Opinion

In this proceeding, we consider constitutional challenges to an initiative measure adopted at the November 6, 1990, General Election. This measure, entitled by its framers as “The Political Reform Act of 1990,” was designated on the ballot as Proposition 140. Its stated purpose is to “restore a free and democratic system of fair elections, and to encourage qualified candidates to seek public office” by limiting “the powers of *500 incumbency.” (Cal. Const., art. IV, § 1.5, added by Prop. 140.) The measure seeks to accomplish these goals by these three separate reforms: “Retirement benefits [of legislators] must be restricted, state-financed incumbent staff and support services limited, and limitations placed upon the number of terms which may be served.” (Ibid.)

Petitioners herein include the California Legislature (both Senate and Assembly), certain individual legislators from both houses, and various citizens, voters and taxpayers. Petitioners jointly assert a variety of challenges to the constitutionality of Proposition 140.

Respondents are various public officials (including the Secretary of State, the state Controller, and the Board of Administration of the Public Employees Retirement System [PERS]) responsible for implementing, enforcing or applying the new measure. Intervener, Californians for a Citizen Government, is the organization that sponsored Proposition 140. Several amici curiae have filed briefs supporting the various parties.

Only intervener (represented by the Pacific Legal Foundation) and respondent Secretary of State (represented by the Attorney General) have filed briefs adverse to petitioners. Respondent state Controller remains neutral, while respondent PERS supports petitioners’ challenge to the pension limitations of Proposition 140, but takes no position on the remaining issues. Nonetheless, for convenience, we sometimes refer to intervener and respondent Secretary of State jointly as “respondents.”

The petition for mandate sought original relief in this court. Although we customarily decline to exercise such jurisdiction, preferring initial disposition by the lower courts, the present case involves issues of sufficient public importance to justify departing from our usual course. As we recently observed in Raven v. Deukmejian (1990) 52 Cal.3d 336, 340 [ 276 Cal.Rptr. 326 , 801 P.2d 1077 ] (assessing the constitutionality of Proposition 115, an initiative measure adopted at the June 1990 Primary Election), quoting from an earlier case, “ ‘It is uniformly agreed that the issues are of great public importance and should be resolved promptly. Accordingly, under well settled principles, it is appropriate that we exercise our original jurisdiction. [Citations.]’ . . .” (See also Planned Parenthood Affiliates v. Van de Kamp (1986) 181 Cal.App.3d 245, 262-265 [ 226 Cal.Rptr. 361 ].)

Pending our review of the issues, on June 14,1991, we temporarily stayed operation of section 5 of Proposition 115, which imposes budgetary restrictions on the Legislature, as hereafter discussed.

The principles that guide us in evaluating the validity of initiative measures such as Proposition 140 are likewise well settled. Although the *501 legislative power under our state Constitution is vested in the Legislature, “the people reserve to themselves the powers of initiative and referendum.” (Cal. Const., art. IV, § 1.) Accordingly, the initiative power must be liberally construed to promote the democratic process. (Raven v. Deukmejian, supra, 52 Cal.3d at p. 341 .) Indeed, it is our solemn duty to jealously guard the precious initiative power, and to resolve any reasonable doubts in favor of its exercise. (Ibid., and cases cited.) As with statutes adopted by the Legislature, all presumptions favor the validity of initiative measures and mere doubts as to validity are insufficient; such measures must be upheld unless their unconstitutionality clearly, positively, and unmistakably appears. (Calfarm Ins. Co. v. Deukmejian (1989) 48 Cal.3d 805, 814 [ 258 Cal.Rptr. 161 , 111 P.2d 1247 ] [evaluating the constitutionality of Prop. 103, an insurance rate initiative measure adopted at the Nov. 1988 Gen. Elec.].)

As will appear, we conclude that although the pension restrictions of Proposition 140 are invalid as to incumbent legislators, the measure is constitutionally valid in all other respects. Before addressing petitioners’ challenges, we briefly outline the basic provisions of the new measure, which consists entirely of various amendments to the state Constitution.

I. Summary of Proposition 140

A. Statement of Purpose and Findings

Section 1.5 is added to article IV of the Constitution (the “Legislative” article), to set forth various findings and statements of the framers’ intent. In pertinent part, this introductory provision recites that although “the Founding Fathers established a system of representative government based upon free, fair, and competitive elections,” nonetheless an “extremely high number of incumbents” are reelected by reason of “[t]he ability of legislators to serve unlimited number of terms, to establish their own retirement system, and to pay for staff and support services at state expense . . . .”

The introductory statement continues by noting that “These unfair incumbent advantages discourage qualified candidates from seeking public office and create a class of career politicians, instead of the citizen representatives envisioned by the Founding Fathers.” The statement concludes by stating that, “To restore a free and democratic system of fair elections, and to encourage qualified candidates to seek public office, the people find and declare that the powers of incumbency must be limited,” as provided by the substantive amendments we now discuss.

B. Term Limitations

The measure imposes specific term limitations for state legislators and various state constitutional officers. Persons elected or appointed on or after *502 November 6, 1990, to the office of Governor (Cal. Const., art. V, § 2), Lieutenant Governor (id., § 11), Attorney General (ibid.), Controller (ibid.), Secretary of State (ibid.), Treasurer (ibid.), Superintendent of Public Instruction (id., art. IX, § 2), or to the State Board of Equalization (id., art. XIII, § 17), or the state Senate (id., art. IV, § 2, subd. (a)), are limited to two terms. Members of the Assembly are limited to three terms (ibid.).

Section 7 is added to article XX of the Constitution (the “Miscellaneous Subjects” article), to explain that the foregoing term limitations “apply only to terms to which persons are elected or appointed on or after November 6, 1990, except that an incumbent Senator whose office is not on the ballot for the general election on that date may serve only one additional term,” and that the foregoing term limits “shall not apply to any unexpired term to which a person is elected or appointed if the remainder of the term is less than half of the full term.”

C. Budgetary Limitations

The measure imposes a budgetary limitation for the Legislature. Section 7.5 is added to article IV of the Constitution to provide that, for the forthcoming fiscal year, “the total aggregate expenditures of the Legislature for the compensation of members and employees of, and the operating expenses and equipment for, the Legislature may not exceed” $950,000 “per member” for that fiscal year, or 80 percent of the amount of money expended for such purposes in the preceding year, whichever is less.

Additionally, new section 7.5 of article IV of the Constitution provides that, for each fiscal year thereafter, the total aggregate expenditures may not exceed the amount expended during the previous year, “adjusted and compounded by an amount equal to the percentage increase in the appropriations limit for the state established pursuant to Article XIIIB” of the Constitution.

D. Pension Limitations

Finally, the measure imposes limitations on legislators’ pension rights. New section 4.5 is added to article IV of the Constitution to provide that the state will contribute the employer’s share to the federal Social Security system on behalf of participating legislators “elected to or serving in the Legislature on or after November 1, 1990,” but “[n]o other pension or retirement benefit shall accrue as a result of service in the Legislature, such service not being intended as a career occupation.”

This same provision further provides that “This Section shall not be construed to abrogate or diminish any vested pension or retirement benefit *503 which may have accrued under an existing law . . . , but upon adoption of this Act no further entitlement to nor vesting in any existing program shall accrue to any such person, other than [federal] Social Security . . .

Respondent state Controller confirms that, as of the effective date of Proposition 140, he ceased making deductions from the salaries of incumbent legislators for the Legislators’ Retirement Fund in connection with services rendered after that date, and has also terminated paying employer or state contributions to that fund for those services.

E. Severance Clause

In addition to the foregoing provisions, subdivision (d) is added to section 11 of article VII of the Constitution (the “Public Officers and Employees” article). Section 11 of article VII previously had limited the obligations of the Legislators’ Retirement System to persons first entering state office after January 1, 1987. Subdivision (d) now sets forth a severance clause reciting that “If any part of this measure or the application to any person or circumstance is held invalid, the invalidity shall not affect other provisions or applications which reasonably can be given effect without the invalid provision or application.” We discuss the matter of severance below in connection with our discussion of the invalid pension restrictions.

II. Discussion

A. Lifetime Ban or Limit on Consecutive Terms

As a preliminary matter, we must address the interpretive question whether Proposition 140 imposes a “lifetime ban” on officers who have served the specified number of terms, or merely limits the number of consecutive terms they may serve. With respect to state legislators, petitioners and intervener assume that once a legislator has served the prescribed maximum number of terms, the measure would forever bar him or her from running for a seat in the legislative house previously served. Respondent Secretary of State, contends, however, that the term limitation “is directed only at the incumbent office holder.” (Italics in original.) In her view, the measure simply limits the number of consecutive terms served, and she suggests that a former legislator might run for a new term of office if he or she is not currently holding that office.

We focus first on the language of the new measure. The introduction to Proposition 140 refers to curtailing “[t]he ability of legislators to serve unlimited . . . terms,” by limiting “the number of terms which may be served.” (Cal. Const., art. IV, § 1.5.) Thus, the limitation on the term of state *504 Senators adds the language, “No Senator may serve more than 2 terms,” to the existing language of the Constitution that “The Senate has a membership of 40 Senators elected for 4-year terms, 20 to begin every 2 years.” (Cal. Const., art. IV, § 2, subd. (a).) The limitation on Assembly members is similarly phrased, stating that “No member of the Assembly may serve more than 3 terms.” (Ibid.) As petitioners observe, the foregoing provisions do not expressly refer to the number of consecutive terms served. Moreover, the measure repeatedly announces its intent to eliminate “career politicians” (see Cal. Const., art. IV, §§ 1.5,4.5), language which would support the view that a lifetime ban was intended.

Respondent Secretary of State points out, however, that these limitations are directed to a “Senator” or “member of the Assembly,” rather than a “person,” and she argues that a literal interpretation of the measure thus discloses an intent to limit the right of an incumbent legislator to continue unabated terms in office. In respondent Eu’s view, once the incumbent legislator has left office, he or she can no longer be described as a “Senator” or “member of the Assembly” to which the term limitation provision would apply.

Petitioners argue that even a purely literal reading of Proposition 140 would support their interpretation: A person who, having already served two prior terms as Senator, surrenders his or her seat for one term and thereafter enters on a third Senate term, could be accurately described as a “Senator” who is “serving] more than 2 terms” contrary to the language of Proposition 140. Intervener suggests that respondent Eu’s interpretation, with its emphasis on limiting only incumbents, could lead to absurd results, permitting a legislator to avoid the term limitations by resigning shortly before his or her final term had expired, and thereupon announcing an intent to run for reelection unencumbered by any term limitations applicable to “incumbents.”

Nonetheless, we agree with respondent Eu that the language of Proposition 140 is ambiguous as to its intent to impose a lifetime ban. As we have previously recognized, to help resolve such ambiguities “it is appropriate to consider indicia of the voters’ intent other than the language of the provision itself. [Citation.]” (Kennedy Wholesale, Inc. v. State Bd. of Equalization (1991) 53 Cal.3d 245, 250 [ 279 Cal.Rptr. 325 , 806 P.2d 1360 ].) Such indicia include the analysis and arguments contained in the official ballot pamphlet. (See ibid.; Amador Valley Joint Union High Sch. Dist. v. State Bd. of Equalization (1978) 22 Cal.3d 208, 245-246 [ 149 Cal.Rptr. 239 , 583 P.2d 1281 ] [hereafter Amador].) These materials strongly support the position of petitioners and intervener that a lifetime ban from office was contemplated by the framers of, and voters for, Proposition 140.

*505 First, the analysis by the Legislative Analyst described the term limitations as limiting “the number of terms that an elected state official can serve in the same office . . . (Ballot Pamp., Proposed Stats, and Amends, to Cal. Const, with argument to voters, Gen. Elec. (Nov. 6, 1990) p. 69, italics in original [hereafter Ballot Pamphlet].) No suggestion is made that only a consecutive term limitation was contemplated.

Second, and more significantly, the opponents’ ballot arguments against Proposition 140 forcefully and repeatedly stressed the measure’s “lifetime ban,” and criticized the proponents’ failure to disclose that particular aspect of the measure in their arguments to the voters. Indeed, the primary thrust of the opponents’ ballot arguments was directed to this aspect of the measure. (See Ballot Pamp., supra, at pp. 70-71.) The argument against Proposition 140 used the phrases “lifetime ban,” “banned for life,” or similar terminology 11 times. (Ibid.)

We are mindful of the fact that ballot measure opponents frequently overstate the adverse effects of the challenged measure, and that their “fears and doubts” are not highly authoritative in construing the measure. (DeBartolo Corp. v. Fla. Gulf Coast Trades Council (1988) 485 U.S. 568, 585 [ 99 L.Ed.2d 645, 661 , 108 S.Ct. 1392 ].) Nonetheless, we find it significant that the proponents failed to contradict the opponents’ “lifetime ban” argument. (See D’Amico v. Board of Medical Examiners (1970) 6 Cal.App.3d 716, 725 [ 86 Cal.Rptr. 245 ].) Moreover, the proponents stressed that Proposition 140 was directed at eliminating “career politicians” or “career legislators,” and suggested that “good legislators will always have the opportunity to move up the ladder.” (Italics added.) These arguments seem to reinforce the idea that a lifetime ban was intended.

We think it likely the average voter, reading the proposed constitutional language as supplemented by the foregoing analysis and arguments, would conclude the measure contemplated a lifetime ban against candidacy for the office once the prescribed maximum number of terms had been served. As we stated in In re Lance W. (1985) 37 Cal.3d 873, 889 [ 210 Cal.Rptr. 631 , 694 P.2d 744 ], “In construing constitutional and statutory provisions, whether enacted by the Legislature or by initiative, the intent of the enacting body is the paramount consideration.” (Italics added.)

Respondent Eu suggests that the applicable principle of interpretation requires us to avoid any constitutional “doubts” or “difficulties” by adopting a construction which will render the measure constitutional. (E.g., Kortum v. Alkire (1977) 69 Cal.App.3d 325, 333-334 [ 138 Cal.Rptr. 26 ].) We find no cases suggesting, however, that this interpretive principle precludes resort to extrinsic evidence to resolve possible ambiguities in a measure. Moreover, as *506 we explain in this opinion, the measure’s lifetime ban is constitutional in all respects.

We conclude that Proposition 140’s term limitations extend over the lifetime of each affected officeholder.

B. Constitutional Revision or Amendment

Turning to petitioners’ constitutional challenges to Proposition 140, they first contend that the measure, and particularly its term and budgetary limitations on the Legislature, effected a constitutional revision rather than a mere amendment. Although “[t]he electors may amend the Constitution by initiative” (Cal. Const., art. XVIII, § 3), a “revision” of the Constitution may be accomplished only by convening a constitutional convention and obtaining popular ratification (id., §§ 2, 4), or by legislative submission of the measure to the voters (id., §§ 1,4). (See also Raven v. Deukmejian, supra, 52 Cal.3d at p. 349 .)

Raven suggested that the revision provision is based on the principle that “comprehensive changes” to the Constitution require more formality, discussion and deliberation than is available through the initiative process. (52 Cal.3d at pp. 349-350.) We should bear in mind, however, that the initiative process may represent the only practical means of achieving the kind of “reforms” of the Legislature involved here, because the revision process can be initiated only with the consent of two-thirds membership of each house of the Legislature. (Cal. Const., art. XVIII, §§ 1, 2.)

In Raven , citing earlier cases, we observed, “Although the Constitution does not define the terms ‘amendment’ or ‘revision,’ the courts have developed some guidelines helpful in resolving the present issue. As explained [in prior cases], our revision/amendment analysis has a dual aspect, requiring us to examine both the quantitative and qualitative effects of the measure on our constitutional scheme. Substantial changes in either respect could amount to a revision.” ( 52 Cal.3d at p. 350 .)

Thus, a constitutional “revision” need not involve widespread deletions, additions''and amendments affecting a host of constitutional provisions and resulting in a quantitative revision. “[E]ven a relatively simple enactment may accomplish such far reaching changes in the nature of our basic governmental plan as to amount to a revision also.” (Amador, supra, 22 Cal.3d at p. 223 ; see Raven v. Deukmejian, supra, 52 Cal.3d at pp. 351-352.)

In Raven , we unanimously struck down as an improper revision the provision of Proposition 115 that, in essence, would have vested in the *507 United States Supreme Court all interpretive power as to certain fundamental procedural rights of criminal defendants under the state Constitution. We observed that the challenged provision would have “devastating” effects, by drastically limiting the procedural rights of criminal defendants, by infringing on the power of the state judiciary to protect California citizens from arbitrary or capricious legislation, and by substantially altering the substance and integrity of the state Constitution as a document of independent force and effect. (52 Cal.3d at pp. 352-354.)

Petitioners assert the effects of the term and budgetary limitations of Proposition 140 on the Legislature are similarly dramatic, causing far-reaching changes in our basic governmental plan. As they phrase the argument, “Weakened by budget cuts and harsh term limits, the Legislature will be unable to discharge its traditional duties of policymaker, keeper of the purse, and counterweight to the executive branch in the way the Constitution intends. The result is a change so profound in the structure of our government that it constitutes a revision . . . .”

In petitioners’ words, the new term limitations will “virtually guarantee that California will have a new Legislature—that is, a Legislature in which the vast majority of the members are new—every six years.” Petitioners note that legislators with “substantial experience” and valuable “expertise” ultimately will be required to surrender forever their legislative offices, and persons with minimal legislative experience will be appointed chairpersons of major committees. Petitioners predict that the Governor’s role in the delicately balanced state governmental process will grow as inexperienced legislators become “susceptible to the subtle pressures of the gubernatorial power of appointment.”

Additionally, according to petitioners, the power of special interest groups to defeat legislative programs will be enhanced. Those groups opposing legislation proposed by “lame duck legislators” can merely “wait until they leave to dismantle it.” Petitioners suggest that inexperienced legislators, lacking any background in “dealing with the budget” or “dealing with each other,” will be greatly hindered in their ability to review and pass a satisfactory state budget. They note that several legislators currently serving on key budget committees have done so for periods far in excess of the terms now permitted by Proposition 140.

As for the budgetary limitations imposed by Proposition 140, petitioners observe that the new measure will force a substantial reduction in the funds available for compensating legislators and their staffs, and defraying legislative operating expenses. Approximately $176 million was appropriated for the Legislature for the 1990-1991 fiscal year, but under Proposition 140 only *508 $114 million would be appropriated for the current fiscal year, resulting in a substantial reduction in fünds. As the Legislative Analyst explained in the voters’ ballot pamphlet, “In 1991-92, expenditures by the Legislature would be reduced by about 38 percent, or $70 million.” (Ballot Pamp., supra, at p. 69.)

According to petitioners, and undisputed by respondents, in anticipation of the forthcoming budgetary reduction, “as of March 1, 1991, more than 640 employees will have left the Legislature as a result of Proposition 140.” For the current fiscal year, “all [Senate] standing committee budgets will be reduced by 20% and all subcommittees and select committees will be eliminated. . . . The Assembly expects to have to take similar steps.”

Petitioners note that among those departing from the Legislature are experienced consultants previously hired to provide accurate information concerning proposed or pending legislation. Petitioners predict the Legislature will become “an essentially reactive body,” dependent on lobbyists and special interest groups as a source of information, and lacking sufficient funds to develop its own “legislative agenda.”

After we granted review in this matter, petitioners sought a stay of the budgetary limitation provisions of Proposition 140, claiming that the offices of Legislative Analyst and Auditor General would be jeopardized unless a stay were granted prior to June 15, 1991, when the proposed budget for the forthcoming fiscal year must be sent to the Governor. (Cal. Const., art. IV, § 12, subd. (c).) Although intervener suggested that petitioners’ request amounted to mere “gamesmanship” designed to lend support to their revision challenge, as previously indicated we issued a temporary stay pending our review of the issues.

As previously noted, petitioners contend that the combined effects of the foregoing term and budgetary limitations on California’s “basic governmental plan” will be as devastating and far reaching as those involved in the provision of Proposition 115 invalidated by us in Raven v. Deukmejian, supra, 52 Cal.3d 336 . They thus assert that Proposition 140 has achieved a qualitative revision of the Constitution. We disagree.

First, the basic and fundamental structure of the Legislature as a representative branch of government is left substantially unchanged by Proposition 140. Term and budgetary limitations may affect and alter the particular legislators and staff who participate in the legislative process, but the process itself should remain essentially as previously contemplated by our Constitution. This aspect distinguishes the present case from Raven , in which we struck down a provision that would have fundamentally changed and *509 subordinated the constitutional role assumed by the judiciary in the governmental process. (See also Amador, supra, 22 Cal.3d 208 , 223 [posing hypothetical example of provision vesting all judicial power in Legislature].)

As indicated in Raven , a qualitative revision includes one that involves a change in the basic plan of California government, i.e., a change in its fundamental structure or the foundational powers of its branches. ( Raven, supra, 52 Cal.3d at pp. 352-355.) Raven invalidated a portion of Proposition 115 because it deprived the state judiciary of its foundational power to decide cases by independently interpreting provisions of the state Constitution, and delegated that power to the United States Supreme Court. (Ibid.)

By contrast, Proposition 140 on its face does not affect either the structure or the foundational powers of the Legislature, which remains free to enact whatever laws it deems appropriate. The challenged measure alters neither the content of those laws nor the process by which they are adopted. No legislative power is diminished or delegated to other persons or agencies. The relationships between the three governmental branches, and their respective powers, remain untouched.

Second, although the immediate foreseeable effects of the foregoing term and budgetary limitations are indeed substantial (primarily, the eventual loss of experienced legislators and some support staff), the assertedly momentous consequences to our governmental scheme are largely speculative ones, dependent on a number of as yet unproved premises.

Petitioners assume, for example, that the eventual loss of experienced legislators, and the arrival of their relatively unseasoned replacements, will irreparably hinder and damage the legislative process. Yet respondents argue with equal conviction that Proposition 140’s term limitations will free the entire process from the control of assertedly entrenched, apathetic, veteran incumbents, thereby allowing fresh creative energies to flourish free of vested, self-serving legislative interests. Respondents also note that the office of Governor is likewise subject to the limitation of two terms under Proposition 140, thus lessening the likelihood that the Legislature will become a subordinate branch.

In similar fashion, whereas petitioners forecast bleak consequences from the budgetary limitations and consequent loss of experienced support staff, respondents assume the monetary restraints will contribute in a positive manner toward eliminating excessive legislative spending and terminating surplus or inefficient personnel.

*510 Petitioners view these budgetary limitations as akin to infringements on the Legislature’s inherent “power of self-preservation” as an independent branch of government. (See Millholen v. Riley (1930) 211 Cal. 29, 33-34 [ 293 P. 69 ]; see also Brown v. Superior Court (1982) 33 Cal.3d 242, 248, fn. 5 [ 188 Cal.Rptr. 425 , 655 P.2d 1260 ].) In petitioners’ words, “a 38% reduction in funds for the Legislature threatens the functioning of that branch and alters the structure of government.” Respondents, on the other hand, assert that the remaining budget allocation (nearly $1 million per legislator for the current fiscal year) is more than ample to provide an effective and efficient legislative staff.

Intervener declares that the Legislature’s budget has increased 838 percent during the period from 1968 to 1990, while the Governor’s budget increased only 417 percent during this same period. Thus, even after the 38 percent budget cut attributable to Proposition 140, the Legislature’s budget will have increased at a substantially greater rate than the Governor’s budget over the last 22 years.

We are in no position to resolve the controversy between the parties regarding the long-term consequences of Proposition 140, for the future effects of that measure on our “basic governmental plan” are simply unfathomable at this time. Indeed, that very uncertainty inhibits us from holding that a constitutional revision has occurred in this case. Our prior decisions have made it clear that to find such a revision, it must necessarily or inevitably appear from the face of the challenged provision that the measure will substantially alter the basic governmental framework set forth in our Constitution. (Brosnahan v. Brown (1982) 32 Cal.3d 236, 261 [ 186 Cal.Rptr. 30 , 651 P.2d 274 ] [rejecting argument that Prop. 8 (see Cal. Const., art. I, § 28) involved improper constitutional revision]; id,., at pp. 258-259 [rejecting argument that Prop. 8 would improperly cause impairment of essential governmental functions]; Amador, supra, 22 Cal.3d at pp. 224-226 [nothing on face of Prop. 13 (see Cal. Const., art. XIII A) “necessarily and inevitably” would result in loss of home rule]; see Raven v. Deukmejian, supra, 52 Cal.3d at p. 349 [“nothing on the face of the challenged measures [Prop. 115] ‘necessarily or inevitably’ compels” dire economic consequences predicted by petitioners in context of single-subject rule challenge].)

In Amador, we considered and rejected a similar revision challenge based on the predicted dire economic consequences to home rule in California arising from the property tax limitations of Proposition 13. We recognized the potential “limiting effect” on local government that would result from the substantial reduction in tax revenues contemplated by the measure, but we *511 concluded that such economic consequences were insufficient to accomplish a constitutional revision. ( 22 Cal.3d at p. 225 .)

Similarly, in Brosnahan v. Brown, supra, 32 Cal.3d at page 261 , we observed that “petitioners’ forecast of judicial and educational chaos is exaggerated and wholly conjectural, based primarily upon essentially unpredictable fiscal or budgetary constraints.” In the present case, petitioners’ prediction that Proposition 140 will “profoundly alter the Legislature’s ability to perform its constitutional functions,” seems similarly conjectural and speculative. The same objection disposes of petitioners’ suggestion, raised only in a footnote, that Proposition 140 may threaten to “impair essential governmental functions” (see Brosnahan v. Brown, supra, 32 Cal.3d at pp. 258-260), or to “radically undercut ] representative democracy” (see Amador, supra, 22 Cal.3d at pp. 227-228).

Petitioners assert that, from an historical standpoint, changes similar in nature to Proposition 140 have been achieved only through formal constitutional revisions. They note that in 1878, as a result of convening a constitutional convention, provisions were adopted that “reduced the Legislature from a full-time body to one that met only biennially and then was only allowed to be paid for sixty days. . . . That system prevailed until 1966, when the Constitution was revised a second time and the Legislature once again became a full-time body.” According to petitioners, any such changes “could only have been done as part of a revision.”

Of course, the fact that a particular constitutional change was adopted at a constitutional convention does not necessarily indicate the change was a revision. Mere amendments may also be adopted during the course of such proceedings. But the main difficulty with petitioners’ argument in this regard is that its acceptance could, as a practical matter, insulate the Legislature from any severe reform measures directed at that branch and initiated by the people, because proceedings to adopt constitutional revisions must be initiated by a two-thirds vote of both houses of the Legislature. (Cal. Const., art. XVIII, §§ 1, 2.) To hold that reform measures such as Proposition 140, which are directed at reforming the Legislature itself, can be initiated only with the Legislature’s own consent and approval, could eliminate the only practical means the people possess to achieve reform of that branch. Such a result seems inconsistent with the fundamental provision of our Constitution placing “[a]ll political power” in the people. (Id., art. II, § 1.) As that latter provision also states, “Government is instituted for [the people’s] protection, security, and benefit, and they have the right to alter or reform it when the public good may require.” (Italics added.)

It seems indisputable that Proposition 140 represents an attempt by the people to “alter or reform” their own government. To construe article XVIII *512 as vesting the Legislature with a power to veto such reform measures would be seriously inconsistent with the democratic principles expressed in article II. If, as petitioners predict, Proposition 140 ultimately produces grave, undesirable consequences to our governmental plan, the Legislature (Cal. Const., art. XVIII, § 1) or the people (id., art. XVIII, § 3) are empowered to propose a new constitutional amendment to correct the situation. Resolving, as we must, all doubts in favor of the initiative process, we conclude that nothing on the face of Proposition 140 effects a constitutional revision.

C. The Single-subject Requirement

Petitioners next contend that Proposition 140 violates the single-subject provision (Cal. Const., art. II, § 8, subd. (d)), by combining in a single measure such “disparate” subjects as term and budgetary limitations and pension restrictions. We disagree.

The principles that guide our resolution of the issue are well settled: “ ‘[A]n initiative measure does not violate the single-subject requirement if, despite its varied collateral effects, all of its parts are “reasonably germane" to each other,’ and to the general purpose or object of the initiative. [Citations.]” (Brosnahan v. Brown, supra, 32 Cal.3d at p. 245 ; see also Raven v. Deukmejian, supra, 52 Cal.3d at p. 346 , and cases cited.)

In the two foregoing cases, we upheld separate crime reform initiatives (Propositions 8 and 115) against single-subject challenges, even though each initiative contained widely disparate procedural and substantive provisions. We observed that the various elements of each of these initiatives united to form a comprehensive criminal justice reform package with one single unifying theme: promoting the rights of actual and potential crime victims. (See Raven v. Deukmejian, supra, 52 Cal.3d at p. 347; Brosnahan v. Brown, supra, 32 Cal.3d at p. 247 .)

In Brosnahan , we admonished that the single-subject rule “obviously forbids joining disparate provisions which appear germane only to topics of excessive generality such as ‘government’ or ‘public welfare.’ ” ( 32 Cal.3d at p. 253 .) We continued by referring to our “liberal interpretative tradition ... of sustaining statutes and initiatives which fairly disclose a reasonable and common sense relationship among their various components in furtherance of a common purpose.” (Ibid)

The unifying theme or common purpose of Proposition 140 is incumbency reform, a subject not excessively general when compared with prior measures upheld by this court. (See also Fair Political Practices Com. v. *513 Superior Court (1979) 25 Cal.3d 33, 43 [ 157 Cal.Rptr. 855 , 599 P.2d 46 ] [reasonably germane subject of “political practices”]; Amador, supra, 22 Cal.3d at p. 231 [valid subject of “property tax relief”]; cf. Harbor v. Deukmejian (1987) 43 Cal.3d 1078, 1097, 1100 [ 240 Cal.Rptr. 569 , 742 P.2d 1290 ] [invalid subjects of “fiscal affairs” and “statutory adjustments”].)

The present measure’s introductory statement declares that an “extremely high number of incumbents” are reelected by reason of “The ability of legislators to serve unlimited number of terms, to establish their own retirement system, and to pay for staff and support services at state expense . . . .” (Cal. Const., art. IV, § 1.5.) The declared purpose of Proposition 140 is “To restore a free and democratic system of fair elections, and to encourage qualified candidates to seek public office,” by limiting “the powers of incumbency.” (Ibid.)

Petitioners observe that Proposition 140 is broader than mere reform of the legislative branch, because the measure imposes term limitations on various constitutional officers as well as legislators. Moreover, petitioners dispute the idea that reducing or limiting legislators’ pension rights could effectively assist in achieving incumbency reform. In their view, reducing state contributions to the Legislators’ Retirement System would not eliminate any true advantage incumbents possess over challengers. Petitioners suggest the topic of pension reductions falls into the separate category or subject of “reducing expenditures.”

Petitioners appear to be confusing germaneness with functional relationship. As we have previously held, the single-subject provision does not require that each of the provisions of a measure effectively interlock in a functional relationship. (See Brosnahan v. Brown, supra, 32 Cal.3d at pp. 248-249.) It is enough that the various provisions are reasonably related to a common theme or purpose. (Id., at pp. 246-248; see Raven v. Deukmejian, supra, 52 Cal.3d at pp. 347-348.)

The framers of Proposition 140 evidently believed that “the powers of incumbency” could be reduced or checked by making an extended career in public office both less available and less attractive to incumbent legislators, through term and budgetary limitations as well as reduced pension benefits. Budgetary reductions may have been deemed necessary to reduce the advantages incumbents possess vis-k-vis other candidates, as discussed in part D, post. As for limited pension benefits, the measure expressly states that the limitation was appropriate because service in the Legislature is no longer “intended as a career occupation.” (Cal. Const., art. IV, § 4.5.) The framers presumably did not believe budgetary or pension limitations were needed *514 with respect to the other constitutional officers subject to the term limitations of the measure.

Whether or not these various provisions are wise or sensible, and will combine effectively to achieve their stated purpose, is not our concern in evaluating the present single-subject challenge. (See Amador, supra, 22 Cal.3d at p. 219 [“We do not consider or weigh the economic or social wisdom or general propriety of the [property tax] initiative. Rather, our sole function is to evaluate [the measure] legally in the light of established constitutional standards.”]; see also CalFarm Ins. Co. v. Deukmejian, supra, 48 Cal.3d at pp. 841-842 [single-subject rule does not require determination whether each section effectively will further the measure’s overall purpose]; In re Lance W., supra, 37 Cal.3d at p. 887 [wisdom of Prop. 8 irrelevant to constitutional challenge]; Brosnahan v. Brown, supra, 32 Cal.3d at p.248 [same].) Sensible or not, the separate aspects of Proposition 140 relate to the furtherance of a common purpose.

We conclude that the various provisions of Proposition 140 are reasonably germane to the single subject of incumbency reform.

D. Effect on Voting and Candidacy Rights

Petitioners next assert that the term limitations of Proposition 140 violate the First and Fourteenth Amendments of the federal Constitution. They observe that under Proposition 140, as previously discussed, once the prescribed maximum terms have expired, officeholders are forever barred from running for the office they held. According to petitioners, this lifetime ban substantially burdens two fundamental rights, namely, the right to vote and the right to be a candidate for public office. Petitioners, urging “strict scrutiny” of the new measure, suggest that no “compelling state interest” supports such a lifetime ban. (See Eu v. San Francisco Democratic Comm. (1989) 489 U.S. 214, 222 [ 103 L.Ed.2d 271, 281 , 109 S.Ct. 1013 ] [hereafter Eu]; Johnson v. Hamilton (1975) 15 Cal.3d 461, 466-468 [ 125 Cal.Rptr. 129 , 541 P.2d 881 ] [applying strict scrutiny to review constitutionality of candidates’ durational residence requirement].)

Respondents, on the other hand, assert the measure is valid under the balancing test announced in Anderson v. Celebrezze (1983) 460 U.S. 780 [ 75 L.Ed.2d 547 , 103 S.Ct. 1564 ], discussed below. Respondents, noting certain mitigating aspects of the measure, contend that the public policy served by Proposition 140 is both rational and compelling, having only minimal effects on voting or candidacy rights. We first turn to the question of the proper standard for resolving petitioners’ challenge.

*515 Because Proposition 140 amends the California Constitution itself, it is appropriate that we look to the decisions of the United States Supreme Court for guidance in determining the validity of the measure under the federal Constitution. We have observed that “[i]n analyzing constitutional challenges to election laws, this court has followed closely the analysis of the United States Supreme Court. [Citations.] [f] The high court has generally addressed the validity of election regulations under the equal protection clause. However, the basis for the court’s selection of that approach and the precise nature of the equal protection tests employed has not always been easily discernible. And, as is true for equal protection doctrine in general, the standard of review utilized in voting and election cases has been in a state of flux. [Citations.]” (Canaan v. Abdelnour (1985) 40 Cal.3d 703, 710-711 [ 221 Cal.Rptr. 468 , 710 P.2d 268 , 69 A.L.R.4th 915 ]; see also Munro v. Socialist Workers Party (1986) 479 U.S. 189, 200-201 [ 93 L.Ed.2d 499, 509-510 , 107 S.Ct. 533 ] [dis. opn. of Marshall, J., observing majority’s failure to articulate appropriate level of scrutiny for appraising validity of restrictions on ballot access].)

Petitioners assume that a “compelling interest” standard is required by Eu, supra, in which the justices unanimously invalidated California’s prohibition on primary endorsements by political parties. To the contrary, it seems clear the high court in Eu imposed such a strict standard because of the serious impact on First Amendment freedoms of speech and association, interests not as directly impacted by Proposition 140, which does not affect speech interests and which impacts all political parties on an equal basis. (See 489 U.S. at pp. 222-225 [103 L.Ed.2d at pp. 281-283].) As Eu concluded, “Because the ban [on endorsements] burdens appellees’ rights to free speech and free association, it can only survive constitutional scrutiny if it serves a compelling governmental interest.” (Id. at p. 225 [103 L.Ed.2d at P.283], fn. omitted.) Thus, Eu is probably inapposite. (See also Clements v. Fashing (1982) 457 U.S. 957, 963 [ 73 L.Ed.2d 508, 515-516 , 102 S.Ct. 2836 ] [plurality opinion stating barriers to candidate’s access to ballot do not compel close scrutiny]; Erum v. Cayetano (9th Cir. 1989) 881 F.2d 689, 692, fn. 7 [discussing Eu, but concluding that in cases involving restrictions on ballot access “heightened scrutiny is not the rule”].) In any event, as will appear, it is unlikely we would reach a different result applying strict scrutiny to evaluate Proposition 140.

In Rodriguez v. Popular Democratic Party (1982) 457 U.S. 1 [ 72 L.Ed.2d 628 , 102 S.Ct. 2194 ], the court considered the constitutional validity of Puerto Rico’s system of allowing only certain designated political parties access to the vote needed to fill legislative vacancies on an interim basis. In another unanimous opinion, the court noted that the federal Constitution does not specify the procedure a state must use in filling legislative *516 vacancies (id., at p. 8 [ 72 L.Ed.2d at p. 634 ]), that the Constitution does not compel a fixed method of choosing state or local officers or representatives (id., at p. 9 [ 72 L.Ed.2d at p. 635 ]), that the right to vote for state offices is not a constitutionally protected right (ibid. [ 72 L.Ed.2d at p. 635 ]), but that once a state or commonwealth provides for elected representatives, a citizen has a constitutional right to participate on an equal basis with other citizens (id., at p. 10 [ 72 L.Ed.2d at p. 635 ]).

Rodríguez v. Popular Democratic Party, supra, 457 U.S. 1 , indicates the high court will give wide latitude to state election laws, even those that may restrict the electorate’s choice of representatives, so long as those laws are applied in an even-handed manner without discriminating against particular citizens or classes of citizens. The incumbency limitations involved here satisfy that standard. (See also Burdick v. Takushi (9th Cir. 1991) 927 F.2d 469 , 473-474 [voters have no constitutional right to vote for particular candidate]; Stiles v. Blunt (8th Cir. 1990) 912 F.2d 260, 266, fn. 10 , cert. den. (1991) _ U.S. _ [ 113 L.Ed.2d 241 , 111 S.Ct. 1307 ] [same]; Zielasko v. State of Ohio (6th Cir. 1989) 873 F.2d 957, 961 [same]; but see Caanan v. Abdelnour, supra, 40 Cal.3d at p. 713 .)

In Canaan v. Abdelnour, supra, 40 Cal.3d 703 , we struck down a city’s blanket prohibition against write-in voting in municipal elections. In so doing, we applied the balancing test set forth in Anderson v. Celebrezze, supra, 460 U.S. 780 , wherein the court held unduly burdensome an Ohio law requiring independent candidates for the November 1980 Presidential Election to file their statements of candidacy by March of that year. The high court in Anderson , acknowledging that a state’s regulatory interests in determining the eligibility of candidates “are generally sufficient to justify reasonable, nondiscriminatory restrictions” (id., at p. 788 [ 75 L.Ed.2d at p. 557 ], fn. omitted), announced the following test:

With respect to Anderson's requirement of showing the “necessity” of the particular burden imposed by the state, we must also consider whether there are any less drastic alternatives to a lifetime ban. (See Anderson v. Celebrezze, supra, 460 U.S. at p. 806 [ 75 L.Ed.2d at p. 568 ]; Canaan v. Abdelnour, supra, 40 Cal.3d 703, 719, fn. 13 .)

Thus, Anderson v. Celebrezze, supra, 460 U.S. 780 , requires us to consider three separate elements in ascertaining the constitutionality of state laws restricting access to the ballot: (1) the nature of the injury to the rights affected, (2) the interests asserted by the state as justifications for that injury, and (3) the necessity for imposing the particular burden affecting the plaintiff’s rights, rather than some less drastic alternatives. Lacking any more specific guidance from the high court, we now apply Anderson's balancing test to the challenged provisions of Proposition 140.

1. Character and Extent of Injury to Protected Rights

Two important rights are affected by Proposition 140, namely, the incumbent’s right to run for public office, and the voters’ right to reelect the incumbent to that office. Consequently, the “injury” to those rights resulting from the application of Proposition 140 is also twofold, namely, lifetime exclusion of the incumbent from the office previously held, and a corresponding permanent inability of the voters to return the incumbent to that office.

a. Effect on Candidates

As previously explained, Proposition 140 imposes a lifetime ban on legislators once they have completed the maximum number of terms. Petitioners argue, “In the long run, the term limitations permanently ban those who are arguably the most qualified candidates—incumbents with the experience and expertise in the legislative process necessary to the most effective representation of their constituencies.” According to petitioners, qualified incumbents will be “purged” solely to seat “massive numbers” of inexperienced “newcomers.” Petitioners predict that only a few qualified persons will be attracted to short term public office.

Respondents, of course, dispute petitioners’ premise that long-term legislators are inevitably better qualified than other candidates, and they believe that term limitations will encourage, rather than inhibit, new qualified *518 candidates seeking short term public service. They characterize the term limitations of Proposition 140 as additional candidacy requirements, akin to age, integrity, training or residency, which have generally been upheld. (See, e.g., Zeilenga v. Nelson (1971) 4 Cal.3d 716, 721 [ 94 Cal.Rptr. 602 , 484 P.2d 578 ]; 25 Am.Jur.2d, Elections, § 175, pp. 870-871, and cases cited; Note, Developments in the Law: Elections (1975) 88 Harv.L.Rev. 1111, 1217 et seq.)

Respondents also stress three features of Proposition 140 that assertedly serve to mitigate the severity of its lifetime ban: First, the affected incumbent is not barred from seeking any other public office, including a seat in another legislative house or a statewide constitutional office. A former Senator may seek a seat in the Assembly, and vice versa. Second, the term limitations arise only after the incumbent already has had the opportunity to serve a significant period in office (i.e., eight years for a Senator, and six years for a member of the Assembly). Finally, the term limitations are generally applicable to persons elected or appointed on or after November 6, 1990. Except for some incumbent Senators, past terms served do not count in calculating the limitation. Thus, by the time the term limitations of Proposition 140 come into play, the incumbent will have already served, and indeed may continue to serve, several terms in public office.

b. Effect on Voters

Petitioners also stress the impact on the voters who are prevented from casting their ballots for the particular candidate of their choice. Just as incumbent legislators are permanently barred from running for another term once they have served the prescribed numbers of terms, the voters are permanently barred from voting for such persons, at least for the legislative office they once held. According to petitioners, the voters thus will be denied the right to vote for those persons who arguably possess the best qualifications.

Additionally, petitioners note that because Proposition 140 was adopted on a statewide basis, “the disability [on candidates and voters] is imposed not by those who have the right to vote for the candidate, but rather by those outside the district.” Petitioners thus suggest the resulting impact on or injury to the voters is aggravated or enhanced by reason of the ability of voters residing outside a particular voting district to essentially “veto” particular candidates within that district.

Respondents reply by citing federal court cases (e.g., Burdick v. Takushi, supra, 927 F.2d 469,473-474 ) stating that voters have no constitutional right to vote for particular candidates. Additionally, respondents observe that the *519 challenged measure does not fall into any of the categories of prior cases in which the right to vote was found impermissibly infringed. No identifiable groups of voters are excluded from voting or otherwise unduly burdened in the exercise of their franchise. Characterizing the term limitations of Proposition 140 as additional candidacy qualifications akin to age or residency, respondents submit that Proposition 140 does not truly impair the franchise, for the voters retain the basic fundamental right to cast their ballots for the qualified candidate of their choice.

Moreover, respondents observe that neither voter choice nor candidate eligibility is restricted based on the content of protected expression, political affiliation, or inherently arbitrary factors such as race, religion or sex. The only criterion used is incumbency. Voters retain the ability to vote for any qualified candidate holding the beliefs or possessing the attributes they may desire in a public officeholder. Under these circumstances, First Amendment protection of political expression and promotion of the marketplace for ideas continue unabated.

Respondents further note that petitioners have cited no case supporting their theory that a voting restriction on “local” offices would be invalid if imposed by voters on a statewide basis. Indeed, such a rule would seemingly call in question any statewide legislation affecting the qualification of candidates for local elections, such as age or residency requirements.

Finally, respondents suggest that because Proposition 140 was an initiative measure adopted by the people at a statewide election, any resulting injury to the exercise of the franchise should be deemed self-inflicted, and thus not constitutionally protected.

c. Summary of Impact on Candidates and Voters

In sum, although Proposition 140 does affect the rights of voters and candidates to a degree, there are several mitigating aspects, including the voters’ continued right to vote for any qualified candidates, as well as the candidates’ ability to run for other public offices, their entitlement to a significant period of service in office before the term limitations apply, and the “prospective” application of the limitation provision. Additionally, we should bear in mind that it is presently unclear under federal law whether and to what extent voters retain a constitutional right to vote for particular candidates such as the incumbent legislators affected by the challenged measure. Thus, the legal impact of Proposition 140 on the voters remains uncertain.

Having discussed the extent of the “asserted injury to the rights protected” (Anderson v. Celebrezze, supra, 460 U.S. at p. 789 [ 75 L.Ed.2d at p. 558 ]), *520 we next analyze the “precise interests put forward by the State as justifications for the burden imposed by its rule” (ibid.).

2. The Interests of the State

Balanced against the foregoing negative impact on candidates and voters flowing from the challenged measure are the considerable state interests assertedly promoted thereby. In the words of new article IV, section 1.5, of the state Constitution, term limitations are deemed necessary to restore “free, fair, and competitive elections,” to “encourage qualified candidates to seek public office,” and to eliminate “unfair incumbent advantages” that have resulted in an “extremely high number of incumbents” and created “a class of career politicians” instead of the “citizen representatives envisioned by the Founding Fathers.”

According to respondents, the state’s interests in limiting incumbency should support measures considerably stronger than a mere temporary disability from holding office. As respondents argue, the state’s strong interests in protecting against an entrenched, dynastic legislative bureaucracy, and in thereby encouraging new candidates to seek public office, are both legitimate and compelling ones that support a lifetime ban from the office and outweigh any interest the incumbent legislators, or the voting public, may have in perpetuating the incumbents’ positions of control.

The legitimacy of the foregoing asserted state interests in limiting incumbency are well recognized in analogous contexts. As stated by the West Virginia Supreme Court of Appeals in rejecting a similar challenge to a state constitutional amendment limiting the right of the Governor to seek a third consecutive term, “Constitutional restrictions circumscribing the ability of incumbents to succeed themselves appear in over twenty state constitutions, and exist in the Twenty-second Amendment to the Constitution of the United States with regard to the Presidency. The universal authority is that restriction upon the succession of incumbents serves a rational public policy and that, while restrictions may deny qualified men an opportunity to serve, as a general rule the over-all health of the body politic is enhanced by limitations on continuous tenure. [Citations and fn. omitted].” (State ex rel. Maloney v. McCartney (1976) 159 W.Va. 513 [ 223 S.E.2d 607, 611 ] [hereafter Maloney], app. disn. sub nom. Moore v. McCartney (1976) 425 U.S. 946 [ 48 L.Ed.2d 190 , 96 S.Ct. 1689 ]; see Maddox v. Fortson (1970) 226 Ga. 71 [ 172 S.E.2d 595, 598-599 ], cert. den. 397 U.S. 149 [ 25 L.Ed.2d 183 , 90 S.Ct. 999 ]; cf. Chemerinsky, Protecting the Democratic Process: Voter Standing to Challenge Abuses of Incumbency (1988) 49 Ohio St. L.J. 773 et seq.; Tribe, American Constitutional Law (2d ed. 1988) § 13-18, at p. 1097 [“Democracy envisions rule by successive temporary majorities. The capacity to displace *521 incumbents in favor of the representatives of a recently coalesced majority is, therefore, an essential attribute of the election system in a democratic republic.”]; cf. Annot. (1958) 59 A.L.R.2d 716 [construction and effect of incumbency limitation laws].)

The Maloney decision continues by describing at length the substantial reasons for limiting the right of incumbents to succeed themselves. These include “The power of incumbent officeholders to develop networks of patronage and attendant capacities to deliver favorably disposed voters to the polls,” “fears of an entrenched political machine which could effectively foreclose access to the political process,” and the belief that regularly disrupting those “machines” “would stimulate criticism within political parties” and “insure a meaningful, adversary, and competitive election.” ( 223 S.E.2d at p. 611 .)

In addition, Maloney explains that “it has long been felt that a limitation upon succession of incumbents removes the temptation to prostitute the government to the perpetuation of a particular administration. [Citation.] . . . Meretricious policies which sacrifice the well-being of economic, social, racial, or geographic minorities are most likely where a political figure, political party, or political interest group can rely upon electorate inertia fostered by the hopelessness of encountering a seemingly invincible political machine.” (223 S.E.2d at pp. 611-612.)

Petitioners observe that Maloney involved a limitation on consecutive terms of a Governor, rather than a lifetime ban on incumbent legislators. They suggest that term limitations on the executive branch are justified by the need to check the substantial concentration of power that the chief executive possesses, a consideration assertedly not applicable to the legislative branch. But we think that many, if not all, of the considerations mentioned in Maloney (e.g., eliminating unfair incumbent advantages, dislodging entrenched political machines, restoring open access to the political process, and stimulating electorate participation) would apply with equal force to the legislative branch.

In connection with petitioners’ argument that Proposition 140’s lifetime ban is unconstitutional, two other cases are instructive, though factually distinguishable. In Clements v. Fashing, supra, 457 U.S. 957 , cited with apparent approval in Anderson v. Celebrezze, supra, 460 U.S. at page 789 , footnote 9 [ 75 L.Ed.2d at page 558 ], the high court upheld the validity of a Texas statute that rendered incumbent justices of the peace ineligible for the Texas Legislature. The disability from office extended only during the term for which the justices were elected or appointed. A plurality of the high court took the position that barriers to a candidate’s access to the ballot do not *522 compel close scrutiny (457 U.S. at p. 963 [ 73 L.Ed.2d at p. 516 ]), and stressed the “de minimis” nature of the restriction, noting that the act merely imposed a brief “waiting period” on current officeholders, and therefore could be sustained by a mere showing of some “rational predicate” to support it (id. at pp. 967-968 [73 L.Ed.2d at pp. 518-519]).

The Clements v. Fashing plurality did not affirmatively indicate that a lifetime bar to legislative service necessarily would be invalid. Significantly, unlike Proposition 140, the Texas act apparently was not aimed at limiting the powers of incumbency, but was based on the “rational predicate” that an affected justice will be less inclined to abuse his position or neglect his duties because of the justice’s aspirations for higher office. (457 U.S. at p. 968 [ 73 L.Ed.2d at p. 519 ].)

In De Bottari v. Melendez (1975) 44 Cal.App.3d 910 [ 119 Cal.Rptr. 256 ], the Court of Appeal struck down a local ordinance prohibiting recalled council members from running for city council within a year of the recall. Although petitioners believe the case supports their position, closer scrutiny indicates otherwise. Finding the candidacy restriction too severe, the De Bottari court observed, “Cases in other jurisdictions upholding limitations on successive terms in office [citations] involve similar restrictions but are not authoritative here since such limitations serve totally different governmental interests.” ( 44 Cal.App.3d at p. 913, fn. 1 , italics added.) De Bottari , using strict scrutiny, reviewed the interests that assertedly supported a temporary ban on candidacy by recalled candidates and found them insufficient to sustain the restriction. The court had no occasion to review the “different” interests served by general limitations on incumbency, as outlined by Maloney, supra.

In sum, despite its distinguishing features, we conclude that Maloney's (supra, 223 S.E.2d 607 ) analysis is quite pertinent to our determination whether permanent incumbency limitations are supported by legitimate and compelling considerations. We conclude they are so supported.

3. Necessity of Imposing Restrictions

We turn next to the “necessity” of imposing the restrictions of Proposition 140 on the dual rights at issue here (see Anderson v. Celebrezze, supra, 460 U.S. at pp. 789-790 [75 L.Ed.2d at pp. 558-559]). Petitioners contend that a lifetime ban on candidacy was unnecessary, and that other less “drastic” alternatives, such as a limitation on consecutive terms, together with additional restrictions on campaign contributions to legislators, decreased fringe and pension benefits, and additional incentives for early retirement, would *523 have been sufficient to promote and accomplish the state interests previously discussed.

As will appear, we conclude that the less drastic alternatives suggested by petitioners would have been inadequate to accomplish the declared purpose of Proposition 140 to eliminate the “class of career politicians” that assertedly had been created by virtue of the “unfair incumbent advantages” referred to in that measure. (Cal. Const., art. IV, § 1.5.)

Respondents stress the substantial advantages incumbent legislators enjoy in this state, advantages that permitted 92 percent of all incumbents to win reelection at this state’s November 1990 General Election. Indeed, respondents note that nine of these incumbents ran unopposed.

Respondents seem correct in this regard. Whether by reason of superior fund raising ability, greater media coverage, larger and more experienced staffs, greater name recognition among the voters, favorably drawn voting districts, or other factors, incumbents do indeed appear to enjoy considerable advantages over other candidates. (See Service Employees v. Fair Political Practices (E.D.Cal. 1990) 747 F.Supp. 580, 588 ; Watson v. Fair Political Practices Com. (1990) 217 Cal.App.3d 1059,1074 [ 266 Cal.Rptr. 408 ], & fn. 13.) As Proposition 140’s introductory statement indicates, the framers of the measure believed these substantial advantages of incumbency were unfair to other candidates and tended to create “a class of career politicians, instead of the citizen representatives envisioned by the Founding Fathers.” (Cal. Const., art. IV, § 1.5.)

Petitioners suggest that a more reasonable alternative existed to the measure’s lifetime ban: disqualification of the incumbent for the forthcoming term, thus “forcing the legislator to take one term off, before being eligible to run for the body.” Yet, as respondents observe, the framers reasonably could conclude that a lifetime ban was necessary to assure that a former officeholder could not reinvoke at least some of the advantages of incumbency to gain reelection after leaving office for a term or more.

Additionally, we believe the framers might well have reasonably concluded that a mere ban on consecutive terms could encourage popular “career politicians” to trade terms with each other, or to attempt to arrange for a “caretaker” candidate, such as a spouse or relative, to hold office for them during the interrupted term. For example, when in 1966 George Wallace became legally ineligible to run for reelection as Governor of Alabama because of state term limitations applicable to that office, his wife, Lurleen, successfully ran in his place, and served as Governor until her death in 1968. George Wallace was reelected as Governor in 1970 and again in *524 1974. (See 12 The New Encyclopedia Britannica, Micropaedia (15th ed. 1990) Wallace, George C., at p. 467.)

Realistically, only a lifetime ban could protect against various kinds of continued exploitation of the “advantages of incumbency” captured through past terms in office. The remainder of petitioners’ suggested “alternatives” essentially involve narrow changes in the system of providing contributions or compensation for legislators, changes that would afford “career politicians” with independent resources little incentive to voluntarily terminate public service.

4. Conclusion

On balance, we conclude the interests of the state in incumbency reform outweigh any injury to incumbent officeholders and those who would vote for them. As Maloney observed ( 223 S.E.2d at p. 612 ), no decisions of the United States Supreme Court have been found that suggest a limitation on incumbency would be unconstitutional. Although such limitations may restrict the franchise (but see cases indicating voters have no right to vote for particular candidates, e.g., Burdick v. Takushi, supra, 927 F.2d 469, 473-474 ), if we use a balancing test that weighs “the enlargement of the franchise by guaranteeing competitive primary and general elections” against “incidental disenfranchisement” of some voters, the court “must conclude that restrictive provisions on the succession of incumbents do[ ] not frustrate but rather further[ ] the policy of the Fourteenth Amendment.” ( Maloney, supra, 223 S.E.2d at pp. 612-613.)

It is true, as petitioners observe, that respondents have not offered evidence to support all of the various premises on which Proposition 140 is based. But as the United States Supreme Court pointed out in Munro v. Socialist Workers Party, supra, 479 U.S. 189 , upholding state restrictions on minority party candidacy, a state need not demonstrate empirically all of the various evils that its regulations seek to combat: “Such a requirement would necessitate that a State’s political system sustain some level of damage before the legislature could take corrective action. Legislatures, we think, should be permitted to respond to potential deficiencies in the electoral process with foresight rather than reactively, provided that the response is reasonable and does not significantly impinge on constitutionally protected rights.” (Id., at pp. 195-196 [ 93 L.Ed.2d at p. 506 ].) We have no reason to doubt that the high court would give similar leeway to state regulations imposed directly by the people through the initiative process.

In sum, it would be anomalous to hold that a statewide initiative measure aimed at “restor[ing] a free and democratic system of fair elections,” and *525 “encouraging] qualified candidates to seek public office” (Cal. Const., art. IV, § 1.5), is invalid as an unwarranted infringement of the rights to vote and to seek public office. We conclude the legitimate and compelling interests set forth in the measure outweigh the narrower interests of petitioner legislators and the constituents who wish to perpetuate their incumbency.

E. Bill of Attainder

Petitioners next assert that Proposition 140 is, in effect, an unlawful bill of attainder. (See U.S. Const., art. I, § 10 [“No state shall. . . pass any bill of attainder . . . .”].) They contend that although the legislative term, budgetary, and pension limitations of Proposition 140 are facially applicable to all current and future incumbent legislators, the measure was nonetheless intended to punish or penalize the current group of incumbent legislators, and especially such long-term incumbents as Assemblyman Willie Brown (elected in 1964), and Senator David Roberti (elected in 1971).

A bill of attainder has been defined as a “legislative punishment, of any form or severity, of specifically designated persons or groups” (United States v. Brown (1965) 381 U.S. 437, 447 [ 14 L.Ed.2d 484, 491 , 85 S.Ct. 1707 ] [federal law forbidding Communist Party members from serving as union officers], italics added), or a “legislative act[ ] . . . that applies] either to named individuals or to easily ascertainable members of a group in such a way as to inflict punishment on them without a judicial trial” (United States v. Lovett (1946) 328 U.S. 303, 315 [ 90 L.Ed. 1252, 1259 , 66 S.Ct. 1073 ] [federal legislation terminating compensation for certain federal employees accused of subversive activities], italics added). (See also Selective Service v. Minn. Public Int. Res. Gp. (1984) 468 U.S. 841, 851 [ 82 L.Ed.2d 632, 643 , 104 S.Ct. 3348 ] [bills of attainder involve punitive action aimed at identifiable person or group]; Nixon v. Administrator of General Services (1977) 433 U.S. 425, 474 [ 53 L.Ed.2d 867, 910-911 , 97 S.Ct. 2777 ]; Cummings v. Missouri (1867) 71 U.S. (3 Cranch.) 277, 320 [ 18 L.Ed. 356, 362 ] [“disqualification from office may be punishment” amounting to an improper bill of attainder]; Crain v. City of Mountain Home, Ark. (8th Cir. 1979) 611 F.2d 726 , 729 [invalidating city ordinances that removed city attorney from office for remainder of term, reduced future salary for the office, and prohibited private practice by such officers]; Connecticut Judicial Selection Com’n v. Larson (D.Conn. 1989) 745 F.Supp. 88, 94-96 [upholding act shortening terms of incumbent nonattomey members of judicial selection committee].)

Petitioners point to the following language in the ballot arguments by the measure’s proponents: “Proposition 140 will end the reign of the Legislature’s powerful officers—the Assembly Speaker (first elected a *526 quarter of a century ago) and the Senate Leader (now into his third decade in the Legislature). . . .” (Ballot Pamp., supra, at p. 70.) According to petitioners, the foregoing language (together with more explicit references to Brown and Roberti in the proponents’ campaign literature) makes it clear that Proposition 140 was directed at punishing such long-term legislators as Brown and Roberti, and that accordingly the measure must be deemed an improper bill of attainder under the foregoing cases.

We conclude that petitioners’ argument should be rejected. As the high court explained in Nixon v. Administrator of General Services, supra, 433 U.S. 425 , in determining whether particular legislation constitutes a bill of attainder, courts have applied three different tests. We believe the challenged measure meets none of them.

First, an “historical” test has been used to determine whether the subject legislation imposes a kind of punishment traditionally deemed prohibited by the federal Constitution. ( 433 U.S. at p. 475 [ 53 L.Ed.2d at p. 911 ].) Second, the courts have used a “functional test of the existence of punishment, analyzing whether the law under challenge, viewed in terms of the type and severity of burdens imposed, reasonably can be said to further nonpunitive legislative purposes. [Citations.]” (Id., at pp. 475-476 [ 53 L.Ed.2d at p. 911 ], italics added and fa. omitted.) Finally, the courts have used a “motivational” test, “inquiring whether the legislative record evinces a congressional intent to punish. [Citations.]” (Id., at p. 478 [ 53 L.Ed.2d at p. 913 ].)

Petitioners appear to concede that the historical test is inapplicable here, and we have found no cases involving term limitations as a form of “punishment.” As for the functional test, the measure itself expresses broad, nonpunitive purposes, namely, “[t]o restore a free and democratic system of fair elections, and to encourage qualified candidates to seek public office” by limiting “the powers of incumbency.” (Cal. Const., art. IV, § 1.5.)

Although petitioners point to the “self-serving” nature of these declarations of intent, we have no reason to dispute the accuracy of their description of the measure’s primary intent. As in Nixon v. Administrator of General Services, supra, 433 U.S. at page 477 [ 53 L.Ed.2d at page 912 ], “Evaluated in terms of these [express congressional] asserted purposes, the law plainly must be held to be an act of nonpunitive legislative policymaking.” (Italics added.)

Petitioners rely primarily on the motivational test, stressing the framers’ express intent to dislodge such long-term incumbents as Brown and Roberti. Petitioners cite the ballot pamphlet arguments of the measure’s proponents, which describe these legislators in unflattering terms (“legislative dictators” *527 participating in “Sacramento’s web of special favors and patronage”), and urge passage of the measure as a means of ending their long incumbency.

But the ballot arguments contain no indication of an intent to punish those individuals for any particular past misconduct. Broad reform measures are frequently prompted by particular acts or circumstances involving specific individuáis, but in our view such measures would not constitute improper bills of attainder unless an intent to punish such individuals clearly appears from their face, or from the circumstances surrounding their passage.

As stated in Connecticut Judicial Selection Com’n v. Larson, supra, 745 F.Supp. at page 96 , quoting from an earlier case, “ ‘[i]t is clear that general legislation such as this . . . aimed at the office . . . rather than the incumbent office holders, has none of the objectionable attributes of a bill of attainder.’ [Citation.]” (Italics added.) Proposition 140 applies with equal force to all state legislators, current and future. Although it is unquestionable the proponents sought to limit the terms of incumbent legislators, including such long-term legislators as Brown and Roberti, we find no evidence of an intent to single out and punish those individuals for any supposed misconduct on their part. Accordingly, we conclude that Proposition 140 does not constitute a bill of attainder under the federal Constitution.

F. Impairment of Contracts

Finally, petitioners maintain that Proposition 140’s limitations on the pension rights of incumbent legislators is unconstitutional as an invalid impairment of contract under the federal Constitution. (U.S. Const., art. I, §10 [“No State shall . . . pass any . . . law impairing the obligation of contracts . . . .”].)

As previously indicated, new section 4.5 is added to article IV of the state Constitution to provide that although the state will contribute the employer’s share to the federal Social Security program on behalf of participating legislators “elected to or serving in the Legislature on or after November 1, 1990,” nonetheless “[n]o other pension or retirement benefit shall accrue as a result of service in the Legislature, such service not being intended as a career occupation.”

This same provision further provides that “This Section shall not be construed to abrogate or diminish any vested pension or retirement benefit which may have accrued under an existing law . . . , but upon adoption of *528 this Act no further entitlement to nor vesting in any existing program shall accrue to any such person, other than [federal] Social Security . . .

Petitioners assert that, as to incumbent legislators continuing to serve on or after November 6, 1990 (either by reelection to a new term or continuation of a prior term), Proposition 140 constitutes an improper impairment of their “vested” contractual or statutory right to continued participation in the pension program administered by the Legislators’ Retirement System (LRS). (See Gov. Code, § 9350 et seq.) In petitioners’ view, supported by respondent PERS, incumbent legislators who were members of the LRS immediately prior to November 6, 1990, are entitled to continue to earn pension benefits so long as they remain otherwise qualified to receive such benefits through continued state service. Petitioners (not joined in this respect by PERS) assert similar contract claims on behalf of nonincumbent legislators newly elected on November 6.

1. Incumbent Legislators

Petitioners’ find ample support for their position in California cases confirming that both the federal and state contract clauses protect the vested pension rights of public officers and employees from unreasonable impairment. (See Allen v. Board of Administration (1983) 34 Cal.3d 114, 119-120, 124 [ 192 Cal.Rptr. 762 , 665 P.2d 534 ] [legislators]; Olson v. Cory (1980) 27 Cal.3d 532, 540-541 [ 178 Cal.Rptr. 568 , 636 P.2d 532 ] [judges); Betts v. Board of Administration (1978) 21 Cal.3d 859, 863-864 [ 148 Cal.Rptr. 158 , 582 P.2d 614 ] [state Treasurer]; Miller v. State of California (1977) 18 Cal.3d 808, 814-817 [ 135 Cal.Rptr. 386 , 557 P.2d 970 ] [public employees]; see also Allen v. City of Long Beach (1955) 45 Cal.2d 128, 131 [ 287 P.2d 765 ] [same]; Kern v. City of Long Beach (1947) 29 Cal.2d 848, 852-853 [ 179 P.2d 799 ] [same].)

Petitioners (joined by respondent PERS) argue that when incumbent legislators first assumed office, they were impliedly promised pension benefits substantially equivalent to those offered by the then-existing provisions of the LRS, and that these benefits included both the primary right to receive any vested pension benefits upon retirement (see City of Long Beach v. Allen (1956) 143 Cal.App.2d 35, 38-39 [ 300 P.2d 356 ]), as well as the collateral right to earn future pension benefits through continued service, on terms substantially equivalent to those then offered (see Carman v. Alvord (1982) 31 Cal.3d 318, 325 [ 182 Cal.Rptr. 506 , 644 P.2d 192 ]). Although Proposition 140 by its terms assures that incumbent legislators’ right to receive vested pension benefits will not be lost or diminished, the measure also purports to terminate their collateral right to accrue additional benefits through further contributions and continued service to the state.

*529 According to petitioners, the applicable principle is set forth in Miller v. State of California, supra, 18 Cal.3d at page 817 : “[U]pon acceptance of public employment plaintiff acquired a vested right to a pension based on the system then in effect. That system allowed him to earn successively higher levels of benefits based on his years of service and his highest average salary during three consecutive years.” (See also Olson v. Cory, supra, 27 Cal.3d at p. 541 ; Betts v. Board of Administration, supra, 21 Cal.3d at p. 863 ; Kern v. City of Long Beach, supra, 29 Cal.2d at pp. 855-856.) Similar increased benefits from continued service were provided legislators under the LRS. (See Gov. Code, § 9350.)

In response to petitioners’ assertions, respondent Eu and intervener first contend that incumbent legislators do not have a vested right under the LRS to continue to accrue pension benefits through continued service. In this regard, they suggest that article IV, section 4, of the state Constitution precludes legislators from acquiring any vested right to continue to earn pension benefits. We disagree.

Article IV, section 4 of the state Constitution, provides in pertinent part that “The Legislature may, prior to their retirement, limit the retirement benefits payable to Members of the Legislature . . . .” (Italics added.) That provision, seemingly empowering the Legislature to exercise some measure of control over the pension rights of its own members prior to their retirement, may create some uncertainty as to the full amount or extent of a legislator’s pension rights during his term of office. But the provision neither states nor implies that these rights are thus deemed inchoate and unprotected from impairment by the initiative process. Significantly, we have never suggested that the mere existence of article IV, section 4, precludes legislators from acquiring pension rights protected by the state or federal contract clauses. (Cf. Allen v. Board of Administration, supra, 34 Cal.3d at pp. 119120.)

Petitioners acknowledge that the state as employer is permitted to make reasonable modifications to the pension system during the employment relationship, so long as employees receive “comparable new advantages” in return for any substantial reduction in benefits. (Olson v. Cory, supra, 27 Cal.3d at p. 541; Betts v. Board of Administration, supra, 21 Cal.3d at p. 864 ; Allen v. City of Long Beach, supra, 45 Cal.2d at p. 131 .) As we stated in Olson , “Although an employee does not obtain any ‘absolute right to fixed or specific benefits . . . there [are] strict limitation^] on the conditions which may modify the pension system in effect during employment.’ [Citation.] Such modifications must be reasonable and any ‘ “changes in a pension plan *530 which result in disadvantage to employees should be accompanied by comparable new advantages.” ’ [Citation.]” ( 27 Cal.3d at p. 541 .)

Petitioners correctly observe that the pension provisions of Proposition 140 do not merely modify the LRS pension system; rather they terminate that system entirely as to additional benefits accruing for future services. Intervener argues that the “transfer” or “redirection” of pension funds to the federal Social Security system operates as a “comparable new advantage” which justifies the impairment and consequently sustains the measure. Petitioners respond by asserting that every legislator already possessed the right to join the federal Social Security system, that 99 out of 120 legislators were already contributing to that system when Proposition 140 was adopted, and that the anticipated federal benefits will be far less than those provided by the LRS. Neither respondent Eu nor intervener has disputed those allegations.

We conclude that incumbent legislators had a vested right to earn additional pension benefits through continued service, despite the potential but unexercised limitations contemplated by article IV, section 4, of the state Constitution. We also conclude that the pension restriction of Proposition 140 must be deemed an impairment, not a mere “modification” or “adjustment,” of the vested pension rights of incumbent legislators.

Respondent Eu and intervener next argue that, assuming an incumbent legislator may acquire such vested rights, a different rule should apply to incumbent legislators who commenced new terms after Proposition 140 became effective. They rely primarily on our decision in Olson v. Cory, supra, 27 Cal.3d 532 , which had upheld a statute limiting certain cost-of-living increases in judges’ salaries after they commenced new terms of office. As Olson held, “A judge who completes one term during which he was entitled to unlimited cost-of-living increases and elects to enter a new term has impliedly agreed to be bound by salary benefits then offered by the state for the different term.” ( 27 Cal.3d at p. 540 .) Similarly, it is argued, incumbent legislators commencing new terms on or after November 6,1990, have impliedly agreed to serve under the conditions mandated by the new measure.

Petitioners and respondent PERS contend, however, that pension rights fall into a different category than salary rights. Although a state officer may have no protectible right to continuation of his former salary from term to term, nonetheless on commencing to serve the state the officer thereupon acquires a vested right to earn, through continued service, additional pension benefits in an amount reasonably comparable to those available when he or she first took office. According to petitioners, that right is not extinguished *531 when one term of office ends and another commences. (See In re Marriage of Alarcon (1983) 149 Cal.App.3d 544, 552-553 [ 196 Cal.Rptr. 887 ] [hereafter Alarcon].) To hold otherwise, petitioners argue, could unduly penalize public officers who serve elective or appointive terms.

We agree with petitioners. It would be anomalous to hold that a legislator’s supposedly “vested” right to acquire a reasonable pension is subject to absolute divesting at the conclusion of the period comprising his or her immediate term of office (assuming no right to receive actual benefits has yet accrued). Under such a holding, a legislator could entirely lose future pension benefits on the very eve of entitlement thereto.

The state, in adopting the LRS, elected to treat legislators in a manner similar to other state employees, providing them with a plan whereby pension benefits could be acquired through continued service beyond their initial term of office. Having done so, the state cannot thereafter abandon that plan as to incumbent legislators without providing them comparable new benefits.

As stated in Alarcon, “The state’s obligation to pay a specific salary extends to the end of each term of office. There is no promise express or implied the state will continue to pay an existing salary beyond the end of a term. Thus, no contractual rights are impaired if different salary or cost of living differentials are payable during and for a future term. []Q A pension, however, is different from a salary. A right to pension benefits provided by the state payable upon fulfillment of age, service and other requirements may not be destroyed, once vested, without impairment of the state’s contractual obligation. [Citations.]” ( 149 Cal.App.3d at p. 552 .)

Alarcon is factually distinguishable, for in that case the affected officer (a state judge who later accepted a federal judgeship) had already acquired a vested, though deferred, right under state law to a full pension prior to the enactment of the challenged restriction on pension benefits for persons accepting federal positions. Under such circumstances, the fact the judge served additional terms as a state judge following the adoption of that restriction could not defeat his previously vested pension rights.

Nonetheless, the principle of Alarcon seems controlling here. In the present case, incumbent legislators acquired a vested right to earn increased pension benefits in return for continued service. By analogy to Alarcon, their commencement of a new term following adoption of the pension restriction of Proposition 140 would not defeat that vested right.

Alarcon further observed that although vested pension rights could be modified before retirement “to keep a pension system flexible and permit *532 adjustments to accord with changing conditions, such modifications must be reasonable and may not destroy or impair a vested contractual right to a pension. [Citations.]” ( 149 Cal.App.3d at p. 553 .) As we have previously discussed, the pension provisions of Proposition 140, which abruptly terminate an incumbent legislator’s right to earn future pension benefits through continued service, must be deemed an impairment, not a mere “modification” or “adjustment,” of the vested pension rights of incumbent legislators, whether or not they will enter a new term on or after November 6, 1990.

Thus, we hold that, under California law, all incumbent legislators acquired a vested right to earn additional pension benefits through continued service, a right which Proposition 140 clearly impairs.

After granting review in this case, we asked for further briefing on the additional question whether the contract clause of the federal Constitution protects the salary or pension rights of state public officials such as legislators. (See, e.g., Higginbotham v. Baton Rouge (1939) 306 U.S. 535, 538-539 [ 83 L.Ed. 968, 971-972 , 59 S.Ct. 705 ], rehg. den. 307 U.S. 649 [ 83 L.Ed. 968 , 59 S.Ct. 705 ] [hereafter Higginbotham]; Dodge v. Board of Education (1937) 302 U.S. 74, 78-79 ; Crenshaw v. United States (1890) 134 U.S. 99, 105-106 [ 33 L.Ed. 825, 827-828 , 10 S.Ct. 431 ]; Butler v. Pennsylvania (1850) 51 U.S. (10 How.) 402, 416-418 [ 13 L.Ed. 472, 478-479 ].)

In Higginbotham, supra, a public officer (commissioner of public parks) was discharged in the middle of his term and sued to recover the balance of his salary for the stated term, asserting a federal contracts clause claim. The high court, in a brief, unanimous opinion (citing earlier cases), denied relief, holding that the contract clause does not protect the tenure assigned to a state’s public officers. The court relied on the “familiar principle that ‘the legislative power of a State, except so far as restrained by its own constitution, is at all times absolute with respect to all offices within its reach. It may at pleasure create or abolish them, or modify their duties. It may also shorten or lengthen the term of service.’ [Citations.]” ( 306 U.S. at p. 538 [83 L.Ed at p. 971].)

The Higginbotham (supra, 306 U.S. 535 ) principle, though stated in terms of “tenure,” extends to matters of compensation of public officers. As stated in Dodge v. Board of Education, supra, 302 U.S. at pages 78-79 [83 L.Ed at page 62], cited with approval in Higginbotham, state legislation fixing the compensation of public officers “creates no contract in their favor and the compensation named may be altered at the will of the legislature.” (See also Crenshaw v. United States, supra, 134 U.S. at p. 105 [33 L.Ed at p. 828] *533 [compensation paid to state public officers is not protected by federal contract clause].)

The applicable principle to be gleaned from Higginbotham, supra, 306 U.S. 535 , appears to be that a state (or its local subdivisions) has the power at any time to create, alter or abolish state or local public offices, and thereby reduce the affected officers’ salaries or other compensation, unfettered by federal contract clause restrictions. Respondent Eu and intervener argue that if, under Higginbotham, a state can terminate a public officer in the midst of his or her term without providing compensation for the remainder thereof, then it would follow that the state likewise can terminate future contributions toward his or her pension, assuming state statutory and constitutional guaranties are satisfied. But as petitioners observe, none of these high court cases involved the special area of vested pension rights, or concerned the rights of persons remaining in office for additional terms.

Although the issue is not entirely free of doubt, we conclude that the foregoing federal cases would not withhold federal contract clause protection from incumbent state legislators who have acquired vested pension rights under state law.

Higginbotham involved the tenure of public officials and the consequences arising from prematurely terminating that tenure. The case did not concern the quite distinct matter of the pension rights of public officials or employees. As we explained in Miller , although the Legislature may reduce the tenure of a public employee’s position, and thereby shorten his or her state service, public employee pension rights involve constitutionally protected obligations. “Pension rights, unlike tenure of civil service employment, are deferred compensation earned immediately upon the performance of services for a public employer ‘[and] cannot be destroyed . . . without impairing a contractual obligation. . . .’ [Citation.]” ( 18 Cal.3d at p. 814 .)

Decisions of this court have assumed the federal contract clause protects the vested pension rights of public officers. Olson and Betts , both supra, involved public officers (a state court judge and a state constitutional officer, respectively), and those cases concluded, without discussion of the Higginbotham/Dodge line of decisions, that the federal contract clause would apply to such persons. (See also Allen v. Board of Administration, supra, 34 Cal.3d at p. 119 .)

In Lyon v. Flournoy (1969) 271 Cal.App.2d 774 [ 76 Cal.Rptr. 869 ], appeal dismissed 396 U.S. 274 [ 24 L.Ed.2d 465 , 90 S.Ct. 564 ], a case likewise involving the rights of a state public official, the Court of Appeal *534 acknowledged the federal rule limiting federal contract clause protection for discharged state public officers (citing Dodge v. Board of Education, supra, 302 U.S. 74 ), and further observed that the public employee pension decisions of this court (e.g., Kern v. City of Long Beach, supra, 29 Cal.2d 848) “plac[ed] earned retirement rights within the shelter of the prohibition against contract impairment, without specific citation of either the federal or state clauses.” ( 271 Cal.App.2d at p. 781 .) The Lyon court nonetheless concluded, “Certainly the California pension decisions have never rejected the federal clause as a source of protection. It is now too late to do so. California law places earned pension rights of public officers and employees under the protection of the contract clause regardless of any characterization adopted by the federal courts.” (Ibid., italics added.)

We agree with Lyon v. Flournoy, supra, 271 Cal.App.2d 774 , that, in light of prior California decisions consistently extending federal contract clause protection to state public officers, it is simply “too late” to retreat from the clear implication of those holdings. We conclude that the pension restrictions of Proposition 140 are unconstitutional under the federal contract clause as applied to incumbent legislators because they infringe on the vested pension rights of those persons.

2. Nonincumbent Legislators

As for nonincumbent legislators first assuming office after Proposition 140 became effective, it seems clear they acquired no vested or protectible right to a continuation of the pension system in operation prior to their employment and, accordingly, the measure properly may be applied to them. (See Allen v. Board of Administration, supra, 34 Cal.3d at p. 124 .) Petitioners offer no argument or authorities suggesting a contrary rule as to such persons.

3. Severability

Although a portion of Proposition 140 is invalid as applied to incumbent legislators, its invalidity does not affect the remaining provisions of the measure, for those provisions can be given effect without regard to the validity or operation of the invalid pension restrictions.

Although the drafters of Proposition 140 included a severance clause (section 11 of the measure), by its terms it pertains only to the subject matter of article VII, section 11, of the state Constitution (regarding the obligations of the LRS) rather than to the entire initiative measure. It is possible the framers intended to apply the severance clause to the entire measure, as the clause referred to “this measure” and was placed at the conclusion of the *535 substantive provisions of Proposition 140. But in any event, it is clear that severance of particular provisions is permissible despite the absence of a formal severance clause. (See, e.g., People v. Mirmirani (1981) 30 Cal.3d 375, 387, fn. 9 [ 178 Cal.Rptr. 792 , 636 P.2d 1130 ].)

As we stated in Raven v. Deukmejian, supra, 52 Cal.3d at pages 355-356, quoting from an earlier case, “ *[t]he cases prescribe three criteria for severability: the invalid provision must be grammatically, functionally, and volitionally separable. [Citations.]’ ” In Raven , we invalidated a portion of Proposition 115 (an initiative pertaining to the rights of criminal defendants), but preserved the remainder, relying on the foregoing criteria for severability. The present measure likewise satisfies each of those criteria.

First, the invalid pension restriction, contained in new section 4.5 of article IV of the state Constitution, can be grammatically severed without affecting the operation of the remaining clauses. In order to preserve the section’s application to nonincumbent legislators, such severance would involve (1) striking from that section the words “or serving in,” so that the section applies only to “a person elected to ffi/ the Legislature on or after November 1, 1990,” and (2) construing the resulting language as applicable only to nonincumbent legislators first elected to the Legislature on or after that date. (See Walnut Creek Manor v. Fair Employment & Housing Com. (1991) 54 Cal.3d 245, 266-267 [ 284 Cal.Rptr. 718 , 814 P.2d 704 ].)

Second, the invalid restriction is functionally severable, as its removal would not affect the function or operation of the remaining provisions. The term and budgetary limitations, together with the restriction on nonincumbent legislators’ pension rights, can operate entirely independently of the invalid restriction.

Finally, the invalid portion of the measure is volitionally severable, in that the framers and voters undoubtedly would have adopted the remaining provisions had they foreseen the success of petitioners’ challenge. The framers included language preserving any “vested” rights which may have accrued, disclosing an intent to accommodate any interests protected by federal contract clause principles. As for the voters, there is no indication they placed undue emphasis on the pension rights provision as it applied to incumbent legislators.

III. Conclusion

We conclude that, except for the restriction on pensions of incumbent legislators, Proposition 140 is constitutionally valid. The petition for a *536 peremptory writ of mandate is granted to the extent it seeks to compel respondents to refrain from enforcing section 4 of Proposition 140 as applied to incumbent legislators. In all other respects, the petition for mandate is denied. The temporary stay issued on June 14, 1991, is vacated. The parties should bear their own costs. (See Raven v. Deukmejian, supra, 52 Cal.3d at p. 356 .)

Panelli, J., Kennard, J., Arabian, J., Baxter, J., and George, J., concurred.