We need look no further than the California Constitution to conclude that California’s mandatory continuing legal education (MCLE) program is unconstitutional because it arbitrarily exempts certain classes of active attorneys from its requirements while imposing those requirements on others.
Article I, section 7, subdivision (a), of the California Constitution provides: “A person may not be . . . denied equal protection of the laws . . . .” Subdivision (b) of the same section provides: “A citizen or class of citizens *661may not be granted privileges or immunities not granted on the same terms to all citizens.” Finally, article IV, section 16, subdivision (a), provides: “All laws of a general nature have uniform operation.” Significantly, the latter two provisions find no counterpart in the federal Constitution, and we have interpreted them as providing greater protection than the federal equal protection clause. Specifically, we have held that the state Constitution requires “ ‘a serious and genuine judicial inquiry into the correspondence between [a legislative] classification and the legislative goals’ ” (Newland v. Board of Governors (1977) 19 Cal.3d 705, 711 [139 Cal.Rptr. 620, 566 P.2d 254], quoting Dorrough v. Estelle (5th Cir. 1974) 497 F.2d 1007, 1011) and prohibits legislative classifications, such as those at issue here, that are “grossly overinclusive” or “underinclusive.” (Brown v. Merlo (1973) 8 Cal.3d 855, 877 & fn. 17 [106 Cal.Rptr. 388, 506 P.2d 212, 66 A.L.R.3d 505]; see also State v. Mohi (Utah 1995) 901 P.2d 991, 997 [“Utah’s uniform operation of [the] laws provision establishes different requirements than does the federal Equal Protection Clause.”].) Nevertheless, the majority does not even bother analyzing the requirements of the California Constitution independently from those of the federal Constitution. The majority should apply California’s equal protection standards, in which case it would conclude, as I do, that the exemptions at issue here are unconstitutional. Instead, it simply treats the California standards as if they did not exist.
I
Our state equal protection jurisprudence grew out of a recognition of the inadequacy of federal standards. Like our state cases, the federal cases recognize two primary tiers of review: “rational basis” (see, e.g., FCC v. Beach Communications, Inc. (1993) 508 U.S. 307, 313 [113 S.Ct. 2096, 2100-2103, 124 L.Ed.2d 211]) and “strict scrutiny” (see, e.g., Shaw v. Hunt (1996) 517 U.S. 899, 907 [116 S.Ct. 1894, 1901, 135 L.Ed.2d 207]). But, in practice, “rational basis,” as used in the federal cases, means any basis, and “strict scrutiny” is “ ‘strict’ in theory [but] fatal in fact.” (Gunther, Foreword: In Search of Evolving Doctrine on a Changing Court: A Model for a Newer Equal Protection (1972) 86 Harv. L.Rev. 1, 8 (Evolving Doctrine).) As a result, the outcome in every case turns on how the court chooses to characterize the classification. Suspect classifications, such as those based on race or that impact rights the court deems “fundamental,” warrant strict (read “fatal”) scrutiny; other classifications warrant rational basis (read “anything goes”) review. (FCC v. Beach Communications, Inc., supra, 508 U.S. 307, 313-316 [113 S.Ct. 2096, 2100-2103].) As Justice Mosk so aptly put it, “[t]he vice of the binary theory ... is that it applies either a standard that is virtually always met (the rational relationship test) or one that is almost never satisfied (the strict scrutiny test). [Citation.] Once the test is *662selected, the result of its application is foreordained . . . .” (Hays v. Wood (1979) 25 Cal.3d 772, 796 [160 Cal.Rptr. 102, 603 P.2d 19] (conc. opn. of Mosk, J.).)
Thirty years ago, Professor Gunther recognized the inadequacy of this approach to equal protection and suggested that the high court might be moving toward “means scrutiny.” (Evolving Doctrine, supra, 86 Harv. L.Rev. at p. 24.) Under means scrutiny, courts do not second-guess legislative purposes, unless, of course, those purposes contravene explicit constitutional constraints such as the right to free speech. (Id. at pp. 21-23.) Instead, courts examine the means the legislature chose, in order to ensure that those means adequately advance the legislative purpose. (Id. at pp. 21, 46-47.)
Means scrutiny has the advantage of not being a subjective process that turns merely on the personal philosophies or biases of judges. (Evolving Doctrine, supra, 86 Harv. L.Rev. at pp. 42-43.) Rather, whether a law in fact furthers its purpose—that is, whether it has proved successful over time—is something litigants can prove or disprove based on objective criteria, .and therefore it is the sort of inquiry courts are capable of making. (Id. at p. 24.) Moreover; because means scrutiny merely asks whether a law actually does what the legislature intended it to do, it does not interfere in the legislative process, but only encourages legislatures to do their work carefully so as to avoid unintended results. As Professor Gunther argued in his groundbreaking article, “[m]eans scrutiny . . . can improve the quality of the political process—without second-guessing the substantive validity of its results—by encouraging a fuller airing in the political arena of the grounds for legislative action. . . . fi[] . . . Too often the only assurance that the [legislature] has thought about the issues is the judicial presumption that it has. Means scrutiny would provide greater safeguards that the presumed process corresponds to reality . . . .” (Id. at p. 44.)
The high court adopted Professor Gunther’s proposal in a series of cases interpreting the federal takings clause. Specifically, the court held that a regulation effects a taking if it does not “ ‘substantially advance’ the ‘legitimate state interest’ sought to be achieved.” (Nollan v. California Coastal Comm’n (1987) 483 U.S. 825, 834, fn. 3 [107 S.Ct. 3141, 3147, 97 L.Ed.2d 677]; see also Dolan v. City of Tigard (1994) 512 U.S. 374, 385 [114 S.Ct. 2309, 2316-2317, 129 L.Ed.2d 304]; Lucas v. South Carolina Coastal Council (1992) 505 U.S. 1003, 1016 [112 S.Ct. 2886, 2893-2894, 120 L.Ed.2d 798]; Keystone Bituminous Coal Assn. v. DeBenedictis (1987) 480 U.S. 470, 485 [107 S.Ct. 1232, 1241-1242, 94 L.Ed.2d 472]; United States v. Riverside Bayview Homes, Inc. (1985) 474 U.S. 121, 126 [106 S.Ct. 455, 458-459, 88 *663L.Ed.2d 419]; Agins v. Tiburon (1980) 447 U.S. 255, 260 [100 S.Ct. 2138, 2141, 65 L.Ed.2d 106].) Obviously, this requirement that legislative means “substantially advance” legislative purposes is more demanding than rational basis review, which only requires some “rational relationship” between means and ends. (Santa Monica Beach, Ltd. v. Superior Court (1998) 19 Cal.4th 952, 1027-1028 [81 Cal.Rptr.2d 93, 968 P.2d 993] (dis. opn. of Brown, J.).) More important, in Brown v. Merlo and Hays v. Wood, we adopted means scrutiny as the standard of judicial review applicable under the state equal protection provisions.
In Brown v. Merlo, supra, we struck down an automobile guest statute that prohibited an injured automobile guest from recovering for the negligent driving of his host, except in certain narrow circumstances. (8 Cal.3d at pp. 858-859.) We noted that, if the law’s purpose was to protect hospitality, it was grossly underinclusive because the law singled out “automobile guests as distinguished from all other guests.” (Id. at p. 859; see also id. at p. 864.) On the other hand, if its purpose was to prevent collusive lawsuits between automobile guests and drivers, the law was grossly overinclusive because the vast majority of automobile guests would not file collusive lawsuits. (Id. at p. 859; see also id. at pp. 874-875.) We said, “it is unreasonable to eliminate causes of action of an entire class of persons simply because some undefined portion of the designated class may file fraudulent lawsuits.” (Id. at p. 875.)
More generally, we noted that an “overinclusive classification scheme . . . does not treat similarly situated individuals in like manner . . . .” (Brown v. Merlo, supra, 8 Cal.3d at p. 876.) Accordingly, we held that “[although the constitutional provisions do not require absolute precision in the designation of classifications, they do not tolerate classifications which are so grossly overinclusive as to defy notions of fairness or reasonableness.” (Id. at p. 877.) We reached the same conclusion with respect to grossly underinclusive classifications (id. at p. 877, fn. 17), and we made clear that we were construing the state Constitution and therefore were not constrained by the possibility that more deferential standards might apply under the federal Constitution. (Id. at p. 865, fn. 7.)
Thus, our holding in Brown v. Merlo, supra, 8 Cal.3d 855, instructed courts to scrutinize the means the Legislature chose to advance its purpose. Specifically, we held that if the Legislature relied on grossly overinclusive or underinclusive classifications, then it chose means that are constitutionally inadequate, and courts must strike the law down. And so as to leave absolutely no doubt about this point, we reiterated it seven years later in Hays v. Wood. In that case, we struck down a law that treated public officials who were attorneys or brokers differently from other public officials with *664respect to income disclosure. (Hays v. Wood, supra, 25 Cal.3d at pp. 778-779.) The Attorney General justified the statute as a protection against self-serving bias on the part of public officials, arguing that attorneys and brokers receive greater profit from their business endeavors than other persons. (Id. at pp. 788-789.) We found this justification inadequate because it was underinclusive. Specifically, we saw no reasonable explanation for “the selection of but two of the several professions having relatively high profit margins for . . . special treatment . . . .” (Id. at p. 789.) We noted that the Attorney General’s argument did “no more than shift the focus of inquiry from one distinction to another.” (Ibid.) In other words, the law might be justified in treating attorneys and brokers differently from business people in general, but it was not justified in treating attorneys and brokers differently from other business people who enjoyed comparably high profit, margins. (Id. at pp. 789-790.)
Significantly, we expressly rejected the federal precedents on which the majority relies today, including Williamson v. Lee Optical Co. (1955) 348 U.S. 483 [75 S.Ct. 461, 99 L.Ed. 563], saying, “respondents read these cases too broadly—at least insofar as they relate to the application of the equal protection provisions of our state Constitution.” (Hays v. Wood, supra, 25 Cal.3d at p. 790, italics added.) We continued: “Werner v. Southern Cal. etc. Newspapers [(1950) 35 Cal.2d 121 [216 P.2d 825, 13 A.L.R.2d 252]], which represents the definitive statement of the subject doctrine in this jurisdiction, makes it lucidly apparent that the legislative body, when it chooses to address a particular area of concern in less than comprehensive fashion by merely ‘striking the evil where it is felt most’ [citation] may not do so wholly at its whim.” (Ibid.) Rather “its decision as to where to ‘strike’ must have a rational basis in light of the legislative objectives.” (Id. at p. 791.) We held that the ultimate question is “whether there exists a reasonable basis for the legislative body to conclude that the members of [the legislative classification] made up ‘the most conspicuous example of the danger it sought to preclude’ [citation].” (Id. at p. 792, quoting Werner v. Southern Cal. etc. Newspapers (1950) 35 Cal.2d 121, 132 [216 P.2d 825, 13 A.L.R.2d 252].)
Therefore, though the federal Constitution may permit legislatures to take a shotgun approach to problem solving (see Williamson v. Lee Optical Co., supra, 348 U.S. 483), our state Constitution insists on greater precision, and it does so by requiring courts to scrutinize the means the Legislature chose to advance its purposes. Rather than merely “rubberstamping” the legislative categories at issue here, we should be engaging in “ ‘a serious and genuine judicial inquiry into the correspondence between the classification and the legislative goals’ ” (Newland v. Board of Governors, supra, 19 Cal.3d at p. 711), and, more particularly, we should be asking whether the legislative *665classifications substantially advance the legislative purposes without being “grossly overinclusive” or “underinclusive.” (Brown v. Merlo, supra, 8 Cal.3d at p. 877 & fn. 17.)
n
Applying these standards, I have no trouble concluding that the California MCLE program violates the equal protection guaranties of our state Constitution by exempting some active attorneys without exempting others who are “similarly situated with respect to the legitimate purpose of the law.” (Purdy & Fitzpatrick v. State of California (1969) 71 Cal.2d 566, 578 [79 Cal.Rptr. 77, 456 P.2d 645, 38 A.L.R.3d 1194].)
Rule 958 of the California Rules of Court authorized the State Bar to establish the MCLE program. Pursuant to rule 958, the Board of Governors of the State Bar adopted detailed rules and regulations that govern MCLE. These regulations created five exemptions from the program, three of which are at issue here: (1) retired judges, (2) officers and elected officials of the State of California, and (3) full-time law professors at accredited law schools.
The majority hypothesizes two purposes that it claims justify the exemptions: (1) attorneys in the exempted categories “as a general matter, are less likely ... to represent clients on a full-time basis . . . ,” and (2) these attorneys “(again, as a general matter) are less likely ... to need continuing education . . . .” (Maj. opn., ante, at p. 646, original italics.) The majority is wrong ah initio under our state Constitution to hypothesize purposes that support the exemptions. In Brown v. Merlo, supra, we rejected “a highly fictional approach to statutory purpose.” (8 Cal.3d at p. 865, fn. 7.) We made it clear that, under the state Constitution, “statutory classifications [must] bear some substantial relationship to an actual, not ‘constructive,’ legislative purpose.” (Ibid.) We reaffirmed this principle in Cooper v. Bray (1978) 21 Cal.3d 841, 854 [148 Cal.Rptr. 148, 582 P.2d 604], where we said that the purpose under consideration must be one that “can realistically be ascribed to the Legislature.” The majority points to nothing in the legislative record indicating that the purposes it hypothesizes are the “actual” or “realistic[]” purposes underlying the exemptions, and, as we shall see, they are in fact highly unlikely.
Moreover, even if the State Bar actually had in mind the purposes that the majority hypothesizes, the classifications reflected in the exemptions violate the state equal protection guaranties because they are grossly overinclusive and underinclusive and thus do not afford “similarly situated” persons “like *666treatment.” (Purdy & Fitzpatrick v. State of California, supra, 71 Cal.2d at p. 578.) First, though it may well be true that retired judges, state officers, and full-time law professors are less likely than other active attorneys to represent clients on a full-time basis, this is only one very small selection of the active attorneys for whom this description applies, and it is certainly not “ ‘the most conspicuous example.’ ” (Hays v. Wood, supra, 25 Cal.3d at p. 792, quoting Werner v. Southern Cal. etc. Newspapers, supra, 35 Cal.2d at p. 132.) If the purpose of the exemptions were truly to exempt attorneys who represent clients on a part-time basis, the rule could easily have included a much broader and more precise exemption that named part-time attorneys expressly. Moreover, as Justice Kennard points out, the hypothetical purpose of exempting part-time attorneys simply makes no sense in light of the more general purpose of consumer protection that underlies the MCLE program: an incompetent part-time attorney is just as dangerous to the consumer as an incompetent full-time attorney. Indeed, the State Bar recognizes this fact by removing the exemption for state employees if they work just one hour outside the scope of their state employment. (State Bar Com. on MCLE, Guidelines: MCLE Rules & Regs., § 6.1.4, eff. Sept. 1, 1995.)
Second, the majority speculates, without any empirical verification, that retired judges, state officers, and full-time law professors are “less likely ... to need continuing education.” (Maj. opn., ante, at p. 646.) While this is a comforting thought, it is a questionable premise. The majority cites “the high level of legal skills and qualifications that generally are required to obtain a position as a judge,” and it points out that “the daily work of a judge involves constant exposure to and analysis of ongoing developments in the law.” (Ibid.) True. But high qualifications and exposure to ongoing developments in the law are only one way in which MCLE serves its consumer-protection purpose. MCLE also requires that attorneys study legal ethics, law practice management, and the prevention of substance abuse and emotional distress (Cal. Rules of Court, rule 958(c); State Bar, MCLE Rules & Regs., § 2.1), subjects which judges do not usually encounter in the course of their work. Furthermore, even if judges are somehow less in need of continuing education while sitting as judges, the majority does not explain why it is rational to assume they are still less in need after they have returned to private practice. As Justice Kennard points out, the purpose of continuing education is not to ensure that lawyers are competent and knowledgeable when they first start their practices, but to ensure that they remain competent and knowledgeable as the years progress.
More important, the classifications are once again grossly overinclusive and underinclusive. Many attorneys who are not retired judges have just as much expertise as retired judges, and at least some retired judges lack the *667expertise the majority imputes to them. If the purpose of the exemptions were truly to exempt attorneys who are less in need of continuing education, the rule could easily have allowed attorneys to satisfy MCLE by taking a statewide examination.
The same arguments can be made about the relative knowledge and expertise of state officers and full-time law professors. While individuals in these groups may have more legal expertise than attorneys in general, the majority’s assertion that, as a group, they have more legal expertise would probably fail to withstand empirical verification, particularly in the areas of legal ethics, law practice management, and the prevention of substance abuse and emotional distress. In any case, these employment categories are, at best, an imperfect basis for assessing legal expertise. And the majority’s assertion that state officers are likely to practice only in areas in which they have expertise is mere speculation and probably true about almost every attorney in the state.
If the purposes of the exemptions are in fact those that the majority has hypothesized, then the State Bar rules achieve those purposes in a ludicrously impractical way, though they could just as easily achieve the same purposes directly and precisely. That approach to legislation is not rational, and we exhibit disrespect for the legislative process when we impute such an irrational approach to our policymakers. Moreover, we undermine the integrity of the political process when we hypothesize such highly unlikely purposes in order to uphold arbitrary and self-serving exemptions.
I dissent.
Appellant’s petition for a rehearing was denied October 20, 1999. Kennard, J., and Brown, J., were of the opinion that the petition should be granted.