I dissent. This court has long concluded, in a variety of different contexts, that sex offender registration is a form of punishment. Today, the court concludes to the contrary, and therefore holds that such a requirement may be imposed retrospectively without violating the ex post facto clause of the federal and California Constitutions (U.S. Const., art. I, § 10; Cal. Const., art. I, § 9). Focusing on the United States Supreme Court’s definition of “punishment,” it concludes the sex offender registration requirement is nonpenal, essentially because the Legislature has not professed an intention to punish. But as the lead opinion itself states, “two factors appear important in each case: whether the Legislature intended the provision to constitute punishment and, if not, whether the provision is so punitive in nature or effect that it must be found to constitute punishment despite the Legislature’s contrary intent.” (Lead opn., ante, at p. 795, italics added.) The lead opinion,1 however, all but ignores the second factor. Accordingly, its analysis is one-sided and its conclusion is contrary to precedent and basic ex post facto principles.
At the outset, it is important to recognize the narrowness of the issue in this case. California has had a sex offender registration law since 1947. (Stats. 1947, ch. 1124, § 1, p. 2562.) For many years, those who have committed various forms of sexual assault, child molestation, and other such serious offenses have been subject to these registration laws. The Legislature may expand, within the boundaries of the constitutional prohibition against cruel and unusual punishment (U.S. Const, 8th Amend.; Cal. Const., art. I, § 17; In re Reed (1983) 33 Cal.3d 914 [191 Cal.Rptr. 658, 663 P.2d. 216] (Reed)), the scope of the registration laws. The question in this case is whether the registration requirement may be retroactively imposed, not on anyone who has committed any of the offenses long enumerated in Penal Code section 290,2 but rather on one who fits into a recently created catchall category of offenders. That category provides for mandatory registration “if *807the court finds at the time of conviction or sentencing that the person committed the offense as a result of sexual compulsion or for purposes of sexual gratification.” (§ 290, subd. (a)(2)(E).) Thus, this case will directly affect only a relatively small group of offenders, those who fit into the catchall category, and who committed their offenses before section 290, subdivision (a)(2)(E) was enacted but are sentenced after that enactment.
In Collins v. Youngblood (1990) 497 U.S. 37, 41 [110 S.Ct. 2715, 2718, 111 L.Ed.2d 30], the high court stated: “Although the Latin phrase 'expost facto’ literally encompasses any law passed ‘after the fact,’ it has long been recognized by this Court that the constitutional prohibition on ex post facto laws applies only to penal statutes which disadvantage the offender affected by them.” Accordingly, under the ex post facto clause “Legislatures may not retroactively alter the definition of crime or increase the punishment for criminal acts.” (Id., at p. 43 [110 S.Ct. at p. 2719], italics added.) As this court recently explained: “The primary purpose of the constitutional guarantee is to ensure that the consequences of a particular course of conduct can be meaningfully assessed in advance, without fear that the rules of criminality and punishment will later change.” (People v. Frazer (1999) 21 Cal.4th 737, 760 [88 Cal.Rptr.2d 312, 982 P.2d 180], italics in original.) The fair warning for the penal consequences of one’s conduct that the ex post facto clause requires is “fundamental to our concept of constitutional liberty.” (Marks v. United States (1977) 430 U.S. 188, 191 [97 S.Ct. 990, 993, 51 L.Ed.2d 260].)
It is evident that the imposition of a lifelong sex offender registration requirement changes the rules of punishment and precludes the meaningful assessment of “a particular course of conduct.” For this reason, we have consistently held that in order for a guilty plea to be deemed knowing and intelligent, a defendant, must be informed of any such registration requirements. (People v. McClellan (1993) 6 Cal.4th 367, 376 [24 Cal.Rptr.2d 739, 862 P.2d 739]; Bunnell v. Superior Court (1975) 13 Cal.3d 592, 605 [119 Cal.Rptr. 302, 531 P.2d 1086]; In re Birch (1973) 10 Cal.3d 314, 321-322 [110 Cal.Rptr. 212, 515 P.2d 12] (Birch).) As we explained: “Under Penal Code section 290, a person convicted of one of the enumerated offenses . . . must register for life with the police department in the city in which he lives. He must reregister whenever he moves and must report each change of address within 10 days.[3] Individuals convicted of one of the enumerated *808crimes have been deemed by the Legislature to have a propensity to commit such antisocial crimes in the future and thus are the subject of continual police surveillance. Whenever any sex crime occurs in his area, the registrant may very well be subjected to investigation. Although the stigma of a short jail sentence should eventually fade, the ignominious badge carried by the convicted sex offender can remain for a lifetime.” (Birch, supra, 10 Cal.3d at pp. 321-322.)
Thus in Birch, supra, 10 Cal.3d at page 322, we concluded that a defendant’s guilty plea without the assistance of counsel was invalid because “we cannot believe that he was aware that as a consequence of urinating in a parking lot at 1:30 in the morning he would be required to register as a sex offender” and that “[cjertainly counsel would have advised him of this grave and direct consequence of his guilty plea.” (Italics in original.) Our conclusion in Birch and the other cases cited above was based on the de facto recognition of sex offender registration as a form of punishment. How can it be that we do not regard a defendant entering a plea agreement to have done so knowingly and intelligently if he is not notified of the sex offender registration requirement, but we do not consider advance notice of these requirements necessary for a “meaningful assessment” of “a particular course of conduct” under the ex post facto clause of the United States and California Constitutions?
In Kelly v. Municipal Court (1958) 160 Cal.App.2d 38 [324 P.2d 990] (Kelly), the court considered whether the sex offender registration requirement was one of the “penalties and disabilities resulting from the offense of which [a defendant] has been convicted” that may be discharged upon fulfillment of the conditions of probation under section 1203.4. In concluding affirmatively, the court stated: “The duty to reregister upon changing one’s place of address is a continuing duty, a burden which the convicted person carries with him until his dying day. Being thus severely limited in his freedom of movement and continuously under police surveillance, all stemming from the conviction which has been set aside, the conclusion seems irresistible that this registration requirement is one of the ‘penalties and disabilities resulting from the offense or crime of which he has been convicted’ . . . .”4 (160 Cal.App.2d at p. 41.)
Reed, supra, 33 Cal.3d 914, contains the most extensive discussion of the question of the punitive nature of sex offender registration and for that *809reason will be discussed at length. In that case, we held that the imposition of sex offender registration on a misdemeanant convicted under section 647, subdivision (a) for soliciting sex in a public restroom violates the cruel or unusual punishment clause of the California Constitution, article I, section 17. Before considering whether the requirement was “cruel or unusual,” we first had to answer the question whether the registration requirement was “punishment.” (33 Cal.3d at p. 919.)
In addressing this question, the Reed court cited the test set forth in Kennedy v. Mendoza-Martinez (1963) 372 U.S. 144 [83 S.Ct. 554, 9 L.Ed.2d 644] (Mendoza-Martinez), for whether a nominally civil sanction was in fact punitive: “[T]he United States Supreme Court enumerated the following factors for consideration on this issue: ‘Whether the sanction involves an affirmative disability or restraint, whether it has historically been regarded as a punishment, whether it comes into play only on a finding of scienter, whether its operation will promote the traditional aims of punishment— retribution and deterrence, whether the behavior to which it applies is already a crime, whether an alternative purpose to which it may rationally be connected is assignable for it, and whether it appears excessive in relation to the alternative purpose assigned are all relevant to the inquiry, and may often point in differing directions.’ ” (Reed, supra, 33 Cal.3d at p. 920.)
Applying this test, the Reed court first determined, “that such registration is an ‘affirmative disability or restraint,’ ” citing language in Kelly and Birch quoted above regarding the onerous nature of the requirement. (Reed, supra, 33 Cal.3d at p. 920.) The Reed court added: “Similarly, albeit in a different context, Justice Kaus has characterized the sex offender registration requirement as follows: ‘Apart from the bother and loss of privacy which mere registration entails, the “ready availability” to the police, if it serves its purpose, presumably means a series of command performances at- lineups.’ [Citation.] Needless to emphasize, law enforcement ‘command performances’ involve compulsion and restraint.” (Ibid., fn. omitted.)
Continuing our consideration of the Mendoza-Martinez factors, we stated: “The fact that sex offender registration may not have ‘historically been regarded as punishment’ is not dispositive. The Mendoza-Martinez opinion sets out a number of relevant considerations, not a checklist of absolute requirements. Furthermore, in Trop v. Dulles (1958) 356 U.S. 86 [2 L.Ed.2d 630, 78 S.Ct. 590], the Supreme Court invalidated the penalty of denationalization imposed for wartime desertion. The court denied that the government has ‘a license ... to devise any punishment short of death within the limits of its imagination. . . . [H] . . . Fines, imprisonment, and even execution may be imposed depending upon the enormity of the crime, but *810any technique outside the bounds of these traditional penalties is constitutionally suspect.’ [Citation.]
“Similarly in Weems v. United States (1910) 217 U.S. 349 [54 L.Ed. 793, 30 S.Ct. 544], the high court stressed the severity of the nonphysical punishment imposed. Though no doubt influenced in its decision by the harshness of the physical punishment, the court also addressed the postimprisonment penalties: ‘His prison bars and chains are removed, it is true, after twelve years, but he goes from them to a perpetual limitation of his liberty. He is forever kept under the shadow of his crime, forever kept within voice and view of the criminal magistrate, not being able to change his domicile without giving notice to the “authority immediately in charge of his surveillance,” and without permission in writing. He may not seek, even in other scenes and among other people, to retrieve his fall from rectitude. Even that hope is taken from him and he is subject to tormenting regulations that, if not so tangible as iron bars and stone walls, oppress as much by their continuity, and deprive of essential liberty.’ [Citation.]
“The third, fourth, and fifth factors enumerated in Mendoza-Martinez are readily satisfied here. We have interpreted section 647(a) to require lewd intent and specific sexual touching. [Citation.] Thus registration ‘comes into play only on a finding of scienter.’ (372 U.S. at p. 168 [9 L.Ed.2d at p. 661].) Whether or not recidivism is in fact a problem with section 647(a) misdemeanants, the legislative intent was surely to deter recidivism by facilitating the apprehension of past offenders. (Barrows v. Municipal Court [(1970)] 1 Cal.3d [821,] 825-826 [83 Cal.Rptr. 819, 464 P.2d 483].) And, of course, ‘the conduct to which [registration] . . . applies is already a crime.’
“We may also consider, under Mendoza-Martinez, whether there exists ‘an alternative purpose’ to which the punishment ‘may rationally be connected.’ (372 U.S. at p. 168 [9 L.Ed.2d at p. 661].) Although the Legislature may reasonably have intended that sex offender registration serve as a law enforcement tool to facilitate criminal investigations, it is not clear that the measure is effective in practice. Moreover, the fact that a minimal or ‘rational’ basis may underlie the legislation is outweighed here by the fact that the penalty of registration is ‘excessive in relation to the alternative purpose assigned’ to it.” (Reed, supra, 33 Cal.3d at pp. 921-922, fn. omitted.)
For all of these reasons we concluded “that the sex offender registration compelled by section 290 is a form of punishment within the meaning of *811article I, section 17, of the [California] Constitution.” (Reed, supra, 33 Cal.3d at p. 322.)
The lead opinion attempts to" distinguish Reed. It points out that our conclusion in Reed that sex offender registration was excessive in relation to the alternative, nonpunitive purpose assigned to it was based on our doubt about whether sex offender registration was effective in practice, which in turn was based in part on concessions in the amicus curiae brief of the Los Angeles City Attorney. The city attorney had advised this court of his opinion that registration of misdemeanants overloaded computers with useless information and made it more difficult to identify more serious sex offenders. (Reed, supra, 33 Cal.3d at p. 922, fn. 7.) The lead opinion implies that the fact no comparable concession can be found in this case limits the applicability of Reed. But this point has little merit for two reasons: First, the primary focus of our opinion in Reed was on the punitive effect or impact of sex offender registration as well as the "fact that it is imposed only in connection with criminal conviction. As discussed, the lead opinion barely addresses the punitive effect prong of ex post facto analysis, focusing almost exclusively on legislative intent. Thus, even if the portion of the Reed opinion regarding the practical effectiveness of sex offender registration were discounted, Reed would still stand for the proposition that the registration requirement was punitive in its impact.
Second, the practical effectiveness of the registration requirement remains in doubt. Since registration is voluntary and police resources are limited, it is questionable whether those who are intent on repeating their offenses will comply with registration laws. Registration has therefore been criticized, as one law enforcement official put it, because it “just keeps honest offenders honest.” (Quoted in Comment, Examining Sex Offender Community Notification Laws (1995) 83 Cal.L.Rev. 885, 903.) This criticism is borne out by reports of widespread inaccuracy in the sex offender registration databases. (Id., at pp. 900-902 [recounting various reports of 80-90 percent inaccuracy in the addresses of registered sex offenders in various California locations].) Thus, although the emphasis in Reed was on the burden and stigma of registration and the connection between the registration requirement and criminal conviction, the fact that substantial questions remain about the practical effectiveness of this requirement also supports the continued viability of Reed’s conclusion that sex offender registration is in fact a punitive measure.5
*812The lead opinion also finds Reed inapposite because in that case “the central issue was whether the registration requirement was excessive or disproportionate to the severity of the offenses committed by the defendant” under the cruel or unusual punishment clause (lead opn., ante, at p. 798, italics in original), which is not at issue in this case. But as noted, in order for the Reed court to conclude that the registration requirement was an excessive or disproportionate punishment in relation to the severity of the offense, it first had to conclude, and did conclude, that the registration requirement was a form of punishment. (Reed, supra, 33 Cal.3d at p. 922.) This foundational conclusion cannot be squared with the court’s holding in the present case.
The lead opinion also contends that “reexamination of the decision in Reed in light of . . . more recent cases,” leads to the conclusion that Reed should be disapproved at least in part. (Lead opn., ante, at p. 798.) The two cases cited by way of example, People v. McVickers (1992) 4 Cal.4th 81 [13 Cal.Rptr.2d 850, 840 P.2d 955] and Kansas v. Hendricks (1997) 521 U.S. 346 [117 S.Ct. 2072, 138 L.Ed.2d 501] (Hendricks), do not support the lead’s position. In McVickers, we upheld against an ex post facto challenge a statute, section 1202.1, that requires every person convicted of various sexual offenses to submit to a blood test for acquired immune deficiency syndrome (AIDS). The lead opinion emphasizes our conclusion in that case that AIDS testing under these circumstances “has a legitimate nonpunitive . . . purpose” (McVickers, supra, 4 Cal.4th at p. 89), and that the sex offender registration statute has such a purpose. But just as central to our holding in McVickers was the conclusion that the burden imposed upon the defendant was minimal. “It is well settled that the physically intrusive effects of drawing blood for a test are slight. ‘The blood test procedure has become routine in our everyday life.’ [Citation.] . . . ffl The strictly limited disclosure provisions of the statute likewise do not have the effect of punishment. Defendant’s fears of social ostracism if he tests positive for the AIDS virus and the result of his test becomes known are unwarranted. Until such time as he becomes a repeat offender, the test result will be disclosed to no one other than defendant and his attorney. It will be recorded at the state’s Department of Justice but only for use by the prosecutor in connection with a future criminal offense.” (Id. at p. 88.) In contrast to this onetime minimal intrusion of drawing blood and being tested, confidentially, for *813AIDS, a defendant subject to section 290 bears the lifetime burden of informing police of his presence and being subject to continuing police surveillance.
In Hendricks, the court considered a statute that authorized the civil commitment of a person convicted of a “sexually violent offense,” when it is proven beyond a reasonable doubt that the person “ ‘suffers from a mental abnormality or personality disorder which makes a person likely to engage in the predatory acts of [such] sexual violence.’ ” (521 U.S. at p. 352 [117 S.Ct. at p. 2077].) Civil commitment could also be brought against those charged with a sexually violent offense who have been found incompetent to stand trial, or have been found not guilty by reason of insanity or some other mental disease or defect. (Ibid.) The commitment is reviewed annually to determine if continued detention is warranted. (Id. at p. 353 [117 S.Ct. at pp. 2077-2078].) The court reasoned that this statutory scheme did not differ appreciably from other civil commitment statutes that had long been regarded as nonpunitive. (Id. at pp. 356-357 [117 S.Ct. at pp. 2079-2080].) The court emphasized the disjunction between past criminal liability and commitment under the statute. Prior criminal conduct “is used solely for evidentiary purposes, either to demonstrate that a ‘mental abnormality’ exists or to support a finding of future dangerousness .... In addition, the Kansas Act does not make a criminal conviction a prerequisite for commitment—persons absolved of criminal responsibility may nonetheless be subject to confinement under the Act.” (Id. at p. 362 [117 S.Ct. at p. 2082].) Furthermore, citing Mendoza-Martinez, the court found the existence of “a scienter requirement is customarily an important element in distinguishing criminal from civil statutes,” and concluded that in the case of the Kansas statute “no finding of scienter is required to commit an individual who is found to be a sexually violent predator; instead, the commitment determination is based on a ‘mental abnormality’ or ‘personality disorder’ rather than on one’s criminal intent.” (Ibid.)
By contrast, the sex offender registration statute applies only to those convicted of criminal offenses. Moreover, in the present case it is imposed on one who commits an offense under “sexual compulsion or for purposes of sexual gratification,” and is therefore contingent on a finding of scienter. Thus, although civil commitment can certainly be as disabling a form of restraint as incarceration, its disjunction from criminal liability in Hendricks makes that case inapposite to the present one.
In short, I am not persuaded that our consistent conclusion in Reed, McClellan, Birch, and other cases that the lengthy sex offender registration requirement is a penal sanction, was incorrect. Therefore, to impose such a *814sanction retroactively violates that most basic guarantee found in the ex post facto clause that a person be notified in advance of the penal consequences of his actions. For this reason, I would affirm the judgment of the Court of Appeal.
Reference to the “lead opinion” includes the concurring and dissenting opinion to the extent it agrees with the lead opinion.
All statutory references are to the Penal Code unless otherwise indicated.
Since our decision in Birch, section 290 has been amended to impose on individuals subject to sex offender registration the additional burden of annual reregistration. (§ 290, subd. (a)(1)(C); Stats. 1994, ch. 865, § 1.) This is to say nothing of the recently added public notification provisions of section 290, subdivisions (m) and (n) and section 290.4, which, as the lead opinion acknowledges, are not at issue in this case. (Lead o'pn., ante, at p. 796, fn. 6.)
The Legislature later added section 290.1 to provide that the registration requirement may not be discharged through the procedures of section 1203.4. (Stats. 1981, ch. 105, § 1, p. 794.) This legislative amendment does not alter the relevance for purposes of constitutional analysis of the Kelly court’s conclusion that this requirement is indeed a penalty or disability.
The Legislature has recently increased the penalty for failure to register in many cases from a misdemeanor to a felony with a maximum of 16 months in a state prison. (Stats. 1996, ch. 908, § 2.) These increased sanctions may improve compliance with the registration law, *812although recidivists still have a strong incentive to evade that law. But the fact that the failure of a sex offender to regularly report his whereabouts to the police now results in an additional felony conviction lends further support to the conclusion that, in terms of its effect, section 290 is indeed a punitive measure, i.e., the kind of consequence of criminal activity that, under the ex post facto clause, an individual should be notified of in advance in order to meaningfully assess his course of conduct.