the court upheld the validity of the statute by inserting a judicially created use restriction on the disclosures that it required. The court concluded, however, that it would be “unfair” to punish Byers for his failure to comply with the statute because he could not reasonably have anticipated the judicial promulgation of the use restriction.3 We granted certiorari to assess the validity of the California Supreme Court’s premise that without a use restriction §20002 (a)(1) would violate the privilege against compulsory self-incrimination. We conclude that there is no conflict between the statute and the privilege.

(1)

Whenever the Court is confronted with the question of a compelled disclosure that has an incriminating potential, the judicial scrutiny is invariably a close one. Tension between the State’s demand for disclosures and the protection of the right against self-incrimination is likely to give rise to serious questions. Inevitably these must be resolved in terms of balancing the public need on the one hand, and the individual claim to constitutional protections on the other; neither interest can be treated lightly.

An organized society imposes many burdens on its constituents. It commands the filing of tax returns for income; it requires producers and distributors of consumer goods to file informational reports on the manu*428facturing process and the content of products, on the wages, hours, and working conditions of employees. Those who borrow money on the public market or issue securities for sale to the public must file various information reports; industries must report periodically the volume and content of pollutants discharged into our waters and atmosphere. Comparable examples are legion.4

In each of these situations there is some possibility of prosecution — often a very real one — for criminal offenses disclosed by or deriving from the information that the law compels a person to supply. Information revealed by these reports could well be “a link in the chain” of evidence leading to prosecution and conviction. But under our holdings the mere possibility of incrimination is insufficient to defeat the strong policies in favor of a disclosure called for by statutes like the one challenged here.

United States v. Sullivan, 274 U. S. 259 (1927), shows that an application of the privilege to the California statute is not warranted. There a bootlegger was prosecuted for failure to file an income tax return. He claimed that the privilege against compulsory self-incrimination afforded him a complete defense because filing a return would have tended to incriminate him by revealing the unlawful source of his income. Speaking for the Court, Mr. Justice Holmes rejected this claim on the ground that it amounted to “an extreme if not an extravagant application of the Fifth Amendment.” Id., at 263-264.5 *429Sullivan’s tax return, of course, increased his risk of prosecution and conviction for violation of the National Prohibition Act. But the Court had no difficulty in concluding that an extension of the privilege to cover that kind of mandatory report would have been unjustified. In order to invoke the privilege it is necessary to show that the compelled disclosures will themselves confront the claimant with “substantial hazards of self-incrimination.”

The components of this requirement were articulated in Albertson v. SACB, 382 U. S. 70 (1965), and later in Marchetti v. United States, 390 U. S. 39 (1968), Grosso v. United States, 390 U. S. 62 (1968), and Haynes v. United States, 390 U. S. 85 (1968). In Albertson the Court held that an order requiring registration by individual members of a Communist organization violated the privilege. There Sullivan was distinguished:

“In Sullivan the questions in the income tax return were neutral on their face and directed at the public at large, but here they are directed at a highly selective group inherently suspect of criminal activities. Petitioners’ claims are not asserted in an essentially noncriminal and regulatory area of inquiry, but against an inquiry in an area permeated with criminal statutes, where response to any of the . . . questions in context might involve the petitioners in the admission of a crucial element of a crime.” 382 U. S., at 79 (emphasis added).

Albertson was followed by Marchetti and Grosso where the Court held that the privilege afforded a complete defense to prosecutions for noncompliance with federal *430gambling tax and registration requirements. It was also followed in Haynes where petitioner had been prosecuted for failure to register a firearm as required by federal statute. In each of these cases the Court found that compliance with the statutory disclosure requirements would confront the petitioner with “substantial hazards of self-incrimination.” E. g., Marchetti v. United States, 390 U. S., at 61.

In all of these cases the disclosures condemned were only those extracted from a “highly selective group inherently suspect of criminal activities” and the privilege was applied only in “an area permeated with criminal statutes” — not in “an essentially noncriminal and regulatory area of inquiry.” E. g., Albertson v. SACB, 382 U. S., at 79; Marchetti v. United States, 390 U. S., at 47.

Although the California Vehicle Code defines some criminal offenses, the statute is essentially regulatory, not criminal. The California Supreme Court noted that §20002 (a)(1) was not intended to facilitate criminal convictions but to promote the satisfaction of civil liabilities arising from automobile accidents. In Marchetti the Court rested on the reality that almost everything connected with gambling is illegal under “comprehensive” state and federal statutory schemes. The Court noted that in almost every conceivable situation compliance with the statutory gambling requirements would have been incriminating. Largely because of these pervasive criminal prohibitions, gamblers were considered by the Court to be “a highly selective group inherently suspect of criminal activities.”

In contrast, § 20002 (a) (1), like income tax laws, is directed at all persons — here all persons who drive automobiles in California. This group, numbering as it does in the millions, is so large as to render § 20002 (a)(1) a statute “directed at the public at large.” Albertson v. SACB, 382 U. S., at 79, construing United States v. Sulli*431van, 274 U. S. 259 (1927). It is difficult to consider this group as either “highly selective” or “inherently suspect of criminal activities.” Driving an automobile, unlike gambling, is a lawful activity. Moreover, it is not a criminal offense under California law to be a driver “involved in an accident.” An accident may be the fault of others; it may occur without any driver having been at fault. No empirical data are suggested in support of the conclusion that there is a relevant correlation between being a driver and criminal prosecution of drivers. So far as any available information instructs us, most accidents occur without creating criminal liability even if one or both of the drivers are guilty of negligence as a matter of tort law.

The disclosure of inherently illegal activity is inherently risky. Our decisions in Albertson and the cases following illustrate that truism. But disclosures with respect to automobile accidents simply do not entail the kind of substantial risk of self-incrimination involved in Marchetti, Orosso, and Haynes. Furthermore, the statutory purpose is noncriminal and self-reporting is indispensable to its fulfillment.

(2)

Even if we were to view the statutory reporting requirement as incriminating in the traditional sense, in our view it would be the “extravagant” extension of the privilege Justice Holmes warned against to hold that it is testimonial in the Fifth Amendment sense. Compliance with §20002 (a)(1) requires two things: first, a driver involved in an accident is required to stop at the scene; second, he is required to give his name and address. The act of stopping is no more testimonial — indeed less so in some respects — than requiring a person in custody to stand or walk in a police lineup, to speak prescribed words, or to give samples of handwriting, fingerprints, or *432blood. United States v. Wade, 388 U. S. 218, 221-223 (1967); Schmerber v. California, 384 U. S. 767, 764 and n. 8 (1966); 8 J. Wigmore, Evidence § 2265, pp. 386-400 (McNaughton rev. 1961). Disclosure of name and address is an essentially neutral act. Whatever the collateral consequences of disclosing name and address, the statutory purpose is to implement the state police power to regulate use of motor vehicles.

Section 20002 (a)(1) first requires that a driver involved in an accident “shall immediately stop the vehicle at the scene of the accident. . . .” It is, of course, possible that compliance with this requirement might ultimately lead to prosecution for some contemporaneous criminal violation of the motor vehicle code if one occurred, or an unrelated offense, always provided such offense could be established by independent evidence. In that sense it might furnish the authorities with what might be called “a link in the chain of evidence needed to prosecute ....” Hoffman v. United States, 341 U. S. 479, 486 (1951). In Schmerber v. California, supra, at 764, the Court held that “the privilege is a bar against compelling 'communications’ or 'testimony,’ but . . . compulsion which makes a suspect or accused the source of ‘real or physical evidence’ does not violate it.” There the petitioner had been compelled to undergo the forcible withdrawal of blood samples for alcohol content analysis, and the Court sustained this procedure over petitioner’s claim that he had been compelled to furnish evidence against himself. See also Holt v. United States, 218 U. S. 245, 252 (1910) (Holmes, J.) (requiring defendant to model a blouse would be barred only by “an extravagant extension of the Fifth Amendment”).

Stopping in compliance with § 20002 (a)(1) therefore does not provide the State with “evidence of a testimonial or communicative nature” within the meaning of the Constitution. Schmerber v. California, supra, at 761. *433It merely provides the State and private parties with the driver’s identity for, among other valid state needs, the study of causes of vehicle accidents and related purposes, always subject to the driver’s right to assert a Fifth Amendment privilege concerning specific inquiries.

Respondent argues that since the statutory duty to stop is imposed only on the “driver of any vehicle involved in an accident,” a driver’s compliance is testimonial because his action gives rise to an inference that he believes that he was the “driver of [a] vehicle involved in an accident.” From this, the respondent tells us, it can be further inferred that he was indeed the operator of an “accident involved” vehicle. In Wade, however, the Court rejected the notion that such inferences are communicative or testimonial. There the respondent was placed in a lineup to be viewed by persons who had witnessed a bank robbery. At one point he was compelled to speak the words alleged to have been used by the perpetrator. Despite the inference that the respondent uttered the words in his normal undisguised voice, the Court held that the utterances were not of a “testimonial” nature in the sense of the Fifth Amendment privilege even though the speaking might well have led to identifying him as the bank robber. United States v. Wade, supra, at 222-223. Furthermore, the Court noted in Wade that no question was presented as to the admissibility in evidence at trial of anything said or done at the lineup. Id., at 223. Similarly no such problem is presented here. Of course, a suspect’s normal voice characteristics, like his handwriting, blood, fingerprints, or body may prove to be the crucial link in a chain of evidentiary factors resulting in prosecution and conviction. Yet such evidence may be used against a defendant.

After having stopped, a driver involved in an accident is required by § 20002 (a)(1) to notify the driver of the other vehicle of his name and address. A name, linked *434with a motor vehicle, is no more incriminating than the tax return, linked with the disclosure of income, in United States v. Sullivan, supra. It identifies but does not by itself implicate anyone in criminal conduct.6

Although identity, when made known, may lead to inquiry that in turn leads to arrest and charge, those developments depend on different factors and independent evidence. Here the compelled disclosure of identity could have led to a charge that might not have been made had the driver fled the scene; but this is true only in the same sense that a taxpayer can be charged on the basis of the contents of a tax return or failure to file an income tax form. There is no constitutional right to refuse to file an income tax return or to flee the scene of an accident in order to avoid the possibility of legal involvement.

The judgment of the California Supreme Court is vacated and the case is remanded for further proceedings not inconsistent with this opinion.

Vacated and remanded.

concurring in the judgment.

For the reasons which follow, I concur in the judgment of the Court.

I

The respondent, Byers, as a driver of a vehicle involved in an accident resulting in property damage, was charged in a two-count complaint with overtaking another vehicle in a manner proscribed by § 21750 of the California Vehicle Code (Supp. 1971) and failing to comply with the requirements of § 20002 (a) of the California Vehicle *435Code (Supp. 1971).1 The parties have stipulated that the accident was caused by respondent's violation of § 21750 of the California Vehicle Code. App. 36. The California Supreme Court has held that in circumstances where a driver involved in an accident has reason to believe his compliance with § 20002 (a) creates a substantial risk of disclosure of incriminating evidence, the Fifth Amendment requires that the State must either excuse his noncompliance if he properly pleads the privilege against self-incrimination in a subsequent prosecution for failure to comply or forgo the use of any information disclosed by the State’s compulsion. Construing the state statute as wholly nonprosecutorial in purpose, the court then held that imposition of a restriction on the use of the information or its fruits in a subsequent criminal prosecution for the conduct causing the accident would be consistent with the state legislative purpose.

I cannot separate the requirement that the individual stop from the requirement that he identify himself for purposes of applying either the “testimonial- — non-testimonial” classification of Schmerber v. California, 384 U. S. 757 (1966), or the “substantial danger of incrimination” test of Hoffman v. United States, 341 U. S. 479 (1951). The California Supreme Court treated these requirements, in the primary context in which the statute operates, as compelling identification of oneself as a party involved in the statutorily regulated event. If evidence of that self-identification were admitted at trial, it would *436certainly be “testimonial.” If all that is offered at trial is the identification evidence of third-party witnesses, it still does not follow from United States v. Wade, 388 U. S. 218 (1967), that because the policies of the Fifth Amendment are not significantly affected by state compulsion to cooperate in the production of real evidence where the State has independently focused investigation on the defendant, these policies are similarly unaffected where the State — in pursuit of “real” evidence — demands of the defendant that he focus the investigation on himself. See generally Mansfield, The Albertson Case: Conflict Between the Privilege Against Self-Incrimination and the Government's Need for Information, 1966 Sup. Ct. Rev. 103, 121-124.

It may be said that requiring the defendant to focus attention on himself as an accident participant is not equivalent to requiring the defendant to focus attention on himself as a criminal suspect. And that proposition raises the underlying issue which we must resolve in this case: how do the various verbal formulations for assessing the legal significance of the risk of incrimination, developed by the Court primarily in the context of the criminal process, see Malloy v. Hogan, 378 U. S. 1, 11-14 (1964); Hoffman v. United States, 341 U. S. 479 (1951), operate in the context of the state collection of data for purposes essentially unrelated to criminal prosecution?

II

The California Supreme Court in the present case resolved that issue as follows:

“Decisions of the United States Supreme Court make clear that the privilege against self-incrimination is a personal one, and that whether the government may require a disclosure depends upon the facts of each case. Invocation of the privilege is not limited to situations in which the purpose of the *437inquiry is to get an incriminating answer. It is the efject of the answer that is determinative. ‘To sustain [a claim of] privilege, it need only be evident from the implications of the question, in the setting in which it is asked, that a responsive answer to the question or an explanation of why it cannot be answered might be dangerous because injurious disclosure could result.’ ” [Citing Hoffman, supra, and Mansfield, supra.] Byers v. Justice Court, 71 Cal. 2d 1039, 1046, 458 P. 2d 465, 470 (1969) (emphasis in original).

The California Supreme Court was surely correct in considering that the decisions of this Court have made it clear that invocation of the privilege is not limited to situations where the purpose of the inquiry is to get an incriminating answer. For example, in the context of civil proceedings the privilege is generally available to witnesses as long as a substantial risk of self-incrimination can be made out by the witness. See McCarthy v. Arndstein, 266 U. S. 34 (1924); Murphy v. Waterfront Comm’n, 378 U. S. 52, 94 (1964) (White, J., concurring). And, in fairness to the state tribunal whose decision we are reviewing, it must be recognized that a reading of our more recent cases — especially Marchetti v. United States, 390 U. S. 39 (1968), and Grosso v. United States, 390 U. S. 62 (1968) — suggests the conclusion that the applicability of the privilege depends exclusively on a determination that, from the individual’s point of view, there are “real” and not “imaginary” risks of self-incrimination in yielding to state compulsion. Thus, Marchetti and Grosso (and the cases they overruled, United States v. Kahriger, 345 U. S. 22 (1953), and Lewis v. United States, 348 U. S. 419 (1955)), start from an assumption of a nonprosecutorial governmental purpose in the decision to tax gambling revenue; those cases go on to apply what in another context I have *438called the “real danger v. imaginary possibility” standard, see Emspak v. United States, 349 U. S. 190, 209 n. 1 (1955), to the gambling reporting requirements imposed on the individual in order to determine the constitutionality of those requirements. See Marchetti v. United States, supra, at 48-54; Grosso v. United States, supra, at 66-67. A judicial tribunal whose position with respect to the elaboration of constitutional doctrine is subordinate to that of this Court certainly cannot be faulted for reading these opinions as indicating that the “inherently-suspect-class” factor is relevant only as an indicium of genuine incriminating risk as assessed from the individual’s point of view. See also Haynes v. United States, 390 U. S. 85, 95-98 (1968); Leary v. United States, 395 U. S. 6, 16-18 (1969); and compare the dissenting opinion of Mr. Justice Frankfurter in the Kahriger case, supra.

That inference from qur past cases was the central premise of the California Supreme Court’s opinion. See Byers v. Justice Court, supra, at 1042-1043, 458 P. 2d, at 468. Thus, that tribunal is in agreement with the conclusion reached in today’s opinion of The Chief Justice that the class of accident participants is not an “inherently suspect” class within the meaning of Marchetti, Grosso, and Haynes. But the state court went on to conclude that the widespread prevalence of criminal sanctions as a means of regulating driving conduct cast a substantial shadow of suspicion over the class, and that this circumstance plus the driver’s awareness that his illegal behavior caused the accident rendered the driver’s conclusion that he would incriminate himself by complying with the statute sufficiently plausible to support an assertion of the privilege. Id., at 1045-1046, 458 P. 2d, at 470. Starting from the California Supreme Court’s premise and looking at our cases defining the test *439for risks of incrimination, see Malloy v. Hogan, 378 U. S., at 11-14; Hoffman v. United States, 341 U. S. 479 (1951); Rogers v. United States, 340 U. S. 367, 374 (1951); Brown v. Walker, 161 U. S. 591 (1896), I would have to reach the same conclusion. I am, however, for the reasons stated in the remainder of this opinion, constrained to hold that the presence of a “real” and not “imaginary” risk of self-incrimination is not a sufficient predicate for extending the privilege against self-incrimination to regulatory schemes of the character involved in this case.

Ill

First, it is instructive to consider the implications of adhering to the premise which the California Supreme Court drew from our prior cases. In United States v. Sullivan, 274 U. S. 259 (1927), Mr. Justice Holmes stated his view that “ [i] t would be an extreme if not an extravagant application of the Fifth Amendment to say that it authorized a man to refuse to state the amount of his income because it had been made in crime.” Id., at 263-264.2 Yet- — at least for an individual whose income is largely or entirely derived from illegal activities — it is, I think, manifestly unsatisfactory to maintain that it should be “ ‘perfectly clear [to him], from a careful consideration of all the circumstances in the case [that his statement of the amount of his income] cannot possibly have [a] tendency’ to incriminate.” Hoffman v. United States, 341 U. S., at 488. (Emphasis in original.) Certainly that individual would have good reason to suspect that if the State is permitted to introduce his income tax return into evidence, the informa*440tion. contained therein — even if wholly confined to a statement of his gross income — will, when combined with other evidence derived from independent sources, incriminate him. United States v. Burr, 25 F. Cas. 38, 40 (No. 14,692e) (CC Va. 1807). Nor can the “required records” doctrine of Shapiro v. United States, 335 U. S. 1 (1948), be invoked to avoid that conclusion; for that doctrine, as applied to this situation, would simply mean that the taxing power is of sufficient import to justify compelled self-incrimination.

If Mr. Justice Holmes’ assertion that it would be an extreme, if not extravagant, extension of the Fifth Amendment to apply it in such a situation strikes a responsive chord, it is because the primary context from which the privilege emerges is that of the criminal process, both in the investigatory and trial phases. When applied in that context, the sole governmental interest that the privilege defeats is the enforcement of law through criminal sanctions. And, with regard to the witness’ privilege, the judge can, for the most part, draw the line between “real” and “imaginary” risks of incrimination in the marginal cases, thereby offsetting the tendency for the privilege to become an absolute right not to disclose any information at all.

But, of course, governmental interests other than the enforcement of criminal laws are affected by an extension of the privilege to all instances of governmental compulsion to disclose information. In the present case, the interests of the State of California in a system of personal financial responsibility for automobile accidents are implicated. Indeed, in my Brother Brennan’s view, the price which the State must pay for utilizing compulsory self-reporting to assure personal financial responsibility on the highways is to forgo use of the criminal sanction to regulate driving habits in all cases where the individual would *441be required to comply with § 20002.3 And I emphasize that the logic of the Hoffman standard is such that the result cannot be said to turn on an empirical assumption that there is a lower correlation between criminal prosecutions and highway accidents, than between criminal *442prosecutions and income disclosure. For if the privilege is truly a personal one, and the central standard is the presence of “real” as opposed to “imaginary” risks of self-incrimination, such general empirical differences can only function as evidentiary indicia in assessing the particular individual’s claim, in all the circumstances of his particular case, that if he were to comply with the reporting requirement he would run a genuine risk of incrimination. And, if the Hoffman standard is to be truly applied — as opposed to indulging in a collection of artificial, if not disingenuous judgments that the risks of incrimination are not there when they really are there — we will have to recognize that in the absence of some sort of immunity grant the individual will be required to decide whether or not to comply without the guidance of a judicial decision as to how the standard applies to his personal situation. That means that the marginal cases should be resolved in the individual’s favor. What we are really talking about, then, is either a standard for risks of self-incrimination which protects all personal judgments which are not patently frivolous, or a grant of immunity potentially applicable to all instances of compelled “self-reporting.”

Of course, the California Supreme Court took the position that the permissible state objective in the reporting requirement and the constitutional values protected by the Fifth Amendment could be accommodated by imposing a restriction on prosecutorial use of the disclosed *443information and its fruits. See infra, at 444-447. See generally McKay, Self-Incrimination and the New Privacy, 1967 Sup. Ct. Rev. 193, 229-231; Mansfield, supra, at 163-166. But that accommodation leaves the Government’s capacity to utilize self-reporting schemes practically impaired by the necessary presumption that evidence used in a prosecution after the individual discloses his relationship to the regulated transaction would not have been available if the individual had not complied with the statute. In the context of “hit-and-run” statutes a use immunity — unless honored in the breach by consistent findings of “no taint” — is likely to render doubtful the State’s ability to prosecute in a large class of cases where illegal driving has caused accidents. On the other hand, it would seem unlikely that the state legislature will accept the California Supreme Court’s invitation to override the use requirement if the legislative judgment is that the State’s ability to use the criminal sanction is too severely handicapped by a use restriction. See infra, at 446-447. For the impact of a practically self-executing claim of privilege on the noncriminal objectives of the reporting requirement would be even more severe. Even under a use restriction, then, the choice open to the State is to forgo prosecution in at least a large number of accident cases involving illegal driving — the precise situation where criminal sanctions are likely to be most appropriate — or to forgo self-reporting in a large class of accident cases.

Mr. Justice Brennan argues that to draw this conclusion from the record in this case is to “flout” the conclusion of the California Supreme Court and the California Legislature that imposition of a use restriction as a condition for prosecuting Byers for noncompliance with § 20002 (a) is “not at all inconsistent with the asserted state interests.” Post, at 476. Apparently my Brother Brennan maintains that imposition of a use restriction *444in the circumstances of this case will not, in fact, significantly interfere with the State’s ability to enforce criminal sanctions relating to driving behavior where that behavior culminates in an accident causing property damage.4 That, in any event, seems to be the view he attributes to the California Supreme Court and the California Legislature. See opinion of Mr. Justice Brennan', post, at 475-476 and n. 10.

But that is certainly not the position the California Supreme Court took, nor the position that court attributed to the state legislature. Thus, having first concluded that the Federal Constitution required recognition of an assertion of the privilege in all situations where the individual confronts “real” and not “imaginary” risks of self-incrimination, the California Supreme Court set about the task of ascertaining the legislative preference between legislative goals which, by virtue of the imposition of a federal constitutional requirement, were placed in conflict:

“Finally, it is instructive, in determining legislative intent, to consider an analogous field of legislation involving a similar conflict between requiring disclosures for noncriminal purposes and the privilege against self-incrimination. In the statutes requiring drivers involved in accidents resulting in personal injury or death to file accident reports, the *445Legislature has explicitly subordinated the state’s prosecutorial interest to the interest in obtaining the disclosure.

“In the present case there is no problem of conflicting state and federal interests; it is the state which both demands disclosure of information in ‘hit-and-run’ accidents and prosecutes those who commit criminal acts on the highways. Imposing use-restrictions in the present case merely involves this court in making a judgment, based on an assessment of probable legislative intent, that the Legislature would prefer to have the provisions of section 20002 of the Vehicle Code upheld even in cases involving possible criminal misconduct at the cost of some burden on prosecuting authorities in criminal cases arising out of or related to an accident covered by that section rather than avoid that burden at the cost of significantly frustrating the important noncriminal objective of the legislation. Imposition of use-restrictions in the present case will not preclude the Legislature from overriding our decision if it wishes by simply enacting legislation declaring that information derived from disclosures required by section 20002, subdivision (a), may be used in criminal prosecutions, in which case the privilege could be claimed in appropriate situations.

“There is another significant distinction between the circumstances in Marchetti and the circumstances in the present case. In Marchetti the imposition of use-restrictions on information obtained as a result of compliance with the federal wagering tax would have had a much more sweeping effect on state law enforcement than would the imposition of such restrictions here. It appears that most — perhaps almost all — violators of state criminal prohibitions against wagering and related activities are *446subject to the disclosure requirements of the federal wagering tax. (Marchetti v. United States, supra, 390 U. S. 39, 44-46, fns. 5-6.) Thus, the imposition of use-restrictions in order to permit Congress to compel all wagerers to comply with the wagering tax law would have meant that in almost all state prosecutions for wagering or related illegal activities the state would be forced, if the defendant proved compliance with the federal law, to establish that its evidence was untainted. This situation might indeed seriously hamper such state prosecutions. By contrast, far from all criminal violations committed on the highways by drivers of motor vehicles involved property damage. The burden resulting from the imposition of use-restrictions in the latter situation will exist only in those instances where property damage occurs in the course or as a result of a criminal violation committed on the highways by a driver.

“We conclude that criminal prosecutions of drivers involved in accidents will not be unduly hampered by rules that prosecuting authorities may not use information divulged as a result of compliance with section 20002, subdivision (a), of the Vehicle Code or the fruits of such information and that in prosecutions of individuals who have complied with that section the state must establish that its evidence is not the fruit of such information.

“Since imposition in the present case of use-restrictions as described above will neither frustrate any apparent legislative purpose behind the enactment of section 20002 of the Vehicle Code nor unduly hamper criminal prosecutions of drivers involved in accidents; and since the imposition of such restrictions will not preclude the state Legislature from overriding our decision if it wishes, the reasons im*447pelling the United States Supreme Court to reject the ‘attractive and apparently practical’ suggestion of imposing restrictions in Marchetti v. United States, supra, 390 U. S. 39, 58, are absent in the present case, and we must, in order to fulfill our responsibility to protect the privilege against self-incrimination, hold that where compliance with section 20002 of the Vehicle Code would otherwise be excused by an assertion of the privilege, compliance is, as in other cases, mandatory and state prosecuting authorities are precluded from using the information disclosed as a result of compliance or its fruits in connection with any criminal prosecution related to the accident.” Byers v. Justice Court, supra, at 1055-1057, 458 P. 2d, at 476-477 (footnotes omitted).

It is readily apparent from the above passages that the California Supreme Court recognized, as of course it had to, see n. 4, supra, that imposition of a use restriction would significantly impair the State’s capacity to prosecute drivers whose illegal behavior caused accidents. But that court had decided that the Federal Constitution compelled the State, at the very least, to accept that burden on its prosecutorial efforts in such cases if it wished to pursue its nonprosecutorial goal through compelled self-reporting. Given the availability of the criminal sanction for cases where accidents do not occur the court concluded that interference with prosecutorial efforts in accident cases was not so important that it rendered the use restriction less palatable to the State than recognition of an outright privilege not to disclose. I fail to see how it “flouts” the State’s assessment of its own interests to remove the premise that federal law compels a sacrifice of criminal law enforcement where accidents are involved; I doubt that anyone would maintain that criminal law enforcement goals are not significantly served by imposition of criminal sanctions in the very *448cases where the feared results of dangerous driving have actually materialized. Of course, after the federal law premise has been removed, the State is free to conclude as a matter of state constitutional or legislative policy that continued imposition of use restrictions with respect to this category of cases would still be appropriate in light of the State’s own assessment of the relevant regulatory interests at stake and the personal values protected by the privilege against self-incrimination.

IV

Thus the public regulation of driving behavior through a pattern of laws which includes compelled self-reporting to ensure financial responsibility for accidents and criminal sanctions to deter dangerous driving entails genuine risks of self-incrimination from the driver’s point of view. The conclusion that the Fifth Amendment extends to this regulatory scheme will impair the capacity of the State to pursue these objectives simultaneously. For compelled self-reporting is a necessary part of an effective scheme of assuring personal financial responsibility for automobile accidents. Undoubtedly, it can be argued that self-reporting is at least as necessary to an effective scheme of criminal law enforcement in this area. The fair response to that latter contention may be that the purpose of the Fifth Amendment is to compel the State to opt for the less efficient methods of an “accusatorial” system. But see Schmerber v. California, 384 U. S. 757 (1966). But it would not follow that the constitutional values protected by the “accusatorial” system, see infra, at 450-451, are of such overriding significance that they compel substantial sacrifices in the efficient pursuit of other governmental objectives in all situations where the pursuit of those objectives requires the disclosure of information which will undoubtedly significantly aid in criminal law enforcement.

*449For while this Court’s Fifth Amendment precedents have instructed that the Fifth Amendment be given a construction “as broad as the mischief against which it seeks to guard,” Miranda v. Arizona, 384 U. S. 436, 459-460 (1966) (quoting from Counselman v. Hitchcock, 142 U. S. 547, 562 (1892)), and while the Court in Malloy v. Hogan, 378 U. S. 1 (1964), treated the privilege as one of those fundamental rights to be “selectively incorporated” into the Fourteenth Amendment, it is also true that the Court has recognized that the “scope of the privilege [does not coincide] with the complex of values it helps to protect.” Schmerber v. California, 384 U. S., at 762. And see Me. Justice Brennan’s concurring opinion in Marchetti, supra, and Grosso, supra, 390 U. S., at 72-73. In the Schmerber case the Court concluded that the impact of compelled disclosure of “non-testimonial” evidence on the values the privilege is designed to protect was insufficient to warrant a further restriction on the State’s enforcement of its criminal laws. And the Court in Schmerber explicitly declined reliance on the implication of a “testimonial” limitation to be found in the language of the Fifth Amendment. 384 U. S., at 761 n. 6.

The point I draw from the Schmerber approach to the privilege is that “[t]he Constitution contains no formulae with which we can calculate the areas within this 'full scope’ to which the privilege should extend, and the Court has therefore been obliged to fashion for itself standards for the application of the privilege. In federal cases stemming from Fifth Amendment claims, the Court has chiefly derived its standards from consideration of two factors: the history and purposes of the privilege, and the character and urgency of the other public interests involved. . . .” Spevack v. Klein, 385 U. S. 511, 522-523 (1967). (Harlan, J., dissenting.)

There are those, I suppose, who would put the “liberal construction” approach of cases like Miranda, and Boyd *450v. United States, 116 U. S. 616 (1886), side by side with the balancing approach of Schmerber and perceive nothing more subtle than a set of constructional antinomies to be utilized as convenient bootstraps to one result or another. But I perceive in these cases the essential tension that springs from the uncertain mandate which this provision of the Constitution gives to this Court.

This Court's cases attempting to capture the “purposes” or “policies” of the privilege demonstrate the uncertainty of that mandate. See Tehan v. Shott, 382 U. S. 406, 413-416 (1966); Murphy v. Waterfront Comm’n, 378 U. S., at 55; Miranda v. Arizona, 384 U. S., at 460; Boyd v. United States, supra. One commentator takes from these cases two basic themes: (1) the privilege is designed to secure among governmental officials the sort of respect for the integrity and worth of the individual citizen thought to flow from the commitment to an “accusatorial” as opposed to an “inquisitorial” criminal process; (2) the privilege is part of the “concern for individual privacy that has always been a fundamental tenet of the American value structure.” McKay, Self-Incrimination and the New Privacy, 1967 Sup. Ct. Rev. 193, 210. Certainly, in view of the extension of the privilege to witnesses in civil lawsuits, see McCarthy v. Arndstein, 266 U. S. 34 (1924)—a context in which, in most instances, information is sought by a private party wholly for purposes of resolving a private dispute — it is unlikely that the rationale of the privilege can be limited to preservation of official respect for the individual's integrity. Though the “privacy” rubric is not without its difficulties in the Fifth Amendment area,5 it does, I think, capture an important element of the concerns of the privilege, which accounts in part for our willingness to accept *451its reach beyond the context of the criminal investigation or trial. The premise of the criminal sanction — -and the disgrace that goes with it — is that it is more feared than the mere censure of our fellow members of society; although communal living requires us to be willing to disclose much to the government and our fellow citizens about our private affairs — and although the fear of eventually having to disclose operates as an inhibiting factor on our personal lives — it still makes sense to think of the Fifth Amendment as intended at least in part to relieve us of the very particular fear arising from the imposition of criminal sanctions.

These values are implicated by governmental compulsion to disclose information about driving behavior as part of a regulatory scheme including criminal sanctions. The privacy interest is directly implicated, while the interest in preserving a commitment to the “accusatorial” system is implicated in the more attenuated sense that an officialdom which has available to it the benefits of a self-reporting scheme may be encouraged to rely upon that scheme for all governmental purposes. But, as I have argued, it is also true that, unlike the ordinary civil lawsuit context, special governmental interests in addition to the deterrence of antisocial behavior by use of criminal sanctions are affected by extension of the privilege to this regulatory context.6 If the privilege is extended to the *452circumstances of this case, it must, I think, be potentially available in every instance where the government relies on self-reporting. And the considerable risks to efficient government of a self-executing claim of privilege will require acceptance of, at the very least, a use restriction of unspecified dimensions. Technological progress creates an ever-expanding need for governmental information about individuals. If the individual's ability in any particular case to perceive a genuine risk of self-incrimination is to be a sufficient condition for imposition of use restrictions on the government in all self-reporting contexts, then the privilege threatens the capacity of the government to respond to societal needs with a realistic mixture of criminal sanctions and other regulatory devices.7 To the extent that our Marchetti-Orosso line of *453. cases appears to suggest that the presence of perceivable risks of incrimination in and of itself justifies imposition of a use restriction on the information gained by the Government through compelled self-reporting, I think that line of cases should be explicitly limited by this Court.

V

I would not, however, overrule that line of cases. In each of those cases,8 the Government, relying on its taxing power, undertook to require the individual to focus attention directly on behavior which was immediately recognizable as criminal in virtually every State in the Union. Since compelled self-reporting is certainly essential to the taxing power, those cases must be taken to stand at least for the proposition that the Fifth Amendment requires some restriction on the efficiency with which government may seek to maximize both noncriminal objectives through self-reporting schemes and enforcement of criminal sanctions. If the technique of self-reporting as a means of achieving regulatory goals unrelated to deterrence of antisocial behavior through criminal sanctions is carried to an extreme, the “accusa-torial” system which the Fifth Amendment is supposed to secure can be reduced to mere ritual. And the risk that such a situation will materialize is not merely a function of the willingness of an ill-disposed officialdom to exploit the protective screen of ostensible legislative purpose to bypass the procedural limitations on governmental collection of information in the criminal process. The sweep of modern governmental regulation — and the dynamic growth of techniques for gathering and using information culled from individuals by force of criminal *454sanctions — could of course be thought to present a significant threat to the values considered to underpin the Fifth Amendment, quite apart from any supposed illegitimate motives that might not be cognizable under ordinary canons of judicial review. As uncertain as the constitutional mandate derived from this portion of the Bill of Rights may be, it is the task of this Court continually to seek that line of accommodation which will render this provision relevant to contemporary conditions.

' In other words, we must deal in degrees in this troublesome area. The question whether some sort of immunity is required as a condition of compelled self-reporting inescapably requires an evaluation of the assertedly noncriminal governmental purpose in securing the information, the necessity for self-reporting as a means of securing the information, and the nature of the disclosures required. See generally Mansfield, The Albertson Case: Conflict Between the Privilege Against Self-Incrimination and the Government’s Need for Information, 1966 Sup. Ct. Rev. 103, 128-160.

The statutory schemes involved in Marchetti and related cases, see n. 8, supra, focused almost exclusively on conduct which was criminal. As the opinion of The Chief Justice points out, the gambling activities involved in Marchetti and Grosso which gave rise to the obligation to report under that statutory scheme were illegal under federal law and the laws of almost every State in the Union. See Marchetti, supra, at 44-46. Indeed, Mr. Justice Brennan’s concurring opinion in Marchetti and Grosso, supra, at 74-75, concisely sets forth the precise degree of focus on criminal behavior as the predicate for state compulsion to report information:

“The Court’s opinions fully establish the statutory system’s impermissible invasions of the privilege. Indeed, 26 U. S. C. § 4401 should create substantial suspicion on privilege grounds simply because it is an *455excise tax upon persons ‘engaged in the business of accepting wagers’ or who conduct ‘any wagering pool or lottery.’ The persons affected by this language are a relatively small group, many of whom are engaged in activities made unlawful by state and federal statutes. But § 4401 is actually even more directly confined to that group. Section 4402 (1) exempts from the tax wagers placed with a parimutuel wagering enterprise ‘licensed under State law,’ and § 4421 defines ‘wager’ to exclude most forms of unorganized gambling such as dice and poker, and defines ‘lottery’ to exclude commonly played games such as bingo and drawings conducted by certain tax-exempt organizations. The effect of these exceptions is to limit the wagering excise tax under § 4401 almost exclusively to illegal, organized gambling.

“Moreover the code contemplates extensive record-keeping reporting by persons obligated to pay the tax. But these are records and reports which would incriminate overwhelmingly. Section 6011 (a) requires any person liable to pay a tax to file a return in accordance with the forms and regulations promulgated by the Secretary or his delegate. The regulations promulgating record-keeping requirements and the requirement that taxpayers make a monthly return on Form 730 . . . were therefore formulated pursuant to specific congressional authority. That the return is intended to be a part of the wagering tax obligation is clear from the face of the return itself. . . .”

Although compelled self-reporting is certainly essential to the taxing power, the decision to collect taxes through a special regulatory scheme which conditions the obligation to report on intent to commit a crime or the actual commission of a crime represents a determination to pursue noncriminal governmental purposes to the entire *456exclusion of the values protected by the Fifth Amendment. Cf. Grosso, supra, at 76 (Brennan, J., concurring). In a very real sense, compliance with the statutory requirements involved in Marchetti and Grosso, followed by use of the information in a prosecution, reduced the “accusatorial system” to the role of a merely ritualistic confirmation of the “conviction” secured through the exercise of the taxing power. Those statutory schemes are hardly distinguishable from a governmental scheme requiring robbers to register as such for purposes of paying an occupational tax and a tax on the proceeds of their crimes. Cf. my Brother Brennan’s opinion in the instant case, post, at 473.

In contrast, the “hit and run” statute in the present case predicates the duty to report on the occurrence of an event which cannot, without simply distorting the normal connotations of language, be characterized as “inherently suspect”; i. e., involvement in an automobile accident with property damage. And, having initially specified the regulated event — i. e., an automobile accident involving property damage — in the broadest terms possible consistent with the regulatory scheme’s concededly noncriminal purpose, the State has confined the portion of the scheme now before us, see n. 1 of The Chief Justice’s opinion, to the minimal level of disclosure of information consistent with the use of compelled self-reporting in the regulation of driving behavior. Since the State could — in the context of a regulatory scheme including an otherwise broad definition of the regulated event— achieve the same degree of focus on criminal conduct through detailed reporting requirements as was achieved in Marchetti and Grosso through the definition of the event triggering the reporting duties of the gambling tax scheme, the Court must take cognizance of the level of detail required in the reporting program as well as the *457circumstance giving rise to the duty to report; otherwise, the State, possessed as it is of increasingly sophisticated techniques of information gathering and storage, will, in the zealous pursuit of its noncriminal regulatory goals, reduce the “accusatorial system" which the Fifth Amendment is intended to secure to a hollow ritual.

California’s decision to compel Byers to stop after his accident and identify himself will not relieve the State of the duty to determine, entirely by virtue of its own investigation after the coerced stop, whether or not any aspect of Byer’s behavior was criminal. Nor will it relieve the State of the duty to determine whether the accident which Byers was forced to admit involvement in was proximately related to the aspect of his driving behavior thought to be criminal.9 In short, Byers having once focused attention on himself as an accident participant, the State must still bear the burden of making the main evidentiary case against Byers as a violator of *458§ 21750 of the California Vehicle Code.10 To characterize this burden as a merely ritualistic confirmation of the “conviction” secured through compliance with the reporting requirement in issue would be a gross distortion of reality; on the other hand, that characterization of the evidentiary burden remaining on the State and Federal Governments after compliance with the regulatory scheme involved in Marchetti and Grosso seems proper.

VI

Considering the noncriminal governmental purpose in securing the information, the necessity for self-reporting as a means of securing the information, and the nature of the disclosures involved, I cannot say that the purposes of the Fifth Amendment warrant imposition of a use restriction as a condition on the enforcement of this statute. To hold otherwise would, it seems to me, embark us on uncharted and treacherous seas. There will undoubtedly be other statutory schemes utilizing compelled self-reporting and implicating both permissible state objectives and the values of the Fifth Amendment which will render this determination more difficult to make. A determination of the status of those regulatory schemes must, of course, await a proper case.

On the premises set forth in this opinion, I concur in the judgment of the Court.11

with whom Mr. Justice Douglas and Mr. Justice Brennan join, dissenting.

Since the days of Chief Justice John Marshall this Court has been steadfastly committed to the principle that the Fifth Amendment’s prohibition against compulsory self-incrimination forbids the Federal Government to compel a person to supply information which can be used as a “link in the chain of testimony” needed to prosecute him for a crime. United States v. Burr, 25 F. Cas. 38, 40 (No. 14,692e) (CC Va. 1807). It is now established that the Fourteenth Amendment makes that provision of the Fifth Amendment applicable to the States. Malloy v. Hogan, 378 U. S. 1 (1964). The plurality opinion, if agreed to by a majority of the Court, would practically wipe out the Fifth Amendment’s protection against compelled self-incrimination. This protective constitutional safeguard against arbitrary government was first most clearly' declared by Chief Justice Marshall in the trial of Aaron Burr in 1807. United States v. Burr, supra. In erasing this principle from the Constitution the plurality opinion retreats from a cherished guarantee of liberty fashioned by James Madison and the other founders of what they proudly proclaimed to be our free government. One need only read with care the past cases cited in today’s opinions to understand the shrinking process to which the Court today subjects a vital safeguard of our Bill of Rights.

*460The plurality opinion labors unsuccessfully to distinguish this case from our previous holdings enforcing the Fifth Amendment guarantee against compelled self-incrimination. See, e. g., Albertson v. SACB, 382 U. S. 70 (1965); Grosso v. United States, 390 U. S. 62 (1968); Marchetti v. United States, 390 U. S. 39 (1968); Leary v. United States, 395 U. S. 6 (1969); Haynes v. United States, 390 U. S. 85 (1968). The plurality opinion, ante, at 431, appears to suggest that those previous cases are not controlling because respondent Byers would not have subjected himself to a “substantial risk of self-incrimination” by stopping after the accident and providing his name and address as required by California law. See California Vehicle Code § 20002 (a)(1) (Supp. 1971). This suggestion can hardly be taken seriously. A California driver involved in an accident causing property damage is in fact very likely to have violated one of the hundreds of state criminal statutes regulating automobiles which constitute most of two volumes of the California Code.1 More important, the particular facts of this case demonstrate that Byers would have subjected himself to a “substantial risk of self-incrimination,” ante, at 431, had he given his name and address at the scene of the accident. He has now been charged not only with failing to give his name but also with passing without maintaining a safe distance as prohibited by California Vehicle Code § 21750 (Supp. 1971). It is stipulated that the allegedly improper passing caused the accident from which Byers left without stating his name and address. In a prosecution under § 21750, the State will be required to prove that Byers was the driver who passed without maintaining a safe distance. Thus, if Byers had stopped and provided his name and address as the driver involved in the accident, the State could have used that information to *461establish an essential element of the crime under § 21750. It seems absolutely fanciful to suggest that he would not have faced a “substantial risk of self-incrimination,” ante, at 431, by complying with the disclosure statute.

The plurality opinion also seeks to distinguish this case from our previous decisions on the ground that § 20002 (a)(1) requires disclosure in an area not “permeated with criminal statutes” and because it is not aimed at a “highly selective group inherently suspect of criminal activities.” Ante, at 430. Of course, these suggestions ignore the fact that this particular respondent would have run a serious risk of self-incrimination by complying with the disclosure statute. Furthermore, it is hardly accurate to suggest that the activity of driving an automobile in California is not “an area permeated with criminal statutes.” Ibid. And it is unhelpful to say the statute is not aimed at an “inherently suspect” group because it applies to “all persons who drive automobiles in California.” Ibid. The compelled disclosure is required of all persons who drive automobiles in California who are involved in accidents causing property damage.2 If this group is not “suspect” of illegal activities, it is difficult to find such a group.

The plurality opinion purports to rely on United States v. Sullivan, 274 U. S. 259 (1927), to support its result. But Sullivan held only that a taxpayer could not defeat a *462prosecution for failure to file a tax return on the grounds that his income was illegally obtained. The Court there suggested that the defendant could lawfully have refused to answer particular questions on the return if they tended to incriminate him.3 Here, unlike Sullivan, the only information that the State requires Byers to disclose greatly enhances the probability of conviction for crime. As I have pointed out, if Byers had stopped and identified himself as the driver of the car in the accident, he would have handed the State an admission to use against him at trial on a charge of failing to maintain a safe distance while passing. Thus, Byers’ failure to stop is analogous to a refusal to answer a particular incriminating question on a tax return, an act protected by the Fifth Amendment under this Court’s decision in Sullivan. Cf. Marchetti v. United States, supra; Grosso v. United States, supra.

I also find unacceptable the alternative holding that the California statute is valid because the disclosures it requires are not “testimonial” (whatever that term may mean). Ante, at 431. Even assuming that the Fifth Amendment prohibits the State only from compelling a man to produce “testimonial” evidence against himself, the California requirement here is still unconstitutional. What evidence can possibly be more “testimonial” than a man’s own statement that he is a person who has just been involved in an automobile accident inflicting prop*463erty damage? Neither United States v. Wade, 388 U. S. 218 (1967), nor any other case of this Court has ever held that the State may convict a man by compelling him to admit that he is guilty of conduct constituting an element of a crime. Cf. United States v. Burr, supra. Yet the plurality opinion apparently approves precisely that result.

My Brother Harlan’s opinion makes it clear that today the Court “balances” the importance of a defendant’s Fifth Amendement right not to be forced to help convict himself against the government’s interest in forcing him to do so. As in previous decisions, this balancing inevitably results in the dilution of constitutional guarantees. See, e. g., Konigsberg v. State Bar, 366 U. S. 36, 56 (1961) (Black, J., dissenting). By my Brother Harlan’s reasoning it appears that the scope of the Fifth Amendment’s protection will now depend on what value a majority of nine Justices chooses to place on this explicit constitutional guarantee as opposed to the government’s interest in convicting a man by compelling self-incriminating testimony. ' In my view, vesting such power in judges to water down constitutional rights does indeed “embark us” on Brother Harlan’s “uncharted and treacherous seas.” Ante, at 458.

I can only assume that the unarticulated premise of the decision is that there is so much crime abroad in this country at present that Bill of Rights’ safeguards against arbitrary government must not be completely enforced. I can agree that there is too much crime in the land for us to treat criminals with favor. But I can never agree that we should depart in the slightest way from the Bill of Rights’ guarantees that give this country its high place among the free nations of the world. If we affirmed the State Supreme Court, California could still require persons involved in accidents to stop and give their names and addresses. The State *464would only be denied the power to violate the Fifth Amendment by using the fruits of such compelled testimony against them in criminal proceedings. Instead of criticizing the Supreme Court of California for its rigid protections of individual liberty, I would without more ado affirm its judgment.

with whom Mr. Justice Douglas and Mr. Justice Marshall join, dissenting.

Although I have joined my Brother Black’s opinion in this case, the importance of the issues involved and the wide range covered by the two opinions supporting the Court’s judgment in this case make further comment desirable. Put briefly, one of the primary flaws of the plurality opinion is that it bears so little relationship to the case before us. Notwithstanding the fact that respondent was charged both with a violation of the California Vehicle Code which resulted in an accident, and with failing to report the accident and its surrounding circumstances as required by the statute under review, the plurality concludes, contrary to all three California courts below, that respondent was faced with no substantial hazard of self-incrimination under California law. My Brother Harlan, by contrast, recognizes the inadequacy of any such conclusion. In his view, our task is to make the Bill of Rights “relevant to contemporary conditions” by simply not applying its provisions when we think the Constitution errs. Ante, at 454. In the context of the present case, this appears to mean that current technological progress enabling the Government more easily to use an individual’s compelled statements against him in a criminal prosecution should be matched by frank judicial contraction of the privilege against self-incrimination lest the Government be hindered in using modern technology further to reduce individual privacy. *465Needless to say, neither of these approaches is consistent with the Constitution.

I

This case arises from an attempt by the State of California to punish an assertion of the Fifth Amendment’s privilege against self-incrimination. Respondent Byers was charged with a statutory duty to report his involvement as a driver in an auto accident involving property damage. This he refused to do, and California seeks to impose criminal punishment for his refusal. Unlike the plurality, I believe that analysis of the question whether California may do so is inevitably tied to the circumstances of this case. I therefore turn to the record.

Respondent was initially charged in Justice Court with two violations of California law. The criminal complaint alleged, first, that he violated California Vehicle Code §21750 (Supp. 1971) by improper passing; and second, that he was involved in an accident causing property damage and failed to report his name, address, and the circumstances of the accident to the other driver involved and the California Highway Patrol. California Vehicle Code § 20002, as amended by Cal. Laws 1965, c. 872.1 After a demurrer to the complaint was rejected, respondent sought a writ of prohibition to restrain prosecution of the second charge of the complaint.

The California Superior Court dealt with the statutory reporting requirement only as applied to respondent. It found as a fact that the alleged improper passing with which respondent was charged caused the accident that respondent was charged with failing to report. App. 49. The court found it “hard to imagine a more damaging *466link in the chain in a prosecution under Vehicle Code Section 21750 than that which establishes that the defendant was driving the vehicle involved.” App. 42. Since on these facts it was “obvious,” App. 44, that respondent faced a substantial hazard of self-incrimination if he reported that he was the driver of one of the automobiles involved in the accident, the Superior Court issued the writ to restrain prosecution for failure to make the report.

The California Court of Appeal also dealt with the statute only as applied to respondent. Like the Superior Court, it found it “difficult to imagine a more damaging link in the chain of prosecution” than the requirement of § 20002 that respondent inform the police that he was the driver of one of the vehicles involved in the accident. As the Court of Appeal put the matter, “To compel [respondent] to comply with [ § 20002] and thus to admit a fact essential to his conviction of a violation of section 21750 is to compel him to give a testimonial declaration that falls directly within the scope of the constitutional privilege.” 71 Cal. Rptr. 609, 612 (Ct. App. 1968). It concluded, however, that respondent could be criminally punished for his failure to report because the Fifth Amendment prohibited the use of “admissions and statements made in compliance with” § 20002 “in a prosecution for a criminal offense arising out of the same course of conduct.” Ibid. It reversed the Superior Court’s grant of the writ of prohibition.2

Finally, the California Supreme Court likewise dealt with the statute in the context of its application to respondent. It first identified the “crucial inquiry in determining the applicability of the privilege” as

“whether the individual seeking to avoid disclosure faces ‘substantial hazards of self-incrimination’ because in his particular case there is a substantial *467likelihood that information disclosed by him in compliance with the statute could by itself or in conjunction with other evidence be used to secure his conviction of a criminal offense.” 71 Cal. 2d 1039, 1043, 458 P. 2d 465, 468 (1969) (emphasis added).

Second, it construed the California statute in question to require a person to whom it applies to report not merely his name and address, but also that he was the driver of an automobile involved in a particular accident. 71 Cal. 2d, at 1045, 458 P. 2d, at 470. It held the privilege against self-incrimination applicable to the “driver of a motor vehicle involved in an accident [who] is confronted with [the] statutory requirement . . . [and who] reasonably believes that compliance with the statute will result in self-incrimination.” Id., at 1047, 458 P. 2d, at 471. It agreed with the two courts below that respondent, at the time of the accident, “had reasonable ground to apprehend that if he stopped to identify himself as required ... he would confront a substantial hazard of self-incrimination.” Id., at 1057, 458 P. 2d, at 477. It agreed with the Court of Appeal, however, that the statute could and should be limited by restricting the use of information acquired pursuant to the statutory compulsion in circumstances where the particular individual reporting could demonstrate a substantial risk of self-incrimination.3 Id., at 1050-1056, 458 P. 2d, at 472-477. Contrary to the Court of Appeal, however, the California Supreme Court felt that it would be unfair to punish respondent when he could have had no knowledge that use restrictions would be applied by *468the courts. Accordingly, it affirmed the Superior Court. Id., at 1057-1058, 458 P. 2d, at 477-478. The two dissenting justices took issue with the majority only over the question whether respondent’s punishment would be unfair. Id., at 1059-1060, 458 P. 2d, at 479.

II

The plurality opinion, unfortunately, bears little resemblance either to the facts of the case before us or to the law upon which it relies. Contrary to the plurality opinion, I do not believe that we are called upon to determine the broad and abstract question “whether the constitutional privilege against compulsory self-incrimination is infringed by California’s so-called ‘hit and run’ statute4 which requires the driver of a motor vehicle involved in an accident to stop at the scene and give his name and address.” Ante, at 425. I believe we are called upon to decide the question presented by this case, which is whether California may punish respondent, over his claim of the privilege against self-incrimination, for failing to comply with the statutory requirement that he report his name and address, and the fact that he was the driver of an automobile involved in this particular accident. Despite the plurality’s assurance that its “judicial scrutiny is ... a close one,” ante, at 427, I believe that in the course of explaining its own views regarding “disclosures with respect to automobile accidents” in general, ante, at 431, the plurality has lost sight of the record before us. See ante, at 427, 430-431. Instead of dealing with the “underlying constitutional issues in clean-cut and concrete form,” Rescue Army v. Municipal *469Court, 331 U. S. 549, 584 (1947), the plurality seeks a broad general formula to resolve the tensions “between the State’s demand for disclosures and . . . the right against self-incrimination.” Ante, at 427. But only rivers of confusion can flow from a lake of generalities. Cf. the opinion of my Brother Black, ante, at 460-461.

Much of the plurality’s confusion appears to stem from its misunderstanding of the language, embodied in several of this Court’s opinions, regarding questions “directed at a highly selective group inherently suspect of criminal activities.” Albertson v. SACB, 382 U. S. 70, 79, (1965); see Marchetti v. United States, 390 U. S. 39, 47, 57 (1968). The plurality seems to believe that membership in such a suspect group is somehow an indispensable foundation for any Fifth Amendment claim. See ante, at 429-431. Of course, this is not so, unless the plurality is now prepared to assume that McCarthy v. Arndstein, 266 U. S. 34 (1924), Counselman v. Hitchcock, 142 U. S. 547 (1892), Garrity v. New Jersey, 385 U. S. 493 (1967), and Spevack v. Klein, 385 U. S. 511 (1967), were based, respectively, upon the un-articulated premises that bankrupts, businessmen, policemen, and lawyers are all “group [s] inherently suspect of criminal activities.” Instead, in the words of the California Supreme Court, “in each case the crime-directed character of the registration requirement was . . . important only insofar as it supported the claims of the specific petitioners that they faced 'substantial hazards of self-incrimination’ justifying invocation of the privilege.” 71 Cal. 2d, at 1043, 458 P. 2d, at 468. That this is so is evident from our emphasis in Marchetti that “we do not hold that these wagering tax provisions are as such constitutionally impermissible .... If, in different circumstances, a taxpayer is not confronted by substantial hazards of self-incrimination . . . nothing we decide today would shield him from the *470various penalties prescribed by the wagering tax statutes/’ 390 U. S., at 61. The point is that in both Albertson and Marchetti, petitioners arrived in this Court accompanied by a record showing only that they had failed to register, respectively, as Communists and as a gambler, and that, in fact, they were such. Since neither of these facts was necessarily criminal, we had to determine whether the petitioners faced “real and appreciable” or merely “imaginary and unsubstantial” 5 hazards when they refused to register. That the petitioners belonged in each case to an inherently suspect group was relevant to that question, and that alone. By contrast, in the present case we are dealing with a record which demonstrates, as found by all three courts below, that respondent was charged by California both with illegal passing which resulted in an accident, and with failing to report himself as one of the drivers involved in that accident. It is hard to imagine a record demonstrating a more substantial hazard of self-incrimination than this. Yet the plurality somehow concludes that respondent did not face the “substantial risk of self-incrimination involved in Marchetti.” 6 Ante, at 431.

*471The plurality opinion also places great reliance upon United States v. Sullivan, 274 U. S. 259 (1927). I had understood that case to stand for the proposition that one who desired to raise a Fifth Amendment claim to protect his refusal to provide information required by the Government on a tax return should make specific objection to the particular question on the return. Sullivan’s sole objection was to disclosing the amount of his income, on the ground that it had been made in crime; he did not claim to be entitled to refuse to disclose his name and address. The Court suggested, although it did not decide, that it would in a proper case reject the claim as to amount of income.7 It may be that Sullivan also stands for the proposition that an individual may not refuse, on Fifth Amendment grounds, to file a return disclosing his name and address, and by implication disclosing that he has earned some income during the previous year.8 But that question was not raised in Sullivan, and the Court explicitly noted that it was not *472called upon to decide what information could be withheld; certainly I would expect this Court to hesitate before affirming the conviction of a fugitive from justice for filing a tax return which omitted his address. However that may be, I am frankly unable to understand just what the plurality thinks that Sullivan stands for. Rather than pursue the matter further, I simply note below those portions of the Sullivan opinion quoted, paraphrased, or omitted in the plurality opinion. Portions there quoted are in roman type; portions there paraphrased are enclosed in brackets; portions there omitted are in italics.

“As the defendant's income was taxed, the statute of course required a return. ... In the decision that this was contrary to the Constitution we are of opinion that the protection of the Fifth Amendment was pressed too far. If the form of return provided called for answers that the defendant was privileged from making he could have raised the objection in the return, but could not on that account refuse to make any return at all. We are not called on to decide what, if anything, he might have withheld. ... [It would be] an extreme if not an extravagant application of the Fifth Amendment [to say that it authorized a man to refuse to state the amount of his income because it had been made in crime.] But if the defendant desired to test that or any other point he should have tested it in the return so that it could be passed upon.” United States v. Sullivan, 274 U. S., at 263-264.

Cf. the plurality opinion, ante, at 428-429, 433-434.

I find even less persuasive the plurality’s alternative suggestion, see ante, at 431-434, that the California statute involved here does not require individuals to “provide the State with ‘evidence of a testimonial or communicative nature’ within the meaning of the Con*473stitution.” Ante, at 432. To begin with, the plurality opinion states that “[cjompliance with §20002 (a)(1) requires two things: first, a driver involved in an accident is required to stop at the scene; second, he is required to give his name and address;” it later suggests that, conceivably, “it [could] be . . . inferred” that such a driver “was indeed the operator of an 'accident involved’ vehicle.” Ante, at 431, 433. But, the plurality opinion continues, United States v. Wade, 388 U. S. 218, 221-223 (1967), rejects the notion that “such inferences are communicative or testimonial.” Ante, at 433. Putting aside the plurality’s misreading of Wade, adequately dealt with by my Brother Harlan, ante, at 435-436, the initial problem with the plurality opinion is that it adopts a construction of the California statute that was explicitly rejected by the California Supreme Court. That court specifically stated that the statute involved here “requires drivers involved in accidents to identify themselves as involved drivers.” 71 Cal. 2d, at 1045, 458 P. 2d, at 470 (emphasis added in part). We have no license to overrule the California Supreme Court on a question of the construction of a California statute. Even if we did, however, I would still not be persuaded by the plurality’s reasoning that since “[ disclosure of name and address is an essentially neutral act,” ante, at 432, any inferences which may be drawn from that disclosure are not “communicative or testimonial” in nature. Ante, at 432, 433. Apparently the plurality believes that a statute requiring all robbers to stop and leave their names and addresses with their victims would not involve the compulsion of “communicative or testimonial” evidence.

Ill

Similarly, I do not believe that the force of my Brother Black’s reasoning may be avoided by my Brother Har*474lan’s approach. He quite candidly admits that our prior cases compel the conclusion that respondent was entitled to rely on the privilege against self-incrimination as a defense to prosecution for failure to stop and report his involvement in an accident. Ante, at 438-439. He would simply limit those cases because he believes that technological progress has made the privilege against self-incrimination a “threat” to “realistic” government that we can no longer afford.9 To the extent that this argument calls for refutation, it is adequately disposed of in Mr. Justice Brandéis’ dissenting opinion in Olmstead v. United States, 277 U. S. 438, 472-477, 479 (1928). Our society is not endangered by the Fifth Amendment. “The dangers of which we must really beware are . . . that we shall fall prey to the idea that in order to preserve our free society some of the liberties of the individual must be curtailed, at least temporarily. How wrong that kind of a program would be is surely evident from the mere statement of the proposition.” J. Harlan, Live and Let Live, in The Evolution of a Judicial Philosophy 285, 288 (D. Shapiro ed. 1969).

In any event my Brother Harlan’s opinion is consistent neither with the present record nor its own premises. As to the first, my Brother Harlan appears to believe that the imposition of use restrictions on the *475present statute would threaten the capacity of California “to respond to societal needs with a realistic mixture of criminal sanctions and other regulatory devices.” Ante, at 452. If so, this threat passed unperceived by the California Supreme Court: that court stated that its imposition of a use restriction “will neither frustrate any apparent significant legislative purpose nor unduly hamper criminal prosecutions of drivers involved in accidents resulting in damage to the property of others.” 71 Cal. 2d, at 1054, 458 P. 2d, at 475.10 It seems to have passed unnoticed by the California Legislature as well. The present statute applies to drivers involved in accidents causing property damage. California Vehicle Code § 20012 (Supp. 1971) requires similar, albeit more detailed, reports from drivers involved in accidents resulting in either personal injury or death. Yet the California Legislature itself imposed use restrictions upon the use of such reports. Ibid.11 It is one thing to respect a State’s assertion that imposition of a particular requirement will unduly hamper a legitimate state *476interest. It is quite another to flout the conclusion of the State’s Supreme Court — and, so far as appears, of the state legislature as well — that imposition of a particular requirement is not at all inconsistent with the asserted state interests.

Moreover, I think my Brother Harlan’s opinion falls on its own premises. For he recognizes, and apparently would follow, our cases holding that the privilege against self-incrimination may be claimed by a witness in a noncriminal proceeding who is asked to give testimony that might indicate his commission of crime. E. g., Counselman v. Hitchcock, 142 U. S., at 562; McCarthy v. Arndstein, 266 U. S. 34 (1924); Hutcheson v. United States, 369 U. S. 599 (1962) (Harlan, J.). See ante, at 450-451. He appears to believe that these cases are different from the one before us, because they involve information “sought by a private party wholly for purposes of resolving a private dispute,” ante, at 450, where no “special governmental interests in addition to the deterrence of antisocial behavior by use of criminal sanctions are affected.” Ante, at 451. Yet this is precisely the case before us. For the only noncriminal interest that has ever been asserted to justify the California reporting statute at issue here is the State’s interest in providing information “sought by a private party wholly for purposes of resolving a private dispute.” Of course, state policy is exercised, in part, through the resolution of otherwise private disputes through the judicial process. But this is true of every civil case, whether it involves tort liability for negligent driving, the ability of private individuals to inherit from one another, Labine v. Vincent, 401 U. S. 532 (1971), or the right of private parties to dissolve a previous marriage, Boddie v. Connecticut, 401 U. S. 371 (1971) (Harlan, J.). To distinguish the ordinary “civil lawsuit context,” ante, at 451, from the civil lawsuit context in which the pres*477ent statute is involved is simply to indulge in the sort of “artificial, if not disingenuous judgments” against which my Brother Harlan’s opinion otherwise warns. Ante, at 442.

Finally, even if everything else in my Brother Harlan’s opinion be accepted, I cannot understand his concurrence in the judgment. For the California Supreme Court agreed with his conclusion that the privilege against self-incrimination does not provide a defense to an individual who fails to comply with the statutory reporting requirement. 71 Cal. 2d, at 1057, 458 P. 2d, at 477. But it nevertheless concluded that respondent should not be punished because it would be “unfair” to do so. 71 Cal. 2d, at 1058, 458 P. 2d, at 478. Although my Brother Harlan concludes that the Fifth Amendment does not excuse compliance with the California reporting requirements for reasons quite different from those relied upon by the California Supreme Court, the point is that both have reached the same conclusion. Of course, we have already held that due process is denied an individual if he is led to believe that the privilege against self-incrimination applies when he refuses to answer questions, and subsequently prosecuted on the grounds that it does not. Raley v. Ohio, 360 U. S. 423 (1959). One would assume that in such circumstances my Brother Harlan would, although for very different reasons, agree that the judgment of the California Supreme Court should be affirmed.

IV

Although, strictly speaking, the only question before us is whether respondent may be punished for failing to comply with the statutory requirement at issue,12 I am *478constrained to add that I cannot agree with the California Supreme Court’s conclusion that the-requirement may be enforced if the State is merely precluded from using the compelled evidence and its fruits in a criminal prosecution. When, as in the present case, the statute requires an individual to admit that he has engaged in conduct likely to be the subject of criminal punishment under the California traffic laws, the requirement in my view may be enforced only if those reporting their involvement in an accident pursuant to the statutory command are immune from prosecution under state law for traffic offenses arising out of the conduct involved in the accident. See Piccirillo v. New York, 400 U. S. 548, 550-551 (1971) (Douglas, J., dissenting); id., at 561—573 (Brennan, J., dissenting); Mackey v. United States, 401 U. S. 667, 702 (1971) (Brennan, J., concurring in judgment).