Richard BARNETT, personally and as class representative, and Mary Bonilla, personally and as class representative, Plaintiffs-Appellants, v. CITY OF CHICAGO, et al., Defendants-Appellees, Carole Bialczak, et al., Intervening Defendants-Appellees.
After the 1990 decennial census, the City of Chicago was required by Illinois law to redistrict its 50 aldermanic wards in order to adjust to population shifts since the last census. The City Council was unable to agree on a new ward map, so alternative maps were submitted to the electorate in a referendum held in 1992. The winning map was promptly challenged in separate suits by representatives of black and Latino (Hispanic) voters. The district judge dismissed the blacks' suit on the pleadings shortly after it was filed. We reversed, Barnett v. Daley, 32 F.3d 1196 (7th Cir.1994), and the case was later tried along with the Latinos' case, which we had suggested be consolidated with it. The trial took 48 days, and it was almost a year before the district judge rendered his decision, upholding the City's ward map. 969 F.Supp. 1359 (N.D.Ill.1997). Both sets of plaintiffs have appealed. The only issue they raise is whether the City's map violates section 2 of the Voting Rights Act, which provides relief to minority groups, such as blacks and Latinos, that “have less opportunity than other members of the electorate ․ to elect representatives of their choice.” 42 U.S.C. § 1973(b). The first aldermanic elections based on the new map were held in 1995, and the next aldermanic elections will be held next year.
If a districting plan drew district lines in such a way that one group, even though it had a substantial fraction of the total electorate (say 49 percent), was a minority in every district, and if in addition the majority group adamantly refused to vote for any candidate that belonged to or was supported by the minority group-which in turn wanted to be represented only by one of its members-then the members of the minority would have no “opportunity ․ to elect representatives of their choice” even though they constituted almost half the electorate. From such a diabolical plan the inference of intentional discrimination would be overwhelming, and so the plan would violate the equal protection clause of the Fourteenth Amendment as well as section 2 of the Voting Rights Act. The Act, however, unlike the constitutional provision, does not require any showing of intentional discrimination, Milwaukee Branch of NAACP v. Thompson, 116 F.3d 1194, 1199 (7th Cir.1997); Barnett v. Daley, supra, 32 F.3d at 1201; NAACP v. City of Niagara Falls, 65 F.3d 1002, 1006 (2d Cir.1995), and by the same token is not limited to impairments of minority voting power so egregious as to compel an inference of such discrimination.
The plaintiffs have abandoned their effort to prove intentional discrimination, and with that out of the picture the question becomes, much as in a case of disparate impact under Title VII of the Civil Rights Act of 1964, whether the challenged districting plan impairs the voting power of minorities more than it has to. That depends on the alternatives. Sometimes there won't be any nondiscriminatory alternative. Imagine a minority group that accounted for less than 1 percent of Chicago's population and was scattered evenly throughout the City. Unless more populous segments of the community were allied with or otherwise supportive of this minority, it would be helpless to elect representatives of its choice to the City Council. Yet there would be no violation of the Voting Rights Act, because it would be infeasible to devise a plan that was more favorable to this minority group. Expanding the Council to 1,000 members, or to 100 and shifting to at-large elections for alderman, perhaps combined with a system of proportional representation, might do the trick. But it would be the kind of radical remedy that the Act does not require, since the protection of minorities is not the sole factor to be considered in a decision on districting.
The plaintiff is not required to propose an alternative map that is “final” in the “final offer” arbitration sense, where the parties cannot modify their offers once they have denominated them final and the tribunal is confined to choosing which of the final offers is better and cannot formulate its own, best remedy. But the plaintiff must show that there is a feasible alternative to the defendant's map, an alternative that does a better job of balancing the relevant factors, although the fine-tuning of the alternative can be left to the remedial stage of the litigation. Sanchez v. Colorado, 97 F.3d 1303, 1314-15 (10th Cir.1996); Nipper v. Smith, 39 F.3d 1494, 1533 (11th Cir.1994) (en banc).
Section 2 unfortunately provides no guidance on how to balance the factors and thus determine whether a challenged plan needlessly impairs a minority group's voting power. The statute tells the courts to consider “the totality of circumstances,” and that has turned out to be, if anything, worse than useless advice, as it has discouraged the Supreme Court from trying to particularize the standard. Johnson v. De Grandy, 512 U.S. 997, 1017-21, 114 S.Ct. 2647, 2659-62, 129 L.Ed.2d 775 (1994). All that is clear under the statute and caselaw at present is that if voting is “polarized,” in the sense that members of various racial or ethnic groups have a strong preference for a candidate that belongs to their group, and if it would be possible to reconfigure a districting map so as to give one of the minority groups one or more districts (in addition to what the map gives it, if the map gives it anything) in which the group would be a majority, then a more searching inquiry into whether the challenged map denies a minority group a fair opportunity to elect the candidates that its members prefer becomes necessary. Id. at 1011-13, 114 S.Ct. at 2657-58; Thornburg v. Gingles, 478 U.S. 30, 50-51, 106 S.Ct. 2752, 2766-67, 92 L.Ed.2d 25 (1986); Clark v. Calhoun County, 88 F.3d 1393, 1395-96 (5th Cir.1996); NAACP v. City of Niagara Falls, supra, 65 F.3d at 1007; Little Rock School District v. Pulaski County Special School District, # 1, 56 F.3d 904, 911-12 (8th Cir.1995). It is not even certain whether in making the threshold inquiry the court should use total population, voting-age population, or voting-age population of citizens only. The issue was expressly left open in Johnson v. De Grandy, supra, 512 U.S. at 1008-09, 114 S.Ct. at 2655-56. A larger fraction of both the black and the Latino populations of Chicago is under voting age than of the white population. The question of citizenship is also important in this case, because more than 40 percent of the Latinos in Chicago are not U.S. citizens.
Voting for alderman in Chicago is polarized. Blacks are almost never elected alderman in wards that do not contain a black voting-age majority and Latinos are almost never elected in wards that do not contain a Latino majority of citizens of voting age. Conversely, whites are rarely elected in wards in which either blacks or Latinos have an electorally effective majority. The implication is that unless the ward map is drawn in such a way that blacks and Latinos are concentrated (but not too concentrated, for any minority votes not needed to elect the minority candidate in the ward of concentration would be “wasted” from the standpoint of enabling the members of the minority group to elect candidates of their choice) rather than spread out, they will not be able to elect aldermen of their choice-their choice being aldermen of their own race or ethnicity. How concentrated these groups must be in order to constitute an effective majority in the sense just indicated depends on voting-related characteristics of the population, notably age, citizenship, registration, and turnout. Because blacks have a younger age distribution than whites, and a smaller percentage of eligible blacks are registered to vote than of eligible whites, it is a rule of thumb that blacks must be at least 65 percent of the total population of a district in order to be able to elect a black. Ketchum v. Byrne, 740 F.2d 1398, 1413-16 (7th Cir.1984); African American Voting Rights Legal Defense Fund, Inc. v. Villa, 54 F.3d 1345, 1348 n. 4 (8th Cir.1995); Prosser v. Elections Board, 793 F.Supp. 859, 869 (W.D.Wis.1992) (3-judge district court) (per curiam). Likewise, because of both age and the percentage of noncitizens, Latinos must be 65 to 70 percent of the total population in order to be confident of electing a Latino. Ketchum v. Byrne, supra, 740 F.2d at 1415. A black-majority ward, then, is one that is at least 65 percent black on a total-population basis, and a Latino-majority ward is one that is at least 65-70 percent Latino on the same basis. Insofar as this approach to determining majority status seems to reward groups whose members do not bother to register or vote though eligible to do so, it can be questioned as an original matter; but the approach is well entrenched in the cases, and the City does not question it.
An important consideration in deciding whether a districting plan discriminates is whether the plan creates majority districts for each group in proportion to each group's electoral potential. If (given polarized voting) a group that has 40 percent of the electorate has a working majority in only 10 of 100 districts, something is likely to be amiss, although it is possible that because of the geographical distribution of the group, differences in registration or turnout, or other factors, the situation cannot be rectified without undue sacrifice of other important values.
The City's map, the map challenged in this case, creates 24 majority-white wards, 19 majority-black wards, and 7 majority-Latino wards, calculated by the method used in the preceding paragraphs. The qualification is important. In one of the wards that we have called a white-majority ward, blacks have 54 percent of the voting-age population. It is nevertheless a white-majority ward because blacks constitute only 55 percent of the ward's total population, which is much less than 65 percent, and the remaining population is mainly white (44 percent of the voting-age population) and in practice (that is, with due regard for registration and turnout) controls the ward. The district judge, however, classified it together with four other wards in which whites have an absolute voting-age majority as “multi-racial,” thus reducing the number of majority-white wards to 19, compared to the 19 black and 7 Latino. He based this reclassification in part on the politics of these wards, finding that their white aldermen are attentive to the interests of the black voters. No doubt they are. But the blacks would prefer to elect black aldermen and cannot. It may be highly regrettable that a candidate's race should matter to the electorate; but it does; and the cases interpreting the Voting Rights Act do not allow the courts to ignore that preference. The judge may have been right or wrong as an original matter, but under the caselaw he clearly erred in taking these five wards out of the white column.
We want to make clear, however, that in so ruling we do not reject the concept of an “influence ward”-a ward in which a minority group has enough political heft to exert significant influence on the choice of candidate though not enough to determine that choice-or suppose the concept forever inapplicable to Chicago. It is merely inapplicable at this time on this record, because of the rigid racial bloc voting on all sides and the district court's flawed methodology of relying on its own estimation of the responsiveness of particular incumbents to particular groups.
Since there are 50 wards in all, the 24-19-7 racial and ethnic breakdown of the wards that is seen to be produced by the City's map once the erroneous analysis by the district court is set to one side means that 48 percent of the wards are white majority wards, 38 percent are black majority wards, and 14 percent are Latino majority wards. The population of Chicago, in contrast to the allocation of wards by the City's map, is 38 percent white, 39 percent black, and 20 percent Latino (the other 3 percent are mainly Asian, and even if Asians voted as a bloc their distribution throughout the City makes it impossible to create an Asian-majority ward). This makes it seem as if the Latino population is severely shortchanged, to the benefit of the white, who appear also to have gotten all the wards that would have gone to Chicago's Asian-American residents if the Asian-Americans had a more concentrated residential pattern.
The picture changes, however, if voting-age, or citizen voting-age, population is used as the benchmark for determining proportional equality of voting power. Whites have 44 percent of the voting-age population, blacks 36 percent, and Latinos 17 percent. And of just the citizen voting-age population, 46 percent are white (and remember that 48 percent of the wards on the City's map are white), 40 percent are black (38 percent of the wards), and only 11 percent are Latino (14 percent of the wards). We should decide, therefore, before going any further, which population basis should be used in determining whether the distribution of effective majority status is proportional to population. The concept of virtual representation argues for using total population. Children cannot vote, but their parents can, and if the parents can be assumed to be faithful agents of their children, then boosting parents' voting power by the number of their children-which, though in an extraordinarily rough way, is the effect of using total population as the basis for determining equality of voting power-gives the children a kind of representation in the political process. And similarly, if it is assumed that citizen and noncitizen Latinos have a strong community of interest, then giving citizen Latinos extra voting power based on the number of noncitizen Latinos gives the latter a kind of representation in the political process. The plaintiffs point out that the census, which is the basis on which the U.S. House of Representatives is apportioned, is based on total rather than voting-age or citizen voting-age population.
We think that citizen voting-age population is the basis for determining equality of voting power that best comports with the policy of the statute. We can set to one side both the census and the virtual representation of noncitizens by citizens. The census counts the total population for the same reason that it counts rather than samples: in order to minimize controversy. Cf. Wisconsin v. City of New York, 517 U.S. 1, 10-12, 116 S.Ct. 1091, 1096-97, 134 L.Ed.2d 167 (1996); Tucker v. U.S. Department of Commerce, 958 F.2d 1411, 1413 (7th Cir.1992). To verify the age and citizenship of the population would enormously complicate the decennial census and open the census-takers to charges of manipulation.
Neither the census nor any other policy or practice suggests that Congress wants noncitizens to participate in the electoral system as fully as the concept of virtual representation would allow, although permanent resident aliens are permitted to make federal campaign contributions, 2 U.S.C. § 441e, as are certain other nonvoters. The right to vote is one of the badges of citizenship. The dignity and very concept of citizenship are diluted if noncitizens are allowed to vote either directly or by the conferral of additional voting power on citizens believed to have a community of interest with the noncitizens-that being the very premise of the Latinos' claim in this litigation.
As for the virtual representation of children, which today as when we wrote our previous opinion seems to us a plausible ground for the use of total population (or at least total citizen population) as the basis for determining voting power, 32 F.3d at 1200, we do not consider the argument in favor of it strong enough to override the statutory language, which in its reference to “members of the electorate” excludes children. It is not as if the proposal were to give extra votes to families with more than the average number of children, a bizarre suggestion in our political culture but one that could be defended with reference to the concept of virtual representation. (Yet its bizarreness is a clue that the concept of virtual representation is unlikely to have been adopted by the Voting Rights Act.) The proposal is to give extra voting power to a racial or ethnic group which happens to have a higher than average number of children, as if the entire group bore some kind of attenuated parental relation to the children of members of the group. This is not a ridiculous idea. Some children will come of voting age during the period in which the City's ward map is in effect (by the same token, some Latino noncitizens will become citizens). And we are already departing from the literal meaning of “electorate” in using citizen voting-age population rather than registered voters-perhaps corrected for differences in turnout-as the base. But in addition to its being in tension with the statutory language, the idea of virtual representation for children is not one that the plaintiffs in this protracted litigation, now in its seventh year, have ever tried to develop or substantiate. In its present embryonic state the idea constitutes too speculative a ground for departing from a narrow reading of the statute.
Our conclusion that the proper benchmark for measuring proportionality is citizen voting-age population is consistent with the caselaw. Negron v. City of Miami Beach, 113 F.3d 1563, 1567-69 (11th Cir.1997); Campos v. City of Houston, 113 F.3d 544, 547-48 (5th Cir.1997); Romero v. City of Pomona, 883 F.2d 1418, 1425-26 (9th Cir.1989). It is true that some cases, including one of our own, use voting-age population rather than citizen voting-age population. But they are cases in which, as far as appears, noncitizens were not a significant part of the relevant population. McNeil v. Springfield Park District, 851 F.2d 937, 944-45 (7th Cir.1988); African American Voting Rights Legal Defense Fund, Inc. v. Villa, supra, 54 F.3d at 1352-53; see also Little Rock School District v. Pulaski County Special School District, # 1, supra, 56 F.3d at 911 n. 9. It is also true that only Villa discusses the issue of the proper population base in the specific context of determining proportional equality of voting power; but the considerations that all these cases discuss are applicable to that context as well. A group that has electoral power, in the sense of power to elect the representatives of its choice, equal to its fraction of the electorate will be hard-pressed to demonstrate, and has not demonstrated in this case, that its members nevertheless “have less opportunity than other members of the electorate ․ to elect representatives of their choice.”
When citizen voting-age population is used for determining proportional equality of voting power, the deviations from equality in the City's ward map are seen to be modest but not trivial. Blacks are underrepresented, whites and Latinos overrepresented. As each ward contains 2 percent of the City's population (though not necessarily of the citizen voting-age population) an alteration in the boundaries in the City's map that created one more black ward and one less white ward might well achieve an almost exact proportional equality of voting power between black and white. It would leave Latinos overrepresented. But only blacks and Latinos are plaintiffs in this case; no other group is complaining about its share of wards; and Latinos, it should be plain by now, have no basis for complaining about the City's ward map if we are correct that citizen voting-age population is the correct basis under the Voting Rights Act for determining proportional equality of voting power.
Have the black plaintiffs made a case for another ward? One of the maps they submitted would create 23 white wards, 20 black wards, and 7 Latino wards, which if Latino overrepresentation is ignored does track citizen voting-age population. Unfortunately from the standpoint of winding up this unduly prolonged litigation, this map was not the focus of the proceedings in the district court and is not discussed in the district court's opinion or in the briefs filed by the parties in this court. And because it was not, we cannot determine its appropriateness. Remember that proportionality is not the be-all or the end-all of redistricting under the Voting Rights Act. Deviations from proportionality, especially small ones, can be justified by reference to other factors, such as the compactness of districts and the desirability of preserving continuity and recognizing topographical, cultural, and economic factors that may make one ward mapping preserve communities of political interest better than another. The tricky part is maintaining equality of population and the desired racial and ethnic balance without creating wards of grotesque rather than merely irregular shape. But the maps submitted by the parties suggest that the creation of one additional ward for blacks, which is to say redrawing the map to give them 20 wards instead of 19, would be feasible in the sense of not trampling on any values that deserve weight in redistricting. We do not think that the plaintiffs' failure to highlight this alternative to the City's map in their briefs in this court constitutes a waiver. It was enough, we think, that they showed that the reasoning by which the district judge approved the City's map, which created only 19 black-majority wards, was deeply flawed.
We shall, therefore, while affirming the judgment for the City in the suit brought by the Latino plaintiffs, vacate the judgment in the other suit and remand the case to the district court for a prompt determination of whether the City's map violates the section 2 rights of the black plaintiffs. It may not-even though a map that shifts a ward from the white to the black column is feasible and would bring these groups' power to elect aldermen into line with their respective proportions of the citizen voting-age population-because, as we have emphasized, proportionality is not the only consideration in determining a violation of section 2. The district court may if it wishes take additional evidence. Although the Latinos' suit is dismissed, they should be permitted to intervene in the blacks' suit to protect their interests in any redrawn map.
In order that a new map (if the City's map is determined to violate the section 2 rights of the black plaintiffs) can be in place in time for the next aldermanic elections, in 1999, we instruct the district court to complete the proceedings on remand within 90 days. Should any of the parties appeal from the decision on remand, we shall set an expedited schedule for briefing and argument. The protraction of this litigation has been absurd, and the 90-day estimate for the proceedings on remand is generous. In Prosser v. Elections Board, supra, a suit to reapportion the entire Wisconsin legislature was filed on January 30, 1992; a two-day trial was held in April; and the court issued its decision, reapportioning the legislature, on June 2, 1992. Thus the entire proceeding was completed in four months. After eight years and two decisions by this court, the parties should be able to complete the litigation in the district court (including if necessary the drawing of a new map) within the period that we have allotted. In view of the time factor, we direct the clerk of our court to issue the mandate one week from today unless a petition for rehearing is filed within that period, in which event the mandate shall be issued upon the disposition of the petition. Boucher v. School Board, 134 F.3d 821, 829 (7th Cir.1998). There will be no award of costs in this court.
affirmed in Part, Vacated in Part, and Remanded with Directions.
POSNER, Chief Judge.