Calderon v. City of Los Angeles , 4 Cal. 3d 251 ( 1971 )

4 Cal. 3d 251 (1971) 481 P.2d 489 93 Cal. Rptr. 361

RICHARD M. CALDERON et al., Plaintiffs and Appellants,
v.
CITY OF LOS ANGELES, Defendant and Respondent.

Docket No. L.A. 29752.

Supreme Court of California. In Bank.

March 2, 1971.

COUNSEL

Lawrence William Steinberg, A.L. Wirin, Fred Okrand and Laurence R. Sperber for Plaintiffs and Appellants.

Mario G. Obledo and Joe C. Ortega as Amici Curiae on behalf of Plaintiffs and Appellants.

Roger Arnebergh, City Attorney, Bourke Jones, James A. Doherty, Assistant City Attorneys, and Thomas C. Bonaventura, Deputy City Attorney, for Defendant and Respondent.

OPINION

SULLIVAN, J.

We are called upon to decide whether the "one man, one vote" command of the equal protection clause requires that councilmanic voting districts be apportioned according to population or whether it is satisfied if each district contains a substantially equal number of registered voters. In short, is the constitutional imperative fulfilled by a standard of "one voter, one vote," rather than "one person, one vote"?^[1] As we explain infra, we have concluded that an apportionment plan based on registered voters will satisfy the equal protection clause only if it produces districts containing roughly equal numbers of people. Since, in the circumstances before us, the "registered voter" standard deviates sharply from population equality, it is constitutionally defective and must fall.

At issue is article II, section 6, subdivision 2(a) of the Charter of the City of Los Angeles (City), which provides, in pertinent part: "Between the dates of July 1 and September 15 of each fourth year, commencing with the year 1964, the Council shall ... redistrict the City into fifteen (15) districts and such districts shall be used for all elections of councilmen.... Districts so formed shall not deviate in the number of voters by more than ten percent above or below one-fifteenth of the total number of registered voters in the City of Los Angeles at the close of registration for the direct primary state elections held during the year in which such districts are to be established, and as nearly as practicable such districts shall be bounded by natural boundaries or street lines." (Italics added.)

Plaintiffs, Los Angeles residents and voters, brought this class action in separate counts for mandate and declaratory relief on behalf of themselves and all other Los Angeles citizens. They contended that the use of registered voters, rather than total population, as the standard for councilmanic apportionment is a direct violation of the equal representation principles established by such cases as Reynolds v. Sims (1964) 377 U.S. 533 [12 L. Ed. 2d 506, 84 S. Ct. 1362] and Avery v. Midland County (1968) 390 U.S. 474 [20 L. Ed. 2d 45, 88 S. Ct. 1114]. They argued that the "registered voter" basis results in substantial inequities, including over-representation of some districts and severe underrepresentation of others, particularly those populated largely by racial and ethnic minorities. Plaintiffs sought a declaration that the present apportionment scheme is unconstitutional and a writ of mandate commanding the City of Los Angeles to cease using voter registration as the basis for its councilmanic districting and to substitute population as the standard therefor. The trial court issued an alternative writ of mandate.

The City's answer to the petition and complaint (which also appears to be a return to the alternative writ) asserted in essence that while population is a valid apportionment basis, voter registration is also a constitutionally permissible standard. At the same time, the City moved for judgment on the pleadings on the ground that plaintiffs had stated no cause of action. Pursuant to stipulation of the parties, the court ordered off calendar the hearing on the return to the alternative writ and heard argument on the City's motion, which was granted. Judgment on the pleadings was entered in favor of defendant. This appeal followed.

I

The genesis of the present controversy lies in a seeming looseness in the language of the myriad United States Supreme Court decisions which have sought to interpret the "one man, one vote" principle since the issue of electoral apportionment was first held justiciable nine years ago. (Baker v. Carr (1962) 369 U.S. 186 [7 L. Ed. 2d 663, 82 S. Ct. 691].) Over the years the court has "used the words ``inhabitant,' ``citizen,' ``resident,' and ``voter' almost interchangeably in describing those who deserve representation, without indicating which of these bases for measuring substantial equality is most appropriate."^[2] (Note, Reapportionment (1966) 79 Harv.L.Rev. 1226, 1254.) In particular, the court has tended to treat as fungible two quite distinct concepts: first, that each district should contain equal numbers of people — a "population" standard; and second, that each voter is entitled to have his ballot carry an equal impact — a "voter" standard. (See Note, Reapportionment on the Sub-State Level of Government: Equal Representation or Equal Vote? (1970) 50 B.U.L.Rev. 231, 238-247.)

Thus, in Reynolds v. Sims, supra, 377 U.S. 533, the seminal decision in defining the "one man, one vote" doctrine, the high court stated, at one point, that "the fundamental principle of representative government in this country is one of equal representation for equal numbers of people, without regard to race, sex, economic status, or place of residence within a State." (Id. at pp. 560-561 [12 L.Ed.2d at p. 526].) (Italics added.) Yet a few pages later the court declared: "Simply stated, an individual's right to vote for state legislators is unconstitutionally impaired when its weight is in a substantial fashion diluted when compared with votes of citizens living in other parts of the State."^[3] (Id. at p. 568 [12 L.Ed.2d at p. 531].) (Italics added.) Later in the opinion the two concepts were fused and treated as indistinguishable: "Whatever the means of accomplishment, the overriding objective must be substantial equality of population among the various districts, so that the vote of any citizen is approximately equal in weight to that of any other citizen in the State." (Id. at p. 579 [12 L.Ed.2d at p. 537].) (Italics added.)

Although these principles present a surface similarity, their operative effect is quite different. "Equal representation for equal numbers of people" implies that a districting plan must count every resident, voter or not. By contrast, the doctrine that each voter's ballot should carry equal weight is satisfied so long as all districts contain the same number of registered voters, regardless of total population. Where the percentage of registered voters varies among districts, these two standards will produce dissimilar results.

This dichotomy was recognized and resolved in Burns v. Richardson (1966) 384 U.S. 73 [16 L. Ed. 2d 376, 86 S. Ct. 1286], where the court was confronted with a Hawaii apportionment scheme based on registered voters. Because of the large number of military personnel stationed in Hawaii and the sizable influx of tourists, many of the people physically present in the state were not permanent residents and hence ineligible for Hawaiian citizenship. The court began by asserting that the state was free to limit its apportionment base solely to state citizens, rather than using total population figures, thus excluding "aliens, transients, short-term or temporary residents, or persons denied the vote for conviction of crime...." (Id. at p. 92 [16 L.Ed.2d at pp. 390-391].)

However, the court reached a different conclusion on the validity of the registered voter standard. "Such a basis," it stated, "depends not only upon criteria such as govern state citizenship, but also upon the extent of political activity of those eligible to register and vote. Each is thus susceptible to improper influences by which those in political power might be able to perpetuate underrepresentation of groups constitutionally entitled to participate in the electoral process, or perpetuate a ``ghost of prior malapportionment.'" (Fn. omitted.) (Id. at pp. 92-93 [16 L.Ed.2d at p. 391].) In addition, the court pointed out that the number of registered voters might fluctuate considerably in a given election depending on "``such fortuitous factors as a peculiarly controversial election issue, a particularly popular candidate, or even weather conditions.' [Citation.]" (Fn. omitted.) (Id. at p. 93 [16 L.Ed.2d at p. 391].) "In view of these considerations," the court declared, "we hold that the present apportionment satisfies the Equal Protection Clause only because on this record it was found to have produced a distribution of legislators not substantially different from that which would have resulted from the use of a permissible population basis." (Id.) (Italics added.)

(1) We have no doubt that the holding of Burns binds local governmental entities like the Los Angeles City Council. Although decided after Burns, Avery v. Midland County, supra, 390 U.S. 474 and Hadley v. Junior College District, supra, 397 U.S. 50 have established beyond dispute that the "one man, one vote" principle and its subsequent interpretations are fully applicable to local units exercising general governmental powers.^[4] The Los Angeles City Council exercises far more comprehensive powers than did the Midland County Commissioners Court, which was largely concerned with building rural roads, or the Kansas City Junior College District, whose authority was confined to financing and operating an educational institution.

Even more compelling evidence of the applicability of Burns to the Los Angeles City Council is the fact that in Burns the Supreme Court cited approvingly Ellis v. Mayor and City Council of Baltimore (4th Cir.1965) 352 F.2d 123. That decision invalidated the Baltimore councilmanic districting scheme, which utilized a registered voter standard, producing a distribution that markedly diverged from one based on population. (See also Kapral v. Jepson (D.Conn. 1967) 271 F. Supp. 74, 80.) The high court approval of Ellis establishes conclusively that Burns applies to councilmanic apportionment.

The City relies on two federal district court decisions, Holt v. Richardson (D.Hawaii 1965) 238 F. Supp. 468 and Buckley v. Hoff (D.Vt. 1965) 243 F. Supp. 873, to support its contention that the equal protection clause permits the use of a registered voter apportionment base. (2) Such reliance is not merely misplaced; it also reveals a cavalier disregard for that basic jurisprudential principle that the decision of a court of superior jurisdiction supersedes contrary holdings of inferior tribunals.

Holt was the trial court decision in the case which, on appeal, became Burns v. Richardson. Consequently, any language in Holt which appears to validate a registered voter standard in the abstract — and a close reading of that case reveals little such textual support — is modified by the Supreme Court's holding in Burns. The other case, Buckley, was decided 10 months before Burns, and is therefore also of little precedential value. Furthermore, the trial courts in both Holt and Buckley approved a registered voter basis only after expressly ascertaining that in the particular jurisdiction, such a standard would produce only minor deviations from a population-based plan.^[5] Consequently, on their facts Holt and Buckley support the plaintiffs', not the City's, position.

(3) Although we are, of course, constrained by the supremacy clause (U.S. Const., art. VI, cl. 2) to follow decisions of the Supreme Court on matters of constitutional interpretation, we emphasize that we do so here not only from constitutional compulsion but also as a matter of conviction.^[6] Adherence to a population standard, rather than one based on registered voters, is more likely to guarantee that those who cannot or do not cast a ballot may still have some voice in government.

Thus a 17-year-old, who by state law is prohibited from voting, may still have strong views on the Vietnam War which he wishes to communicate to the elected representative from his area.^[7] Furthermore, much of a legislator's time is devoted to providing services and information to his constituents, both voters and nonvoters. A district which, although large in population, has a low percentage of registered voters would, under a voter-based apportionment, have fewer representatives to provide such assistance and to listen to concerned citizens. (See Note, Equal Representation and the Weighted Voting Alternative (1969) 79 Yale L.J. 311, 319.)

(4) In summary, as the Supreme Court has recently commented, "Equal representation for equal numbers of people is a principle designed to prevent debasement of voting power and diminution of access to elected representatives." (Kirkpatrick v. Preisler (1969) 394 U.S. 526, 531 [22 L. Ed. 2d 519, 525, 89 S. Ct. 1225].) (Italics added.) Crucial though voting is as a method of participation in representative government (Kramer v. Union School District (1969) 395 U.S. 621, 626-627 [23 L. Ed. 2d 583, 588-589, 89 S. Ct. 1886]; Carrington v. Rash (1965) 380 U.S. 89, 96 [13 L. Ed. 2d 675, 680, 85 S. Ct. 775]; Harper v. Virginia Bd. of Elections (1966) 383 U.S. 663, 667 [16 L. Ed. 2d 169, 172, 86 S. Ct. 1079]; Reynolds v. Sims, supra, 377 U.S. 533, 555, 561-562, 565 [12 L. Ed. 2d 506, 523, 526-527, 529]; Wesberry v. Sanders, supra, 376 U.S. 1, 17-18 [11 L. Ed. 2d 481, 491-492]; Castro v. State of California (1970) 2 Cal. 3d 223, 234 [85 Cal. Rptr. 20, 466 P.2d 244]), access to elected officials is also an important means of democratic expression — and one that is not limited to those who cast ballots.^[8]

Our preference for a population, rather than a registered voter, standard is rooted in another important consideration. The instant complaint alleges that the present voter-based system results in severe underrepresentation of those districts with heavy concentrations of blacks and Mexican-Americans. According to the figures presented, if the 15 councilmanic districts were precisely equal in population, each would contain 194,496 persons. However, District 9 in southeast Los Angeles, whose residents are largely black and Mexican-American, has 260,503 people, or about 33.9 percent above the ideal. Similarly, District 15, which comprises the Watts area and is mostly black, has a population of 228,814 persons, or 17.6 percent above the average. By comparison, District 2, located in the San Fernando Valley and containing few minority residents, has 164,850 people, or 15.2 percent below the ideal figure. Yet each of these districts is entitled to one representative on the city council.

Although such a discriminatory effect would have to be proved at trial,^[9] the United States Supreme Court has asserted that an otherwise acceptable apportionment plan may fail to pass constitutional muster if "designedly or otherwise," it operates "to minimize or cancel out the voting strength of racial or political elements of the voting population." (Fortson v. Dorsey (1965) 379 U.S. 433, 439 [13 L. Ed. 2d 401, 405, 85 S. Ct. 498]; see Burns v. Richardson, supra, 384 U.S. 73, 88 [16 L. Ed. 2d 376, 388]; Chavis v. Whitcomb (S.D.Ind. 1969) 305 F. Supp. 1364, prob. juris. noted sub nom. Whitcomb v. Chavis (1970) 397 U.S. 984 [25 L. Ed. 2d 392, 90 S. Ct. 1112, 1113, 1125].) Thus far this issue has been raised only in the context of dilution of minority voting strength by use of multi-member districts, but we see no reason why it is not relevant in testing the validity of a voter-based districting standard. (5) Racial or ethnic minorities often have distinct political interests, not shared by the general public, for which they seek political redress through their elected representatives.^[10] (See Chavis v. Whitcomb, supra, 305 F. Supp. 1364, 1386, prob. juris. noted sub nom. Whitcomb v. Chavis, 397 U.S. 984 [25 L. Ed. 2d 392, 90 S. Ct. 1112, 1113, 1125].) Where, for any reason, a non-population-based scheme tends sharply to reduce the representation of such groups, it must be regarded as constitutionally suspect.^[11]

We now apply the above principles to the instant case. Plaintiffs have presented statistics showing the number of registered voters and the population in each of the 15 councilmanic districts in Los Angeles.^[12] According to these figures, the breakdown, by population and voters, in each district is as follows (for convenience, we have listed the districts according to population, starting with the largest):

District               Population             Registered Voters
--------               ----------             -----------------
9                   260,503                     74,624
15                   228,814                     72,416
1                   227,776                     79,615
7                   205,423                     78,550
8                   205,385                     82,578
12                   204,854                     81,636
14                   189,876                     77,276
3                   189,723                     80,348
6                   186,783                     79,850
13                   183,202                     77,114
10                   179,649                     77,217
11                   175,022                     82,287
2                   164,850                     83,275
5                   162,123                     80,864
4                   153,462                     75,063

These figures demonstrate, with stark clarity, how sharply a voter-based apportionment may diverge from an acceptable population distribution. The number of voters in each district is roughly similar. (But see part II, infra.) Yet in terms of population, the districts are grossly unequal. As we have already noted, the two largest districts contain 33.9 and 17.6 percent more people than the ideal figure.^[13] The smallest district has 21 percent fewer residents than the ideal. When the largest and the smallest districts are compared with each other, we find that the largest has nearly 70 percent more people than the smallest. In light of repeated United States Supreme Court decisions that electoral districts must contain substantially equal numbers of people (see part II, infra), this variation is simply unacceptable. (6) Thus, on the pleadings before us, we can only conclude that the Los Angeles voter-based apportionment produces a distribution markedly different from that which would have resulted from use of a population basis. (Burns v. Richardson, supra, 384 U.S. 73, 93 [16 L. Ed. 2d 376, 391].) Consequently, it denies plaintiffs and other residents of Los Angeles the equal protection of the laws and is invalid.

The City contends, however, that plaintiffs fail to state a cause of action simply by alleging a variance between total population figures and voter registration. Its position is that since Burns held that a "citizen population," as well as a total population, standard could satisfy the requirements of "one man, one vote," plaintiffs should be required to show that the present voter-based system leads to an apportionment substantially different from the distribution produced by the use of any other "permissible population basis."

This argument entirely misconstrues the significance of Burns and is devoid of merit. The essence of Burns, and of nearly every other apportionment decision, has been the repeated assertion of the primacy of population as the keystone of electoral districting. (7) "Population is, of necessity, the starting point for consideration and the controlling criterion for judgment in legislative apportionment controversies." (Fn. omitted.) (Reynolds v. Sims, supra, 377 U.S. 533, 567 [12 L. Ed. 2d 506, 530-531].) As we shall discuss in part II, infra, the burden is thus placed on the jurisdiction whose apportionment is challenged to justify any significant deviation from population equality.

(8a) When plaintiffs set forth the district-by-district figures for total population and voters — with the wide variances between them — they stated a prima facie case for the invalidity of a voter standard. They were not obliged to negate the existence of any possible population basis which might correlate more closely with the voter-based districts. That task should fall to the City, to be presented by way of defense. In the absence of such defense, plaintiffs' allegations clearly stated a cause of action, and entry of judgment on the pleadings in favor of the City was erroneous.

We find direct support for our conclusion in Kapral v. Jepson, supra, 271 F. Supp. 74, where voters challenged the constitutional validity of a provision in the city charter of Milford, Connecticut, which required each voting district to contain "as nearly as possible an equal number of electors." Plaintiffs there submitted statistics showing large disparities in the number of residents in the five aldermanic districts, and moved for summary judgment. Apparently, no figures on the distribution of registered voters were even furnished. Thus, the court was faced with only the city charter's requirement of a registered voter apportionment and allegations of a jaggedly uneven distribution of people among the districts.

On this sparse record, the court granted plaintiffs' motion for summary judgment. While acknowledging the holdings of Burns and Ellis that a voter-based districting plan would satisfy the equal protection clause if it produced a distribution similar to that resulting from use of a permissible population basis, the court asserted, "But a specific factual finding to that effect would be required before apportionment based on numbers of registered voters could be said to comport with federal constitutional standards. The record in the instant case, at this stage of the proceedings, does not support such a finding." (Id. at p. 80.) It is clear, therefore, that a cause of action is stated merely by demonstrating that an existing districting scheme produces a markedly unequal distribution of people.^[14]

The City also makes much of the fact that the only population figures which plaintiffs have produced are estimates prepared by the Los Angeles Planning Commission.^[15] Strangely, the City contends that these statistics, prepared by its own planners, are an insufficient basis for demonstrating the inequality of the present districting scheme, without proof as to how they were arrived at and how accurate they are.

Although we agree that such information should be supplied at trial,^[16] we cannot accept the City's implied premise that population estimates can never be the basis of apportionment, nor its corollary that registration figures are a preferable standard because they are more frequently available and consequently more accurate. (9) Since the Supreme Court has stated that reapportionment more often than every 10 years is desirable (Burns v. Richardson, supra, 384 U.S. 73, 96 [16 L. Ed. 2d 376, 393]; Reynolds v. Sims, supra, 377 U.S. 533, 584 [12 L. Ed. 2d 506, 540]), the court necessarily implies that between decennial federal censuses, such redistricting may be based on reliable estimates.^[17]

(10) Furthermore, although it is true that registration figures are more frequently updated than the federal census, we reject the City's contention that this alone makes them a preferable, or even acceptable, apportionment standard. Since population is the controlling factor in any constitutionally valid districting scheme, the fact that registration statistics are current is legally irrelevant if they fail to reflect population distribution. This correlation cannot be determined unless appropriate population figures are available.

Nor is the City's position bolstered by its demonstration that two California code sections (Sts. & Hy. Code, § 2107 and Gov. Code, § 36516) provide that an area's population may be computed for certain purposes by multiplying the number of registered electors by three. In Burns the Supreme Court clearly rejected the proposition that voter registration was necessarily an accurate reflection of population. (See also Klahr v. Williams (D.Ariz. 1970) 313 F. Supp. 148, 151-152, juris. noted sub nom. Ely v. Klahr (1970) 400 U.S. 963 [27 L. Ed. 2d 382, 91 S. Ct. 364].) (11) When interpreting the federal Constitution, we are bound by the high court's holding, notwithstanding state statutory presumptions to the contrary. Furthermore, the allegations in the instant complaint strongly belie the legislative assumption which apparently underlies these code sections — that voter registration rates are uniform. Although on a statewide level considerable uniformity may appear,^[18] within a single city sizable fluctuations are likely to exist. (See Note, supra, 50 B.U.L.Rev. at pp. 243-244.)

(12) By our holding today we should not be understood to condemn a voter-based apportionment in all circumstances and for all time. Where such a plan is shown to fairly reflect population distribution, it may withstand constitutional attack.^[19] The Supreme Court in Burns isolated several factors which would contribute to such a result (Burns v. Richardson, supra, 384 U.S. 73, 96-97 [16 L. Ed. 2d 376, 393-394]), and we present these as guidelines for the future.

First, voter registration is likely to show a significant correlation to population distribution if a high percentage of those eligible to vote are in fact registered. In Burns, 80 to 90 percent of the potential voters had registered. Second, reapportionment should take place frequently — perhaps every four or eight years — in order to keep up with registration changes. We note that the Los Angeles Charter requires redistricting every four years, a commendable condition. Third, use of registration statistics from presidential elections will tend to assure a high level of participation and reduce the likelihood that fluctuations in local interest in the outcome of a particular race will cause erratic patterns of political activity which may then be frozen into an apportionment scheme for many years. Although the Los Angeles Charter requires redistricting in presidential years, it utilizes registration figures from the direct primary state election, rather than the general election, in those years. Such statistics do not maximize registration because not infrequently persons fail to register for a primary election for want of a contest in one or more offices. Since a race for every office is generally assured in the general election — particularly in a presidential year — those figures will produce the greatest registration. Finally, introduction of a system of permanent personal registration would help to ensure the stability and accuracy of the registration rolls, thus causing them to better reflect population distribution.

II

Although we hold that, on the pleadings before us, article II, section 6, subdivision 2(a) of the Los Angeles City Charter is invalid because of its reliance on a registered voter standard, we feel compelled to consider plaintiffs' other constitutional challenge to that provision. We do so now in order to anticipate a charter amendment merely substituting "people" in place of "registered voters," thus correcting one error while leaving another intact.^[20]

The remaining defect is the command that "Districts so formed shall not deviate in the number of voters by more than ten percent above or below one-fifteenth of the total number of registered voters in the City of Los Angeles...." (Italics added.) Plaintiffs contend that the 10 percent deviation thus approved is too large to meet the standards established by recent Supreme Court decisions. We agree that the formula is constitutionally faulty, although partly on other grounds.

Despite occasional statements in early apportionment decisions that "[m]athematical exactness or precision is hardly a workable constitutional requirement" (fn. omitted) (Reynolds v. Sims, supra, 377 U.S. 533, 577 [12 L. Ed. 2d 506, 536]; see Wesberry v. Sanders, supra, 376 U.S. 1, 18; Hadley v. Junior College District, supra, 397 U.S. 50, 58 [25 L. Ed. 2d 45, 52]), in recent years the Supreme Court has shown a steadfast determination to tolerate progressively smaller deviations from numerical equality in apportionment plans. (Comment, Effective Representation and Multimember Districts (1970) 68 Mich.L.Rev. 1577, 1591; see generally, Dixon, The Warren Court Crusade for the Holy Grail of "One Man-One Vote," in the S.Ct. Rev., 219-224 (P. Kurland ed. 1969).^[21]) For example, in Kirkpatrick v. Preisler, supra, 394 U.S. 526, the high court invalidated a congressional redistricting statute where the most populous district was 3.13 percent above the mathematical ideal and the least populous was 2.84 percent below. In a companion case, Wells v. Rockefeller (1969) 394 U.S. 542 [22 L. Ed. 2d 535, 89 S. Ct. 1234], a New York congressional apportionment plan which permitted variances of 6 percent above and below perfect equality was also rejected.

Those decisions involved congressional districting, which is based on article I, section 2, of the federal Constitution^[22] and for which the court has prescribed a rule of equality "as nearly as is practicable." (Wesberry v. Sanders, supra, 376 U.S. 1, 7-8 [11 L. Ed. 2d 481, 486].) By contrast, state and local apportionment is guided by the dictates of the equal protection clause; for such jurisdictions, the constitutional mandate, enunciated first in Reynolds, has been "substantial equality." (Reynolds v. Sims, supra, 377 U.S. 533, 579 [12 L. Ed. 2d 506, 537].) Despite the different constitutional bases and the varying formulae, however, "it has never been apparent that the Court sees these two clauses as producing different yardsticks for districting matters." (Dixon, op. cit. supra, at p. 222.)^[23]

We find it particularly significant that in Kirkpatrick the court repeatedly cited Swann v. Adams, supra, 385 U.S. 440, as authority for its strict rule of numerical equality. Swann, which invalidated a plan permitting deviations of up to 18 percent, involved redistricting of the Florida State Legislature. Similarly, on the same day that the court decided Swann, it reversed and remanded a congressional apportionment case "for further consideration in light of Swann v. Adams. ..." (Duddleston v. Grills (1967) 385 U.S. 455 [17 L. Ed. 2d 508, 87 S. Ct. 611]; see also, Kilgarlin v. Hill (1967) 386 U.S. 120 [17 L. Ed. 2d 771, 87 S. Ct. 820], voiding a state legislative apportionment plan with variations up to 14 percent.) Thus, the court clearly seems to imply that decisions concerning state legislatures are interchangeable with those involving congressional districts so far as the constitutionally required standard of mathematical uniformity is concerned. (See Dixon, op. cit. supra, at p. 222 and fn. 16; Comment, supra, 68 Mich. L.Rev. at p. 1591, fn. 51.)

Does this strict rule of equality apply to local government bodies like the Los Angeles City Council? No, contends the City, citing the following language from Avery v. Midland County, supra, 390 U.S. 474: "This Court is aware of the immense pressures facing units of local government, and of the greatly varying problems with which they must deal. The Constitution does not require that a uniform straitjacket bind citizens in devising mechanisms of local government suitable for local needs and efficient in solving local problems." (Id. at p. 485 [20 L.Ed.2d at p. 53].)

However, that passage was illustrated by citation of decisions involving either representatives chosen at large (Dusch v. Davis (1967) 387 U.S. 112 [18 L. Ed. 2d 656, 87 S. Ct. 1554]) or officials who were appointed, not elected. (Sailors v. Board of Education (1967) 387 U.S. 105 [18 L. Ed. 2d 650, 87 S. Ct. 1549].) Such cases hardly support a contention that popularly elected representatives, chosen from individual districts, may represent significantly unequal numbers of people. Furthermore, the Avery court announced a standard — "the Constitution permits no substantial variation from equal population in drawing districts for units of local government having general governmental powers" (390 U.S. at pp. 484-485 [20 L.Ed.2d at p. 53]) — which is virtually identical to the "substantial equality" rule applied to state legislatures. Interestingly, the other "local government" apportionment case, Hadley v. Junior College District, supra, 397 U.S. 50, appears to adopt the "as nearly as is practicable" standard hitherto reserved for congressional apportionment. (Id. at p. 56 [25 L.Ed.2d at p. 51].)^[24]

While the above analysis alone might not convince us of the per se invalidity of the charter provision here at issue, when it is coupled with another factor, we are satisfied that the section cannot stand. That factor is the Supreme Court's frequent rejection of any mathematical formula which purports to establish an "acceptable" variance from ideal equality. Correspondingly, the court has insisted that any such deviation be justified by the governmental unit involved on the basis of "legitimate considerations incident to the effectuation of a rational state policy...." (Reynolds v. Sims, supra, 377 U.S. 533, 579 [12 L. Ed. 2d 506, 537].)

These twin concepts were forcefully expressed in Roman v. Sincock (1964) 377 U.S. 695 [12 L. Ed. 2d 620, 84 S. Ct. 1462], a state legislative apportionment case where the court stated: "Our affirmance of the decision below is not meant to indicate our approval of the District Court's attempt to state in mathematical language the constitutionally permissible bounds of discretion in deviating from apportionment according to population. [Fn. omitted.] In our view the problem does not lend itself to any such uniform formula, and it is neither practicable nor desirable to establish rigid mathematical standards for evaluating the constitutional validity of a state legislative apportionment scheme under the Equal Protection Clause. Rather, the proper judicial approach is to ascertain whether, under the particular circumstances existing in the individual State whose legislative apportionment is at issue, there has been a faithful adherence to a plan of population-based representation, with such minor deviations only as may occur in recognizing certain factors that are free from any taint of arbitrariness or discrimination." (Id. at p. 710 [12 L.Ed.2d at pp. 629-630].)

Similarly, in Kirkpatrick, where even a 3 percent deviation was held excessive in the absence of a valid justification, the court stated: "We reject Missouri's argument that there is a fixed numerical or percentage population variance small enough to be considered de minimis and to satisfy without question the ``as nearly as practicable' standard. The whole thrust of the ``as nearly as practicable' approach is inconsistent with adoption of fixed numerical standards which excuse population variances without regard to the circumstances of each particular case." (Kirkpatrick v. Preisler, supra, 394 U.S. 526, 530 [22 L. Ed. 2d 519, 524].) Once again the court emphasized that the burden was on the state to justify each variance, no matter how small. (See also Kilgarlin v. Hill, supra, 386 U.S. 120, 122 [17 L. Ed. 2d 771, 774]; Swann v. Adams, supra, 385 U.S. 440, 443-444 [17 L. Ed. 2d 501, 504-505].)

The reasons for eschewing such formulae are not far to seek. First, it is practically impossible, without being arbitrary, to choose a cutoff point at which population deviations suddenly become de minimis. Second, use of such yardsticks encourages drafters of apportionment plans to employ the "acceptable" variations as a starting point, instead of striving for equality. (Kirkpatrick v. Preisler, supra, 394 U.S. 526, 531 [22 L. Ed. 2d 519, 524-525].)^[25]

These rationales are obviously applicable to the case at bench. The Los Angeles Charter provision flatly validates variations of 10 percent from the ideal per-district figure. We are not surprised, consequently, to observe that, even under the present registered voter standard, at least 5 of the 15 districts deviate nearly 5 to 8 percent from numerical equality. Furthermore, since the charter writes into law a permissible variance, it clearly dispenses with any legal requirement that the City justify each deviation, as the Supreme Court has repeatedly held to be constitutionally required. Nor, in the case before us, has the City made any effort to present an explanation for the existing discrepancies, preferring instead to rest on vague protestations that mathematical exactness is not constitutionally required and that the charter provision results in no "invidious discrimination."^[26]

We note that the provision permitting a 10 percent divergence is a relatively recent addition to the Los Angeles Charter. From 1925 until 1963, the charter required that each councilmanic district contain "as nearly as practicable equal numbers of voters" — the precise rule which in 1964 the Supreme Court would apply to congressional districting. Thus, at the very time that the Supreme Court was formulating constitutional standards of equality for electoral districts, Los Angeles was amending its charter to relax its own apportionment mandate.

We do not, of course, require prescience on the part of the City of Los Angeles. In decisions in 1965 and 1966 this court announced mathematical guidelines for redistricting which we said would carry a strong presumption of validity — standards of the very sort which the Supreme Court condemned in Roman, Kilgarlin, Swann and Kirkpatrick. (Wiltsie v. Board of Supervisors, supra, 65 Cal. 2d 314, 315-316; Miller v. Board of Supervisors, supra, 63 Cal. 2d 343, 350; Silver v. Brown (1965) 63 Cal. 2d 270, 279 [46 Cal. Rptr. 308, 450 P.2d 132].) One of our guidelines — that no district depart from numerical equality by more than 15 percent — required even less precision than the Los Angeles Charter provision currently before us.^[27] In Silver v. Reagan, supra, 67 Cal. 2d 452, however, we recognized that such a formula contravened recent Supreme Court emphasis that deviations from equality "cannot be presumed valid but must be justified by a specific showing that a permissible state policy is thereby promoted." (Id. at p. 458.) Consequently, we discarded our previous guidelines as the basis for permanent redistricting plans and have not subsequently enunciated any other such mathematical yardstick.

We think the City of Los Angeles must do the same. As we have previously stated, such a formula encourages city officials to accept the tolerable instead of striving for the perfect. Whether the City wishes to amend its charter to substitute its old "as nearly as practicable" standard for the present 10-percent rule or whether it opts simply to delete any numerical reference is a matter which we leave to its discretion. We do so because its choice between those two alternatives is constitutionally irrelevant. (8b, 13) The equal protection clause requires that city councilmanic apportionment produce districts of "substantially equal" population. (Avery v. Midland County, supra, 390 U.S. 474; Reynolds v. Sims, supra, 377 U.S. 533.) If the City complies with this mandate — as indeed it must — it will satisfy the constitutional imperative, regardless of the precise wording of its charter.

In summary, we hold that on the record before us, the judgment on the pleadings in favor of the City cannot be sustained for two reasons: first, the registered voter apportionment standard appears constitutionally faulty in view of the substantial variance between population and registered voters on a per-district basis; and second, the 10 percent provision authorizes too great a deviation from equality.

The judgment is reversed.

Wright, C.J., McComb, J., Peters, J., Tobriner, J., Mosk, J., and Burke, J., concurred.

NOTES

[1] Both these phrases have appeared in Supreme Court apportionment decisions, in addition to the more common "one man, one vote." "One voter, one vote" was coined by Justice Stewart, in his concurring opinion to Gray v. Sanders (1963) 372 U.S. 368, 382 [9 L. Ed. 2d 821, 831, 83 S. Ct. 801]. "One person, one vote" was the earliest formulation of the "one man, one vote" slogan. It occurs in a memorable passage in Justice Douglas' opinion for the court in Gray: "The conception of political equality from the Declaration of Independence, to Lincoln's Gettysburg Address, to the Fifteenth, Seventeenth, and Nineteenth Amendments can mean only one thing — one person, one vote." (Id. at p. 381 [9 L.Ed.2d at pp. 830-831].)

[2] For example, in Reynolds the court stated: "[I]t is a practical impossibility to arrange legislative districts so that each one has an identical number of residents. or citizens, or voters." (377 U.S. at p. 577 [12 L.Ed.2d at p. 536].) This semantic variation has been reflected in the court's actions. On the same day it decided Reynolds, approving apportionment apparently based on total population, it also handed down WMCA, Inc. v. Lomenzo (1964) 377 U.S. 633 [12 L. Ed. 2d 568, 84 S. Ct. 1418], in which the court voiced no objection to New York's use of citizen population, rather than total population, for purposes of legislative apportionment. (Id. at p. 641 [12 L.Ed.2d at p. 573].)

[3] Such "voter" language is by no means confined to the early reapportionment cases. For example, in the most recent such decision, Hadley v. Junior College District (1970) 397 U.S. 50 [25 L. Ed. 2d 45, 90 S. Ct. 791], the Supreme Court stated, "We ... hold that the Fourteenth Amendment requires that the trustees of this junior college district be apportioned in a manner that does not deprive any voter of his right to have his own vote given as much weight, as far as is practicable, as that of any other voter in the junior college district." (Id. at p. 52 [25 L.Ed.2d at p. 48]; see id. at pp. 56, 58 [25 L.Ed.2d at pp. 50, 52].) (Italics added.)

[4] We reached the same conclusion in Miller v. Board of Supervisors (1965) 63 Cal. 2d 343 [46 Cal. Rptr. 617, 405 P.2d 857], decided three years before Avery. There we applied the equal protection clause to invalidate districts used to elect the Santa Clara County Board of Supervisors. (See also Wiltsie v. Board of Supervisors (1966) 65 Cal. 2d 314 [54 Cal. Rptr. 320, 419 P.2d 440].)

[5] In Buckley, for example, the court found that use of a registered voter standard caused only a minimal change in the apportionment base. Taken together, Vermont's 10 largest cities had 35.8 percent of the state's residents and 36 percent of its voters.

[6] History, too, suggests that a population standard was intended by framers of the Constitution, at least in congressional elections. In Wesberry v. Sanders (1964) 376 U.S. 1 [11 L. Ed. 2d 481, 84 S. Ct. 526], the court held that congressional districts must contain "as nearly as is practicable" equal numbers of people. It based its conclusion on article I, section 2 of the federal Constitution, which provides, in part, that members of the House of Representatives shall be chosen "by the people of the several states, and ... shall be apportioned among the several States ... according to their respective numbers...." After a detailed examination of the debates of the Constitutional Convention, the court stated that when the Convention delegates agreed that the House should represent "people," "they intended that in allocating Congressmen the number assigned to each State should be determined solely by the number of the State's inhabitants." (Fn. omitted.) (376 U.S. at p. 13 [11 L.Ed.2d at pp. 489-490].) (Italics added.)

Furthermore, article I, section 2 spells out very carefully how a state's inhabitants were to be computed: "by adding to the whole number of free persons, including those bound to service for a term of years, and excluding Indians not taxed, three-fifths of all other persons." Thus, except for Indians and slaves, all residents of a state were to be counted in determining that state's quota of congressmen.

Yet during this period suffrage was by no means universal. Many states had restrictions based on wealth, and few granted the vote to women. Consequently, it seems clear that total population — not voters — was the apportionment criterion envisioned by the framers of the Constitution. As James Wilson, convention delegate from Pennsylvania, observed, "equal numbers of people ought to have an equal no. of representatives...." (1 The Records of the Federal Convention of 1787 (Farrand ed. 1911) at p. 180, quoted in Wesberry v. Sanders, supra, 376 U.S. 1, 11 [11 L. Ed. 2d 481, 488].)

When after the Civil War, the congressional apportionment formula was amended slightly to reflect the ex-slaves' new status as citizens, nearly identical language was used. This new formula, contained in section 2 of the Fourteenth Amendment, states: "Representatives shall be apportioned among the several States according to their respective numbers, counting the whole number of persons in each State, excluding Indians not taxed." Congressional debates over that amendment once again make clear that apportionment was viewed as reflecting all people, not just voters. Thus, Senator Johnson of Maryland responded to the comment of one of his colleagues as follows: "The honorable member seems to suppose that representation and the franchise are identical. They are as different as light from darkness. The Constitution says so; your own amendment proclaims it. You say that representation is to depend upon numbers. So did your fathers say so. They said it and you have followed their teaching, because they said it was a right to be represented, but not a right to vote." (Cong. Globe, 39th Cong., First Sess. 3029 (1866), quoted in Note, supra, 50 B.U.L.Rev. at p. 243.)

[7] We have recognized that an alien, who is also barred from voting, may still play a significant role in his community. In Purdy & Fitzpatrick v. State of California (1969) 71 Cal. 2d 566 [79 Cal. Rptr. 77, 456 P.2d 645], where we held that the equal protection clause forbade discrimination against aliens in public employment, we stated: "[A]ny classification which treats all aliens as undeserving and all United States citizens as deserving rests upon a very questionable basis. The citizen may be a newcomer to the state ... the alien may be a resident who has lived in California for a lengthy period, paid taxes, served in our armed forces, demonstrated his worth as a constructive human being, and contributed much to the growth and development of the state." (Fns. omitted.) (Id. at p. 582.)

[8] One form of such access is embodied in the First Amendment's guarantee of "the right of the people peaceably to assemble, and to petition the Government for a redress of grievances." There is nothing in that amendment to limit its protection to registered voters. (See, e.g., Tinker v. Des Moines School Dist. (1969) 393 U.S. 503, 506 [21 L. Ed. 2d 731, 737, 89 S. Ct. 733].)

[9] The City contests plaintiffs' allegation that a registered voter standard reduces the political power of minorities. It argues that other districts with large black and Mexican-American populations have a higher than average percentage of registered voters, and are thus overrepresented. This is an issue of fact to be determined at trial. Our statements herein are based on the assumption, which we must make for purposes of considering the City's motion for judgment on the pleadings, that the facts alleged in the complaint are true. (Colberg, Inc. v. State of California ex rel. Dept. Pub. Wks. (1967) 67 Cal. 2d 408, 412 [62 Cal. Rptr. 401, 432 P.2d 3].)

[10] We do not suggest that such group interests can justify variations from population equality. "[N]either history alone, nor economic or other sorts of group interests, are permissible factors in attempting to justify disparities from population-based representation. Citizens, not history or economic interests; cast votes." (Fn. omitted.) (Reynolds v. Sims, supra, 377 U.S. 533, 579-580 [12 L. Ed. 2d 506, 537-538].) Rather, our meaning is that within the framework of population-based apportionment, group interests may not be ignored.

[11] The City contends that plaintiffs must bear the burden of showing that any variance between registered voters and actual population distribution is the product of some "oppressive action" by city officials, apparently an impermissible gerrymander (Gomillion v. Lightfoot (1960) 364 U.S. 339 [5 L. Ed. 2d 110, 81 S. Ct. 125]; cf. Jackson v. Pasadena City School Dist. (1963) 59 Cal. 2d 876 [31 Cal. Rptr. 606, 382 P.2d 878]) or some other device purposely employed to reduce the voting power of minority citizens. We think this contention ignores the language of Fortson that districting schemes which reduce the electoral strength of particular population groups are suspect, whether they operate "designedly or otherwise." (379 U.S. 433, 439 [13 L. Ed. 2d 401, 405].) That phrase, we think, makes clear that no discriminatory motive on the part of districting officials need be shown to call into question such an apportionment.

[12] Upon our inquiry concerning the current accuracy of the statistics in the complaint, the parties stipulated to presentation of the updated estimates on which we base our decision.

[13] The mathematical ideal is computed by dividing the number of districts into the jurisdiction's total population. In Los Angeles that quotient would be 194,496 people per district.

[14] A similar result was reached in Ellis v. Mayor and City Council of Baltimore, supra, 352 F.2d 123, where, in voiding an apportionment based on voters, the court stated: "This is not to say that no departure whatsoever from strict population count is ever permissible; but when the formula adopted results in more than a minor deviation, it is at once suspect. [Citation.]" (Id. at p. 129.)

[15] In making these estimates, the Planning Commission apparently uses the most recent federal census as a starting point, and then computes subsequent changes based on such current criteria as residential electric meter installations, active residential electric meters, building permits, and school enrollment. (Respondent's brief in Court of Appeal, p. 15.)

[16] At this stage of the proceedings, the City has foreclosed its opportunity to attack the reliability of the plaintiffs' statistics by moving for judgment on the pleadings, a procedural device which admits, for purposes of the motion, the truth of the opponent's pleading but asks for judgment as a matter of law. (Colberg, Inc. v. State of California ex rel. Dept. Pub. Wks., supra, 67 Cal. 2d 408, 412.)

[17] Even the federal census, of course, is not an entirely accurate computation of the total population. (See Kirkpatrick v. Preisler, supra, 394 U.S. 526, 539, fn. 3 [22 L. Ed. 2d 519, 529] (Fortas, J., concurring).) Some five million persons were not counted in the 1960 census, and, as Justice Fortas points out, particular population groups, especially young black male adults, are more likely than others to be missed.

[18] We made such an assumption in Silver v. Reagan (1967) 67 Cal. 2d 452, 457, and footnote 3 [62 Cal. Rptr. 424, 432 P.2d 26], in the context of statewide congressional redistricting.

[19] Plaintiffs contend that apportionment based on registered voters can never be constitutionally valid as a permanent arrangement. They point out that the plan approved in Burns was an interim one, adopted by the Hawaii Legislature after another districting scheme had been invalidated. We do not think Burns is clear on this issue. At one point that opinion states: "We are not to be understood as deciding that the validity of the registered voters basis as a measure has been established for all time or circumstances, in Hawaii or elsewhere." (384 U.S. at p. 96 [16 L.Ed.2d at p. 393].) Later, the court declared: "We hold that, with a view to its interim use, Hawaii's registered voter basis does not on this record fall short of constitutional standards." (Id. at p. 97 [16 L.Ed.2d at p. 393].)

These statements show only that an interim plan based on registered voters may be sustained; they indicate nothing about the validity of a permanent plan, for the Burns court was not confronted with such a scheme. However, this issue may soon be raised in the Supreme Court, because upon remand, the federal district court approved a permanent districting scheme based on registered voters. (Burns v. Gill (D. Hawaii 1970) 316 F. Supp. 1285, 1293-1294.) At any rate, we need not decide this issue now.

[20] Such an amendment was recently attempted as part of an overall revision of the Los Angeles Charter. (§ 2.02, proposed charter.) However, the revised charter was defeated in the November 1970 election.

[21] Professor Dixon has suggested that the trend toward requiring equality was by no means presaged by the early apportionment decisions. He points out that as the court has tightened its arithmetic standards, the number of justices dissenting from the majority opinion has increased. (Dixon, op. cit. supra, at pp. 222-223.) However, we think that the court's apparently increasing severity may be a product of the policy that a court confines its holding to the facts before it. Thus, the early apportionment cases involved pre-Baker districting schemes which permitted vast disparities in per-district population. As state legislatures began redistricting in light of the emerging "one man, one vote" principles, the variances from equality slowly decreased. Consequently, the court's attitude appeared to harden, when in fact its decisions may have been merely reflecting the step-by-step approach to apportionment embraced by many jurisdictions. A number of early decisions which appear to validate large divergences from equality have been subsequently explained on other grounds. (See Kirkpatrick v. Preisler, supra, 394 U.S. 526, 532, fn. 2 [22 L. Ed. 2d 519, 525]; Swann v. Adams (1967) 385 U.S. 440, 444-445 [17 L. Ed. 2d 501, 504-505, 87 S. Ct. 569].)

[22] Article I, section 2 provides, in pertinent part: "The House of Representatives shall be composed of members chosen every second year by the people of the several states, and ... shall be apportioned among the several States ... according to their respective numbers...." (See fn. 6, supra.)

[23] The City contends that less precision is required in non-federal districting, citing the court's statement in Reynolds that "Somewhat more flexibility may therefore be constitutionally permissible with respect to state legislative apportionment than in congressional districting." (377 U.S. 533, 578 [12 L. Ed. 2d 506, 537].) But this statement was made in the context of the court's observation that since there are many more state legislative than congressional seats to be distributed in a state, such legislative districts may more feasibly follow political subdivision boundaries than would be the case in congressional districting. (See Dixon, op. cit. supra, at p. 222.)

[24] Hadley invalidated an apportionment plan under which in the particular circumstances of the case, a district containing 60 percent of the total basis was allotted only 50 percent of the elected representatives — the same 10 percent variance at issue in the Los Angeles Charter. However, the court's decision was based, in part, on the fact that the statutory formula by which representatives were distributed contained a built-in bias of up to 16 2/3 percent against the large districts.

[25] The court noted in Kirkpatrick that in drafting the Missouri reapportionment plan there at issue, one leading state legislator "deemed it proper to attempt to achieve a 2% level of variance rather than to seek population equality." (394 U.S. 526, 531 [22 L. Ed. 2d 519, 525].)

[26] The City argues that the equal protection clause goes no further than prohibiting "invidious discrimination." In support of this contention, it cites Williamson v. Lee Optical Co. (1955) 348 U.S. 483 [99 L. Ed. 563, 75 S. Ct. 461], a 16-year-old decision upholding a state's power to bar persons who are not licensed ophthalmologists or optometrists from selling or fitting eyeglasses without a valid prescription. That Williamson has no precedential value whatsoever in the field of political representation and voting — fundamental rights entitled to special judicial protection — should be self-evident.

Although the phrase "invidious discrimination" has been used in some apportionment decisions (see, e.g., Reynolds v. Sims, supra, 377 U.S. 533, 561 [12 L. Ed. 2d 506, 527]; Baker v. Carr, supra, 369 U.S. 186, 244-245 [7 L. Ed. 2d 663, 701-702] (Douglas, J., concurring)), Avery v. Midland County emphasized what was already obvious — that in this context, "invidious" means simply "that bases other than population [are] not acceptable grounds for distinguishing among citizens when determining the size of districts...." (390 U.S. 474, 484 [20 L. Ed. 2d 45, 53].)

[27] When we formulated that standard, we were following the figure suggested for congressional redistricting in legislation then pending before Congress. (H.R. 5505, 89th Cong., 1st Sess. (1965).) That bill did not pass, nor did subsequent legislation which would have limited the permissible deviation to 5 percent after the 1970 census. (H.R. 2508, 90th Cong., 1st Sess. (1967).) The Kirkpatrick decision makes even the latter standard questionable.

Besides the 15 percent guideline, we also required that a majority of members of each house of the Legislature be elected by voters of districts with at least 48 percent of the state's total population. (Silver v. Brown, supra, 63 Cal. 2d 270, 279.)

Even when we announced these guidelines, we recognized that "the United States Supreme Court has eschewed establishing rigid mathematical standards for evaluating legislative apportionments [citations]...." (Id.) However, we thought it only fair to governmental units whose apportionment was challenged to establish some general guidance.