Michael ROBERTS, Plaintiff,
v.
Jerry WAMSER, et al., Defendants.
No. 87-0347C(3).
United States District Court, E.D. Missouri.
December 23, 1987.*1514 Donald Wolff, Wolff & Mass, Clayton, Mo., Richard E. Schwartz, St. Louis, Mo., Grover Hankins, General Counsel, NAACP, Baltimore, Md., Michael A. Middleton, Columbia, Mo., for plaintiff.
Leo V. Garvin, Jr., Richard Alan Cooper, John A. Kilo, Klutho & Flynn, St. Louis, Mo., for defendants.
MEMORANDUM
HUNGATE, District Judge.
This matter is before the Court to determine the merits of plaintiff's claim, as set forth in the fourth amended complaint, after a three-day trial before the Court sitting without a jury.
Pursuant to the Court's federal question jurisdiction, plaintiff brings this action seeking, through a fourth amended complaint, equitable relief and attorney's fees for alleged violations of Section 2 of the Voting Rights Act of 1965, as amended, 42 U.S.C. § 1973.[1] Plaintiff contends certain practices of the Board of Election Commissioners for the City of St. Louis resulted in plaintiff's loss of the March 3, 1987, Democratic primary election for President of the Board of Aldermen in the City of St. Louis, Missouri, and a denial of an equal opportunity for black voters in the City to participate in the electoral process. Defendants deny liability.
Having carefully considered the pleadings, testimony, exhibits, stipulations, memoranda, and relevant record, the Court makes and enters the following findings of fact and conclusions of law.
Findings of Fact
1. Plaintiff, Michael V. Roberts, is a black citizen of the United States of America, and a resident and registered voter of the City of St. Louis, Missouri. Plaintiff was a major black contender in the Democratic primary election held on March 3, 1987, for President of the Board of Aldermen in the City of St. Louis, Missouri ("City").
2. Defendants Jerry B. Wamser, Rita M. Krapf, David A. Robbins, and Walter R. Wrenn, Jr., are the duly appointed and acting members of the Board of Election Commissioners of the City of St. Louis ("Board"). Wamser and Krapf have served as Chairman and Secretary of the Board, respectively, since 1981. Robbins was appointed to the Board in 1985, Wrenn in 1986. Defendants Wamser, Krapf, and Robbins are white; defendant Wrenn is black.
3. Intervenor-defendant Thomas A. Villa ("Villa") is a white citizen residing within the City of St. Louis, Missouri. Villa was a major white contender in and was declared the winner of the Democratic primary election held in the City on March 3, 1987, for President of the Board of Aldermen.
4. General municipal elections are held bi-annually in the City on the first Tuesday in April; a primary election is held on the first Tuesday after the first Monday in *1515 March. See Mo.Rev.Stat. § 115.121, et seq. (1986).
5. In 1983, a primary election was held in the City on March 8 for candidates for the Office of President of the Board of Aldermen; the general election was held on April 5.
In the March 1983 primary election for the Democratic party's nomination for the Office of President of the Board of Aldermen, Roberts was defeated by Thomas Zych (a white candidate) by an official total of 751 votes. Roberts then timely filed an election contest in the Circuit Court for the City of St. Louis, pursuant to Mo.Rev.Stat. § 115.531 (1978, now 1986), in which he sought a new election or a recount. Roberts alleged in his petition, inter alia, that large numbers of voters' ballots had not been counted because either no vote had been recorded on the ballot (an "undervote"), or more than the permissible number of candidates' holes had been punched (an "overvote"). That petition contained no allegations that any alleged errors were racially discriminatory or violated any voter's right to vote. Roberts was not awarded a new election, but a recount was ordered. The recount confirmed the primary election results and actually widened Roberts' margin of defeat. Roberts took no appeal, even though an appeal was available to him pursuant to Mo.Rev.Stat. § 115.551 (1978, now 1986). The Board allowed Roberts' supporters to use stickers for purposes of voting for Roberts as a write-in candidate in the April 1983 general election for President of the Board of Aldermen. Roberts lost and Zych won that election.
6. Roberts was a candidate for the Democratic party's nomination for the Office of President of the Board of Aldermen in the March 1987 primary election in the City. He was opposed by Villa, Geraldine Osborn, and Clifford Wilson. Osborn is a white Alderwoman for the City's 15th Ward; Wilson is black.
7. In his 1987 primary campaign, Roberts made overt racial appeals to black voters. Roberts accused a white opponent Osbornof being backed by "the Klan."
8. Roberts instituted this lawsuit pro se on Friday, February 27, 1987, just four days prior to the primary election. In the first version of the complaint, Roberts and another sought declaratory relief, injunctive relief, and punitive damages from Board defendants and another defendant for alleged violations of federal statutory and constitutional provisions. No claim under state law was presented at that time. Plaintiffs did not seek any temporary relief and did not serve all Board defendants until mid-morning on primary election day. Temporary relief was denied at that time, and a hearing was set on the request for preliminary injunctive relief.
9. The official returns for the March 3, 1987, Democratic primary election for President of the Board of Aldermen in the City reflect that Roberts lost the primary election to Villa by 171 votes out of 77,444 votes cast. On March 4, 1987, the Board announced the official results of the March 3, 1987, primary election. After the general municipal election on April 7, 1987, Villa was declared the winner of the race for President of the Board, and he now holds that position.
10. Missouri law provides that a candidate who loses an election by less than one percent of the total number of votes cast in the election is entitled to an automatic recount. Mo.Rev.Stat. § 115.601 (1986). Since Roberts lost the March 3, 1987, primary by less than one percent of the votes cast, he was entitled to an automatic recount under that statute. Missouri law also provides that a losing candidate who desires to contest the results of a primary election must file a verified petition in the Circuit Court for the county in which the election (and alleged irregularity) occurred within five days after the official announcement of the results of the primary election. Mo.Rev.Stat. § 115.531 (1986).
11. Roberts never filed in the Circuit Court for the City of St. Louis a verified petition contesting the March 3, 1987, primary election. Instead, he filed in this Court a series of amended complaints in *1516 which he alleged numerous historical and longstanding acts of racial discrimination by the Board in connection with, inter alia, the Board's voter registration practices and procedures; the purging of voters from the rolls; the processing of absentee ballots; and the tabulation of votes by the Board's computerized tabulating equipment, all alleged to be in violation of Roberts' rights under the Civil Rights Act, the Voting Rights Act, and Amendments XII, XIV, and XV to the United States Constitution. A second count of his amended complaint, initially submitted to this Court at 7:55 p.m. on Friday, March 6, 1987, and filed March 9, 1987, sought a recount pursuant to Mo. Rev.Stat. § 115.601 (1986), and this Court's pendent jurisdiction. On Monday, March 9, 1987, the Court declined to exercise pendent jurisdiction over Count II of Roberts' amended complaint. This ruling was followed upon subsequent amendments of the complaint.
12. On April 9, 1987, the Court ordered a manual recount of the votes cast in the City's March 3, 1987, primary election for President of the Board of Aldermen from Wards 16, 1, 12, 20, 13, 15, and 22. The recount was to apprise the Court and the parties of the accuracy of the official election returns, and showed a narrowing of the margin between Villa and Roberts by 51 votes (Villa lost 10 votes and Roberts gained 41 votes). While percentage-wise this was a small difference, it was significant when applied to a winning margin of only 171 votes since it was a recount in only one quarter of the City's wards. Therefore, on August 4, 1987, the Court ordered a recount of ballots cast in the remaining wards during that March 3, 1987, election. The Court noted this procedure was not the statutorily authorized state recount procedure, but a pretrial tool to allow the parties to examine the ballots in light of the claims raised in this litigation.
13. Ballots disputed during this recount were examined by a three-member panel. One member was selected by Roberts and one member by Villa. The third member was appointed by the Court upon the failure of the Roberts and Villa representatives to agree on a third member. As the panel reports reveal, the caliber of each of the three panel members was outstanding.
14. The parties disagree as to the precise results of the recount, but it is clear that Roberts would have lost the election by no less than sixty votes even under the most favorable rules applied, either unanimously or by majority vote of the panel.
15. On June 26, 1987, Roberts, through new counsel, filed a fourth amended complaint which deleted most of the earlier allegations. The fourth amended complaint does not present any pendent state claims. By this complaint, plaintiff alleges that the Board had
(a) adopted and implemented ballot casting, handling and counting procedures, practices, and methods which have resulted in the failure to count disproportionately large numbers of ballots cast by black voters, as compared to the number of ballots cast by white voters; [and]
(b) since the March 3, 1987 primary election, up to and including this date, the Board ... continuously refused to count the ballots cast which were not counted electronically.
Generally, then, Roberts complains that uncounted ballots ("over-and-undervoting") occur in disproportionately greater numbers among blacks and that the ballot casting and counting procedures adopted and implemented by the Board (i.e., punch-card voting and computerized tabulation) are to blame.
16. The State of Missouri has enacted a Comprehensive Election Act of 1977, Mo. Rev.Stat. § 115.001, et seq. (1986), which governs all aspects of the election process in the state, and authorizes the use under certain circumstances of automatic tabulating equipment. See Mo.Rev.Stat. § 115.225, et seq. (1986).
17. Prior to 1981, the Board employed the use of Shoup "lever-type" voter machines. Since at least 1981, all ballots used in elections held within the City of St. Louis have been cast on punch cards and tabulated on computerized equipment. While the policy for the new equipment was not clearly *1517 set forth, the Board Chairman noted some advantages of the new system. The punch card voting system equipment is much smaller than the lever machines. Thus, storage problems were alleviated by the change. Also, mistaken ballots may be corrected by obtaining a new ballot.
The record reflects, without dispute, that the lever-type machines did not allow a voter to "overvote."
The Board's present vote tabulating system has been certified for use by the Missouri Secretary of State,[2] who has general supervisory authority over all elections within the State of Missouri.
18. To cast a vote in a punch-card voting system, a voter pushes a sharp pointed metal-tipped device (stylus) through a thin cardboard card (ballot card) in a designated spot. A particular spot corresponds to a particular candidate or issue. The voter determines which candidate or issue to vote for by reading the list of candidates' names or issues that are printed in a small booklet (ballot label) which is affixed to a device in the voting booth. The device is designed to keep the punched out portion of the card, the "chad," trapped inside the voting device and away from the ballot card and the ballot label.
Before voting, the voter slips the ballot card into the voting device so that two large holes at the top of the ballot card will slip over two fixed posts that hold the card in place. If the card is put in backwards or upside down, it will not easily fit into the voting device properly; and mechanical interlocks are provided to prevent the voter from punching holes in such an improperly inserted ballot unless the voter physically continues to hold the ballot in the device. Usually, bold instructions appear on both sides of the card telling the voter what to do. For example, the back side of the card may state, "STOPYOU ARE INSERTING THE CARD WRONG SIDE UP, TURN OVER AND REINSERT." On the front of the ballot card, it may state, "INSERT CARD, THIS SIDE UP. PLACE HOLES OVER POSTS." Further, there is an arrow on the card pointing in the direction of proper insertion. The ballot label also contains instructions such as "VOTE FOR ONLY ONE" or "VOTE FOR TWO" or whatever the particular election may require. The ballot label may be several pages long, and as the voter turns the pages, other areas of the ballot card are exposed where a voter may punch the card. After a voter has completed voting, the voter removes the ballot card from the voting device, folds the ballot card, and takes the ballot card to a locked ballot box where the ballot card is deposited.
The cast ballots are removed from the polling place by a bipartisan team accompanied by a police officer, and taken to the Board of Election Commissioners' office for counting. The counting is done by a computer that "reads" the ballot cards by passing a beam of light over the card. The beam of light shines through the hole in the ballot card and the vote is registered. If there is not a hole in the card ("undervote") or no hole "seen" by the computer, a vote cannot be counted by the tabulating equipment. If there are two holes punched in an election that allows only one selection ("overvote"), the computer will disregard both votes. A vote will also be disregarded in a primary election where the voter cast a vote for an opposite party candidate (i.e., where a Democrat is given a Republican primary ballot or vice versa). At the conclusion of the vote tabulation, the ballot cards are saved pursuant to state law.
19. Since 1981, the Board has had a written policy set forth in its manual of "Instructions for Election Judges" of oral instruction for voters at each polling place. In particular, the manual states:
It is mandatory that you offer to explain and demonstrate each step in voting before a voter is allowed to vote. The *1518 following points should be stressed in the demonstration.
* * * * * *
4. Inform each voter that if he spoils a ballot, he may return the spoiled ballot to the Election Judge and be issued a new ballot card. Only one (1) additional ballot will be provided to any voter.
5. Allow voter to make a punch on Demonstrator to get the feel of the Votomatic.
* * * * * *
7. Advise voter to fold stub over only at perforations. Do not bend, fold, or mutilate ballot card OR TEAR OFF THE STUB.
(emphasis in original).
20. To prepare for an election, the ballot is created using a "ballot management system," the election is "initialized," and the process is tested by running through the equipment for each precinct a ballot card test deck (including one ballot card representing an undervote and one card representing an overvote) with predetermined totals. Additionally, to test the accuracy of the tabulating system, a "logic and accuracy" test is run, in the presence of a bipartisan team and, if desired, a representative for each matter on a ballot, using a ballot card test deck from one precinct for each ballot style throughout the City. After the test, the observers sign a certification that the "logic and accuracy" was performed correctly. The "logic and accuracy" test is run again, in the same manner, the day of the election before any cast ballots are counted.
Upon counting ballots cast by voters, a preliminary audit is conducted on election day. The computer prints a "ballot cast statement" of the number of ballots it considers to have been cast. This statement is compared with the election judges' certification forms, and any discrepancies in the totals disclosed by each are minimized. (If 200 voters in a precinct take ballots but the computer has only counted 190 ballots in that precinct, election board officials seek to reduce this variance. The testimony was, however, that a discrepancy of plus or minus one to three votes per precinct is considered acceptable). The day after the election, a "final audit trail" is conducted, in an effort to eliminate any discrepancies in any precinct. The purpose of the audit is to have the number of ballots counted equal the number of ballots cast in the election. A canvas, which is a total statement of votes cast in every precinct for each candidate or issue in the election, is printed along with the official results. The Board then certifies the canvas. Finally, a random precinct is run again through the equipment to show that the result obtained at certification is the same as the result obtained on election day.
21. With respect to the March 3, 1987, municipal primary election, twenty-three out of the City's 395 precincts had irreconcilable discrepancies.[3]
22. In the March 3, 1987, municipal primary election, the Board duly prepared, provided, and used the ballot cards, voting devices, and tabulating equipment.
23. No evidence of record shows or implies that with respect to the March 3, 1987, municipal primary election, the manner of handling and counting cast ballots diverged from usual procedure or was in some way deficient. Nor is there any evidence of record to suggest that the usual manner of handling cast ballots or of checking the accuracy of the tabulating equipment is improper or needs alteration.
24. The testimony of the Board chairman and of several voters establishes that in actual practice the Board on March 3, 1987, and in other elections did not offer to explain and demonstrate each step in voting before a voter was allowed to vote. This failure affected black and white voters alike.
25. In each municipal primary election conducted since 1981, the Board did not count for any particular candidate approximately *1519 3,000 cast ballots.[4] In the Democratic primary election held in the City on March 3, 1987, the Board did not count for any particular candidate a total of 2,864 cast ballots,[5] with 1,677 of those uncounted ballots being from "black" wards, and 657 of those uncounted ballots being from "white" wards.
26. The Board, despite its accessibility to the above statistics, had initiated no specific voter education activities to minimize or alleviate the high incidence of over/undervoting (uncounted ballots). The Board Chairman did not view the alleged higher incidence of over/undervoting in the black community to be a "problem" or to require attention other than that given to all voters city-wide through education programs, demonstrations, and election judge training.
27. The parties' experts diverged somewhat in the identification of which wards are "black," which are "white," and which are "racially mixed," for purposes of this case. One of plaintiff's experts, Kay Gabbert, a political consultant from St. Louis, Missouri, defined "black" wards as those having a black alderman. Thus, such wards include: 1, 3, 4, 5, 18, 19, 20, 21, 22, 26, and 27 (wards on the north side of the City). She identified "white" wards as wards 9, 10, 11, 12, 13, 14, 15, 16, 23, 24, and 25 (wards on the south side of the City); and "racially mixed" wards as wards 2, 6, 7, 8, 17, and 28 (wards located predominantly in a "corridor" running through the middle of the City between the north side and south side).
Plaintiff's other expert, Dr. Gordon Henderson, a consultant in computers and politics from Richmond, Indiana, found eight wards to be "black." This conclusion was reached after calculating each ward's voting age population, determined from 1980 census information. Based both on professional standards and on an effort to have the percentage race in "black" or "white" wards as similar as possible, Dr. Henderson considered "black" or "white" those wards having 90% or more of such race predominating in the ward's voting age population. Thus, he considered wards 1, 3, 4, 18, 20, 21, 22, and 26 as "black" wards; wards 9, 10, 11, 12, 13, 14, 15, 16, 23, 24, and 25 as "white" wards; and wards 2, 5, 6, 7, 8, 17, 19, 27, and 28 as "racially mixed" wards.
Board defendants' expert, Kenneth Warren, a political science professor at St. Louis University and a political consultant from St. Louis, Missouri, identified "black," *1520 "white," and "racially mixed" wards based on his standard that if 90% of a ward's population is predominantly one race, then that ward may be characterized as a ward of that race. Dr. Warren's identification of "black," "white," and "racially mixed" wards coincides with Ms. Gabbert's.
For purposes of this case, the Court accepts the eleven "black" wards identified by Dr. Warren and Ms. Gabbert, and notes there is no dispute about the identification of "white" wards. Cf. n. 11, infra.
28. The parties' experts testified that professional literature and studies provide several possible reasons for an "undervote" or "overvote," votes not counted by the automatic tabulating equipment. The reasons include: apathy; interest; protest voting; trying to correct a mistaken vote; voter fatigue due to placement on a large ballot; and education of the voter. The Board chairman testified that such votes may also reflect the existence of an expectation by various groups that certain persons must go to the polls. In other words, political parties, unions, or other entities may require those affiliated with them to go to the polls, and those going to the polls may cast "non-votes" by under- or overvoting.
Until we know which is which (over/undervoting), we cannot be certain the actions did not express the voters' desires (e.g., undervoting: not voting for a particular candidate deliberately to show dissatisfaction). Overvoting is a separate problem and the size of that problem is unknown until over and undervoted ballots are distinguished. If there are 2,800 ballots cast but not counted, and 2,700 of those ballots are "undervotes," the problem may be different from a situation in which 2,700 of the uncounted ballots are "overvotes."
29. A certain number of ballots that reflected a clear choice of candidates in the March 3, 1987, Democratic primary election could not be counted by the Board relying on the computerized tabulation equipment alone. Although the appointed panel could ascertain and unanimously agree on for whom a disputed ballot was cast, the same vote could not be and was not counted by the computerized equipment "Deus ex machina."
30. Dr. Henderson, one of plaintiff's experts, used homogeneous analysis and bivariate regression analysis[6] to examine ballots cast and not counted in the City's March 3, 1987, primary election. As a result of these analyses, Dr. Henderson determined that, to a statistically significant degree, votes cast in "black" wards were much more likely not to be counted than those cast in "white wards."[7] Dr. *1521 Henderson did not submit other City elections to such analyses.[8]
31. Neither of defendants' experts subjected the March 3, 1987, election results to either homogeneous ward or bivariate regression analyses, or explicitly disputed the results Dr. Henderson obtained based on the data and analytic methods he used.[9] Instead, Dr. Warren explained calculations showing an estimated drop of approximately 35,000 people in the City's residential population since 1980.[10] Based on these estimates, Dr. Warren challenged Dr. Henderson's reliance on 1980 census data in analyzing a 1987 election.
Dr. Warren acknowledged he would not be able to determine the drop in population for each ward.[11] Thus, the Court is not *1522 persuaded that the analyses provided by Dr. Henderson could be enhanced by taking into account any population decrease. (However, it could be argued that if the population trend for the City of St. Louis from 1950 to 1980 is continued through 2010, there will be nothing left in St. Louis but the Arch.) Moreover, without an analysis using the projected population figures, or a showing of how the population would affect the results obtained by Dr. Henderson based on 1980 census figures, neither of which are present of record, the Court is unwilling to reject Dr. Henderson's analyses on this basis alone.
32. Dr. Warren also reviewed data available for other elections within the City to point out that uncounted ballots (overvoting and undervoting) also occurred in "white" wards. In particular, Dr. Warren looked to U.S. Representative Clay's (1st Congressional District, north side of St. Louis; predominantly "black" wards) and to U.S. Representative Gephardt's (3rd Congressional District, south side of St. Louis; predominantly "white" wards) elections in 1980, 1982, 1984, and 1986. Dr. Warren concluded that in three out of four elections, the number of uncounted votes was lower in Representative Clay's elections than in Representative Gephardt's elections. See, e.g., Defendants' Ex.L.[12] Additionally, after two predominantly "white" wards (Wards 24 and 25) were added to Representative Clay's district in 1982, the next Clay election reportedly had approximately three times more over/undervotes in those "white" wards than the "black" wards within the district. See Defendants' Ex.M.[13] As a result of this analysis and its apparent contradiction of plaintiff's hypothesis, Dr. Warren posited that voter interest might play a key role in the extent of high over- or undervoting.
Therefore, Dr. Warren decided to look to data from a city that he considered predominantly black to see if voter interest would be different. He selected Atlanta, Georgia, a city he described as having a similar population level but dissimilar political status (black mayor, black leader of the city council which has a majority black membership) when compared to St. Louis. In using homogeneous analysis with respect to Atlanta, Dr. Warren reviewed 1980 census data and data from the 1985 mayoral and council president general elections, elections in which the punch card voting system was used. Dr. Warren determined that over/undervoting in Atlanta's "white" voting units was 18.2%, in Atlanta's "black" voting units was 6.9%, and city-wide averaged 10.4%. The three to one ratio "white"-to-"black" over- and undervoting was similar to the ratio of over/undervoting behavior in the white wards in Representative Clay's 1984 election. Thus, Dr. Warren concluded that "interest" is a factor in over- and undervoting.
Dr. Warren acknowledged that, traditionally, professionals determine "interest" by interviewing individuals, a methodology he did not use in either St. Louis or Atlanta. Additionally, Dr. Warren noted that for purposes of his study of Atlanta, he defined "black" or "white" wards as those *1523 having 98% of one race, rather than using the 90% standard he used for purposes of studying St. Louis voting data. Dr. Warren also testified that (a) the 1985 Atlanta mayoral race involved the incumbent black mayor and a white opponent who was characterized by Dr. Warren as "insignificant;" and (b) the 1985 Atlanta council president race involved an unopposed black candidate.
33. Dr. Warren found his hypothesis that voter interest is a factor affecting over/undervoting further borne out by over- and undervoting data for St. Louis' November 1986 general election, see Defendants' Ex. II,[14] and by a review of total votes cast in "black" wards for selected elections in which Dr. Warren considered black candidates or "black issues" significant, see Defendants' Ex. VV.[15] The Court is not persuaded by Dr. Warren's conclusions based on these exhibits. With respect to the graph marked as Defendants' Ex. II, Dr. Warren testified that it showed a small difference in over- and undervoting by blacks and whites both when a race "would not be as interesting to blacks as it would be to whites," Trial Transcript, Vol. III at 30 (reviewing United States Senate race data), and when "blacks were interested in [a] race," id. at 31 (discussing Congressman Clay's race data). Due to his apparently contradictory testimony, the Court does not find interest, or lack thereof, an explanation of his data. With respect to the graph depicting total votes cast by black wards for selected elections, Defendants' Ex. VV, Dr. Warren found that lower turnout in what he considered "important race[s]" (Roberts' 1983 and 1987 primary races for President of the Board) showed disinterest. The record, however, does not reflect how these elections were characterized as "important" to black voters nor the underlying data from which the data was obtained to show, for example, the turnout of white voters.
34. On the present record, the Court finds Dr. Henderson's conclusion that race was a significant factor in the extent of uncounted ballots from the March 3, 1987, St. Louis primary election more credible than Dr. Warren's determination that uncounted ballots are affected more by interest than by race. Dr. Warren's acknowledgement (a) that the usual manner of studying "interest" is by interviewing, and (b) that he had not used the interview method in either St. Louis or Atlanta, render his conclusions not as persuasive to the Court as Dr. Henderson's analyses, which were not questioned with respect to data or methodology. Additionally, the differences in the circumstances of the 1985 Atlanta elections analyzed by Dr. Warren and the 1987 St. Louis election analyzed by Dr. Henderson do not convince the Court that available data from the two cities' elections are proper for comparison.
35. To counter Dr. Warren's attack on plaintiff's reliance on an analysis of only one election, another of plaintiff's experts, Ms. Gabbert, provided undisputed figures regarding votes cast and not counted in the 1981, 1983, 1985, and 1987 municipal Democratic primary elections for the city-wide offices of mayor and president of the board of aldermen. See, e.g., Table, marked as "Appendix I," attached hereto and incorporated herein by reference, from Plaintiff's *1524 Ex. 6 at 1 (source citations omitted).[16] Ms. Gabbert's reported data show:
Uncounted Ballots Uncounted Ballots
In Black Wards As In White Wards As
% of Votes Cast % of Votes Cast
in Black Wards in White Wards
1981 Democratic
Mayoral Primary 4.4% 1.6%
1983 Democratic
Pres of B/A Primary 7.8% 2.4%
1985 Democratic
Mayoral Primary 7.1% 1.9%
1987 Democratic
Pres of B/A Primary 6.4% 1.9%
Thus, as shown above, in each primary, votes cast in "black" wards were about three times more likely not to be counted as votes cast in "white" wards.[17] These results do not contradict Dr. Henderson's analyses of the March 3, 1987, election data.
36. With respect to the City's primary election on March 3, 1987, only, Dr. Henderson used homogeneous ward analysis in the eight wards he found to be "black" and in the eleven wards he found to be "white," to see if there was any relationship between the race of the voter and of the candidate for whom a vote was cast. As a result of this analysis,[18] Dr. Henderson found "clear evidence of substantially significant racially polarized voting as that term is understood in Thornburgh [sic] v. Gingles [478 U.S. 30, 106 S. Ct. 2752, 92 L. Ed. 2d 25 (1986)]. The candidate of choice of minority [black] voters is definitely not the candidate of choice of majority [white] voters." These findings were supported by bivariate regression analysis conducted by Dr. Henderson. That analysis obtained estimates that (a) 98.57% of the white voters voted for Villa, and 1.43% of the white voters voted for Roberts; and (b) 100% of the black voters voted for Roberts, and 0% of the black voters voted for Villa in the March 3, 1987, primary election.
One of defendants' experts, Dr. George Wendel, a professor of political science and public policy at St. Louis University, St. Louis, Missouri, and Director of the Center for Urban Programs at that University, agreed with Dr. Henderson's conclusion that the City's March 3, 1987, primary election for President of the Board of Aldermen was characterized by racially polarized voting. Dr. Wendel further observed there have been elections in the City in which racially polarized voting has occurred.
The Board President also testified that, depending on the candidacy and the campaign, there sometimes is and sometimes is not racially polarized voting in St. Louis.
37. The public school system within the City of St. Louis has been found to be racially segregated, Adams v. United States, 620 F.2d 1277 (8th Cir.) (en banc), cert. denied, 449 U.S. 826, 101 S. Ct. 88, 66 L. Ed. 2d 29 (1980), and remains under a court-ordered desegregation plan, see e.g., Liddell v. Board of Educ., 491 F. Supp. 351 (E.D.Mo.1980) (mandatory plan within city), aff'd, 667 F.2d 643 (8th Cir.), cert. denied, 454 U.S. 1081, 1091, 102 S. Ct. 634, 656, 70 L. Ed. 2d 614, 629 (1981). Prior to 1954, Missouri's constitution required separate *1525 schools for black and white children. Mo. Const. Art. IX § 1(A) (1945). After exhaustively analyzing pertinent historical information, as well as the patterns and effects of student, faculty, and administrator assignments from 1954 to approximately 1980, the United States Court of Appeals for the Eighth Circuit in Adams, supra, found racially segregated schools throughout the City's public school system for which a system-wide remedy was required. Adams, supra, 620 F.2d at 1291. Pursuant to those and other rulings, the City's public school system remains under this court's jurisdiction. Liddell v. Bd. of Educ., No. 72-100C(5) (E.D.Mo., filed Feb. 18, 1972) (Hon. Stephen N. Limbaugh, presiding).
38. For purposes of this case only, the record further reflects that the City historically engaged in or authorized racial segregation in land use, see, e.g., City of St. Louis Revised Code of General Ordinances (1916), Chap. XXXIII, Art. XXI § 3819 (proscribing the use by persons of one race of a residence in any block consisting of residences of the other race); § 3821 (proscribing the use by persons of one race as a church, school, theater, dance hall, or assemblage hall of any building in a block consisting of residences of the other race); §§ 3823 and 3831 (setting forth criminal penalties for violation of the ordinances); and in public accommodations, see, e.g., Draper v. City of St. Louis, 92 F. Supp. 546 (E.D.Mo.1950) (City's prohibition against the use by blacks of the City's open air swimming pools found discriminatory).
39. With respect to recent housing information, plaintiff directed the Court's attention to St. Louis Currents The Community and Its Resources (undated), a booklet prepared by Leadership St. Louis, Plaintiff's Ex. 24 at 49-53. That report states in relevant part that from 1976 through 1983, approximately twice as many blacks as whites lived in substandard housing, whether as renters or as owners. Id. at 51. Additionally, based on the figures available of record regarding the City's residential population by ward, a significantly greater proportion of blacks live in north St. Louis, and a significantly greater proportion of whites live in south St. Louis.
40. The record reflects without dispute that (a) in 1979, the City closed Homer G. Phillips Hospital, located on the north side of the City, prior to the closing in 1985 of the City Hospital located on the south side of the City; (b) public health care operations have been consolidated, with St. Louis County's public health care operations, at St. Louis Regional Medical Center; and (c) for 1984, the Infant Mortality Rate for blacks (18.1%) in the City was approximately twice as high as for whites (10.2%), Plaintiff's Ex. 23 at 1.
41. One of defendants' experts, Dr. Wendel, made an unexplained statement that he did "not belive that [official segregation] has been the case in St. Louis." Trial Transcript, Vol. II at 162. The Court finds this blanket statement insufficient alone to contradict the evidence of record submitted by plaintiff, particularly when Dr. Wendel acknowledged during cross-examination that he had not reviewed a good portion of plaintiff's evidentiary materials on this issue,[19]see, id. at 200-04.
Where segregation was de facto and was not established by constitution or statute, the burden falls on the complainant to show the policy was deliberately established by officials. Where segregation was established by constitution or statute (de jure), the burden of showing its removal falls upon its practitioners. In Missouri's case, school segregation was imposed both by constitution and statute.
42. Census data for 1980, as reported by Dr. Henderson, see Plaintiff's Ex. 8, show that the City's black population between 18 and 65 years of age totals 134,750 persons and consists of 8,376 attending college, and 44.8% of those 25 years old and older being high school graduates. The median number of years of school completed *1526 by black city residents is reportedly 11.4 years. Based on the 1980 census data, there is reportedly a total of 196,977 white city residents between the ages of 18 and 65, consisting of 12,400 individuals attending college, and, of those 25 years old and older, 50.3% of the individuals are high school graduates. The median number of years of school completed by white city residents is reportedly 12.0 years. The reported census data also shows that of those city residents over 25 years old:
Years of School Completed White Black
Persons 25 years old & over 165810 105470
4 years of school or less 5406 7782
Percent 3.3 7.4
5-8 years of school 49084 24584
Percent 29.6 23.3
9-12 years of school 74636 53638
Percent 45.0 50.9
Some college 36684 19466
Percent 22.1 18.5
43. Based on 1980 census information, Dr. Henderson observed that approximately 53% of the "black" households in the City have annual incomes under $10,000, whereas 61% of the "white" households in the City have annual incomes at or above $10,000. Dr. Henderson reported that for the City's black households, the median annual income is $9,410; the mean annual income is $12,437; and the per capita income is $4,116. For "white" households in the City, those figures are $12,829; $16,123; and $7,373 respectively. Dr. Henderson characterized these figures as showing "a fairly dramatic difference" and "considerable disparity" between black and white households. Finally, from available 1980 census data, Dr. Henderson noted that "almost three times as many black as white families have incomes below the poverty level." Plaintiff's Ex. 8 at 15.
44. Dr. Henderson also reviewed census data on the City's labor force for 1979. See Plaintiff's Ex. 8. While the percentages of blacks (56.1%) and whites (55.3%) in the civilian labor force (those 16 years old and over) in St. Louis are very close, Dr. Henderson pointed out that the unemployment rate for blacks (17.7%) is approximately three times the unemployment rate for whites (6.5%). Additionally, Dr. Henderson noted that (a) blacks (at 50.3%) were much more likely to be unemployed for a long period of time (fifteen weeks or more) than whites (35.6%); and (b) "blacks are far more likely than whites to be in the low status low paying jobs." Trial Transcript, Vol. I at 174.
45. As a result of his review of available census data and his determination that on each indicator (education, income, labor force characteristics, occupation, and poverty status) the status of blacks in St. Louis is lower than the status of whites, Dr. Henderson concluded that black persons in St. Louis have a lower socioeconomic status than white persons.
46. No evidence or testimony of record clearly disputes the data on which Dr. Henderson relied or the conclusion he reached regarding City blacks' socioeconomic status. Dr. Wendel, one of defendants' experts, testified to a change in reporting census data for post-high school students, see Trial Transcript, Vol. II at 166-68, but did not directly contradict Dr. Henderson's data or results. For purposes of this case, the Court accepts Dr. Henderson's conclusion on the socioeconomic status of black city residents.
47. The record reflects there has not been a poll tax, literacy test, or ban against single-shot voting[20] affecting City blacks. *1527 Nor does this case involve unusually large election districts or a candidate slating process.
48. Voter registration procedures, such as the number of registration sites, the frequency of registration drives, and the failure to deputize volunteers, have been sustained in the face of constitutional challenges. Coalition for Sensible & Humane Solutions v. Wamser, 590 F. Supp. 217 (E.D.Mo.1984), aff'd, 771 F.2d 395 (8th Cir. 1985).
49. Graphs provided by Dr. Warren depict the following registration statistics for the City in 1980: 96,000 or 62% of the eligible voters in "white" wards were registered; 71,000 or 62% of the eligible voters in "black" wards were registered; and 39,000 or 63% of the eligible voters in "racially mixed" wards were registered. See Defendants' Exs. G and H. By February 4, 1987, those statistics had changed as follows: 90,000 or 61% of the eligible voters in "white" wards were registered; 83,000 or 79% of the eligible voters in "black" wards were registered; and 42,000 or 73% of the eligible voters in the "racially mixed" wards were registered. Id.
50. The present record does not establish any basis for finding voter registration has adversely affected black City voters.
51. It is undisputed that the campaign for the March 3, 1987, race for the Democratic nomination for President of the Board involved racial appeals.[21] While Dr. Wendel, one of defendants' experts, was "inclined to say" political campaigns in the City are not generally characterized by subtle or overt racial appeals, he observed that
electoral campaigns are very bitter in this City. Both blacks and whites are very fractionated in terms of their power basis. Frequently there is very great animosity during a campaign and after it. Certainly there are, frequently after a campaign is over some very bitter statements made back and forth.
Trial Transcript, Vol. II at 168-69. Dr. Wendel's observation that other political campaigns are devoid of racial appeals would be most credible perhaps to persons who were not in St. Louis during the recent campaign for the City school board. The Court finds, for purposes of this case, that racial appeals are used in City political campaigns.
52. While unlike Los Angeles, California; Newark, New Jersey; Atlanta, Georgia; Detroit, Michigan; Gary, Indiana; New Orleans, Louisiana; Chicago, Illinois; and Philadelphia, Pennsylvania, St. Louis blacks have never been able to elect a mayor, nevertheless, in aldermanic, congressional, and citywide races, they have enjoyed some success. Dr. Wendel opined that "[b]lacks have elected a very significant number of members of their own race to major offices in the City," so that they hold positions in numbers approximately proportionate to their population. Id. at 169. The uncontradicted record reflects that at present: (a) eleven of the City's twenty-eight aldermen are black; (b) seven of fifteen state representatives from the City are black; (c) two of four state senators from the City are black; (d) one of two U.S. Representatives from the City is black; (e) three (circuit clerk, recorder of deeds, and treasurer) of eleven (same plus mayor, comptroller, President of the Board, sheriff, circuit attorney, collector of revenue, license collector, and public administration) city-wide offices are held by blacks. Additionally, in the past fifteen years, *1528 blacks have held the positions of comptroller, sheriff, and Board President (by appointment).
The record does not have data clearly showing blacks' lack of success in winning city-wide elections.
53. The residential population of the City is approximately 46% black, and the City's voting age population is approximately 41% black.
54. Dr. Wendel stated there was a high degree of responsiveness by City officials to the needs of the City's black community. While during cross-examination he acknowledged there remain some problems, some of which he thought City officials could not do much about, no contradictory evidence on this issue is available of record.
55. Dr. Wendel also testified that to him, the policy behind the Board's use of the punch card voting system was not "tenuous," e.g., not adopted on a whim, but well-established and well-founded. The record contains no evidence to contradict this opinion.
56. Based on the results of his analyses, Dr. Henderson had the opinion that the operation of the ballot election system on March 3, 1987, denied black voters in the City the candidate of their choice, a "fair counting of their votes," and a "fair and reasonable access to the electoral system."
Dr. Wendel came to the conclusion that the "Zimmer" factors[22] he considered did not obstruct St. Louis "blacks in their exercise of their political rights or participation in the political process.
Dr. Warren opined that Roberts lost the March 3, 1987, primary election in large part because of the apparent lack of support given Roberts by several influential black political leaders and organizations, in particular Congressman Clay.[23] Dr. Warren also stated there is no evidence to indicate the punch-card voting system operates to discriminate against blacks.
Conclusions of Law
A. The Court has jurisdiction of this matter pursuant to 28 U.S.C. §§ 1331 and 1343.
B. The Court is not persuaded by intervenor defendant Villa's contention that plaintiff lacked standing to bring this action. Plaintiff allegedly suffered a distinct and palpable injury, the loss of an election, through the contested practice(s) at issue here and the alleged denial to black voters of an equal opportunity to participate in the political process. Accordingly, the Court finds plaintiff has standing to pursue this action. See Goodloe v. Madison County Bd. of Election Com'rs, 610 F. Supp. 240 (S.D.Miss.1985) (Voting Rights Act lawsuit brought by unsuccessful candidates).
C. Defendants also contend plaintiff's present challenge is untimely since the lawsuit was filed only four days prior to the primary election, and plaintiff allegedly could have sought timely pre-election relief. See, e.g., Toney v. White, 488 F.2d 310, 313-15 (5th Cir.1973) (en banc) (under circumstances, plaintiffs did not deliberately bypass pre-election relief requirement and plaintiffs' claim was properly considered although litigation was filed one month after challenged election occurred). While plaintiff's alleged failure to seek a timely pre-election remedy is not a helpful factor in considering the granting of eleventh hour or post-mortem relief,[24] the *1529 Court finds under the circumstances that defendants have not met the heavy burden needed to establish a deliberate bypass of that requirement. Id. at 315. Accordingly, the Court finds plaintiff's present claims are not barred. As noted herein, plaintiff needed to show a pattern affecting several elections and the new voting procedure was not used in the City until 1981.
D. The Voting Rights Act, 42 U.S.C. § 1973, provides:
(a) No voting qualification or prerequisite to voting or standard, practice, or procedure shall be imposed or applied by any State or political subdivision in a manner which results in a denial or abridgement of the right of any citizen of the United States to vote on account of race or color, or in contravention of the guarantees set forth in section 1973b(f)(2) of this title, as provided in subsection (b) of this section.
(b) A violation of subsection (a) of this section is established if, based on the totality of circumstances, it is shown that the political processes leading to nomination or election in the State or political subdivision are not equally open to participation by members of a class of citizens protected by subsection (a) of this section in that its members have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice. The extent to which members of a protected class have been elected to office in the State or political subdivision is one circumstance which may be considered: Provided, That nothing in this section establishes a right to have members of a protected class elected in numbers equal to their proportion in the population.
E. The central inquiry in, and the essence of, a challenge to an electoral structure or practice under § 2 of the Voting Rights Act, as amended, 42 U.S.C. § 1973, is whether, as a result of the challenged practice or structure, a protected group does not have an equal opportunity to participate in the political process and to elect candidates of that group's choice. Thornburg, supra, 106 S.Ct. at 2763. The issue for resolution here is whether the punch card voting system of voting, as used in the City of St. Louis, denies blacks an equal opportunity with whites to participate in the political process and to elect candidates of their choice.
F. The United States Supreme Court acknowledged that, to determine whether or not the persons protected by the Voting Rights Act have an equal opportunity to participate in the political processes and to elect candidates of their choice, a court must assess the impact of the contested practice or procedure based on several objective factors. Thornburg, supra, 106 S.Ct. at 2763. Those factors, also referred to as the "Zimmer" factors, include:
(a) the history of voting-related discrimination in the state or political subdivision;
(b) the extent to which voting in the elections of the state or political subdivision is racially polarized;
(c) the extent to which the state or political subdivision has used voting practices or procedures that tend to enhance the opportunity for discrimination against the minority group, such as unusually large election districts, majority vote requirements, and prohibitions against single-shot voting;
(d) the exclusion of the minority group from candidate slating processes;
(e) the extent to which minority group members bear the effects of past discrimination in areas such as education, employment, and health, which hinder their ability to participate effectively in the political process;
(f) the use of overt or subtle racial appeals in political campaigns;
(g) the extent to which members of the minority group have been elected to public office in the jurisdiction;
(h) the lack of responsiveness of elected officials to the particularized needs of the minority group's members; and
*1530 (i) the tenuousness of the policy underlying the use of the contested practice or structure.
Id. at 2763-64, citing to S.Rep. No. 97-417, at 28-29, U.S.Code Cong. & Admin.News 1982, pp. 177, 206-07. There is no requirement that a certain number of these factors be established or that these are the sole considerations. Id. at 2764.
G. For a challenge to the use of multi-member districts, a challenge not present here, the United States Supreme Court also requires a minority group to first demonstrate
[(a)] that it is sufficiently large and geographically compact to constitute a majority in a single member district[;] ... [(b)] that it is politically cohesive[; ... and (c)] that the white majority votes sufficiently as a bloc to enable it in the absence of special circumstances, such as the minority candidate running unopposed ... [] usually to defeat the minority's preferred candidate.
Thornburg, supra, 106 S.Ct. at 2766-67 (footnote and citations omitted). In a Voting Rights Act case challenging an at-large election system, a challenge also not present here, the United States Court of Appeals for the Eighth Circuit characterized the existence of these requirements as "a necessary precondition to establishing a claim of vote dilution." Buckanaga v. Sisseton Independent School Dist. No. 54-5, 804 F.2d 469, 474 (8th Cir.1986).
Without deciding whether or not such requirements are prerequisites to a challenge to an official vote counting method and procedure, the challenge present here, see Thornburg, supra, 106 S.Ct. at 2764-65 n. 12 (noting the court does not address whether the standards it finds applicable to plaintiffs' claim would apply to other vote dilution claims), the Court will consider these elements with respect to this case. The Court will then consider each of the "Zimmer" factors, without deciding whether or not they are required in a vote counting challenge such as is presented here. Id.
H. The Court finds the size of the black population in the City, as well as its location predominantly in the northern part of the City, satisfy the prerequisite that the minority population be "sufficiently large and geographically compact" to constitute a majority in a single member district, such as St. Louis for purposes of the elections at issue here. The Court has not been directed to and is unable to find any cases requiring that the minority population actually constitute a majority in the relevant voting unit. Nor does the Supreme Court's language necessarily equate the minority group's existence with a majority percentage in the population. If such a requirement existed, then a minority group would not be able to contest a voting unit procedure unless or until available population data showed a majority percentage for the group. Such a result is not clearly intended by relevant authority. Here, the black population does not constitute such a small percentage or such a dispersed group within the voting unit that it is inconceivable the minority group would ever constitute a voting majority.
I. Additionally, based on the present record, this Court finds the City's black population is politically cohesive and subject to white bloc voting. In Thornburg, the Supreme Court found that these issues are determined by inquiring into the existence of legally significant, racially polarized voting. Thornburg, supra, 106 S.Ct. at 2769. Here, there is no dispute that racially polarized voting, a touchstone of a vote dilution case, existed during the City's March 3, 1987, primary election and that it has existed in various earlier elections for citywide offices. Without dispute then, the Court finds that, for purposes of this case, there exists legally significant, racially polarized voting in St. Louis.
Any city which is unable to pass a bond issue for capital improvements for its schools in twenty-five years gives prima facie indications of an aberrational voting practice. While a two-thirds majority is required for such bond issues, Missouri's other 114 counties have all surmounted this hurdle.
J. The analyses and data available of record establish, and the Court finds, that *1531 the following relevant factors are not present in the City:
(1) a history of voting-related discrimination with respect to issues not now challenged and under consideration;
(2) unusually large election districts, majority vote requirements, prohibitions against single-shot voting, or voting practices or procedures (other than those presently challenged) that may tend to enhance the opportunity for discrimination; and
(3) exclusions of the minority group's members from any candidate slating processes.
K. In Buckanaga, the United States Court of Appeals for the Eighth Circuit noted that in a Voting Rights Act case, a minority group's present socioeconomic disadvantages may reduce that group's participation and influence in political affairs. Buckanaga, supra, 804 F.2d at 474, citing United States v. Marengo County Com'n, 731 F.2d 1546, 1567 (11th Cir.), appeal dism'd and cert. denied, 469 U.S. 976, 105 S. Ct. 375, 83 L. Ed. 2d 311 (1984). For purposes of the claims and relevant standards in this case only, the Court finds the black community in St. Louis City presently bears the effects of past discrimination in such areas as education, employment, housing, and health. This finding is borne out by the record's undisputed evidence that City blacks have a lower socioeconomic status than City whites, as reflected in several indicators (education, income, labor force characteristics, occupations, and poverty level).
L. For purposes of this case, the Court has found that political campaigns in the City are the subject of racial appeals. Such a characterization is undisputed with respect to the March 3, 1987, primary at issue here; and, in any event, it should not be necessary for "a serpent to sting thee twice."
M. Blacks have in recent years elected members to several citywide offices. Defendants point out that when state and federal congressional seats and board of alderman positions are considered, blacks have elected blacks to positions in numbers approximately proportionate to the City's black population. However, the success of some minority candidates does not foreclose a Voting Rights Act claim. Thornburg, supra, 106 S.Ct. at 2780. Here, the record shows the minority candidate's successes are relatively sporadic, rather than sustained, and have been only recently as successful as is evidenced by the present number of elected black officials.
N. In Buckanaga, the Eighth Circuit noted that officials' lack of responsiveness to a minority group is not a necessary factor in a Voting Rights Act case and may be raised "at plaintiff's discretion." Buckanaga, supra, 804 F.2d at 477. Here, this factor was not clearly addressed by plaintiff but was first raised through the testimony of one of defendants' experts, Dr. Wendel. Accordingly, the Court does not consider it with respect to plaintiff's present claim.
O. With respect to the last factor, "tenuousness" of the policy underlying the challenged practice or procedure, the record does not clearly establish the reason for not manually reviewing ballots not counted by the automatic tabulating equipment. The record discloses that the Board's change from the lever voting machines to the present punch card voting system may have been for administrative convenience and efficiency in terms of storage between elections and transportation of the equipment to and from the polling places. Moreover, without dispute the record shows the use of the new system is not "tenuous." That use in and of itself is, however, not at issue here. Rather, the resulting failure of the new system to prevent overvotes, the significant numbers of uncounted ballots in the black community since use of the new system, and the Board's apparent failure explicitly to address that problem are part and parcel of plaintiff's present claim. Without an explanation for the Board's failure manually to review ballots rejected by the equipment, and to see whether such ballots should in fact be counted rather than rejected, the Court is unable to say any such *1532 Board practice or policy is anything but tenuous.
P. In light of the present record, the Court sustains plaintiff's challenge to the Board's failure manually to review cast ballots rejected by the tabulating equipment and finds, under the circumstances, that that failure constitutes a violation of the Voting Rights Act. The record shows the Board's failure to review and, if appropriate, count such ballots has resulted in the City's black voters having less opportunity than other members of the City's electorate to participate in the political process and to elect representatives of their choice. It is not the use of the punch card voting system and automated tabulating equipment alone, but the Board's blanket failure to review rejected ballots, that resulted in the disenfranchisement of City voters. See Goodloe, supra, 610 F.Supp. at 243 (sustaining plaintiff's Voting Rights Act challenge to election officials' failure "to make a case by case or otherwise rational as opposed to arbitrary determination of which ballots were properly voted").
Q. With respect to plaintiff's claim that the Board's arbitrary rejection of cast ballots rejected by the automated tabulating equipment caused his defeat in the March 3, 1987, municipal primary, there is insufficient evidence of record that the rejected ballots resulted in plaintiff's defeat. In fact, there is no dispute that upon the three member panel's manual review of all challenged ballots, plaintiff still lost that election.[25] Thus, the present record does not necessitate either further review of ballots cast but not counted in that election, or other relief from this Court directed solely to that election.
Accordingly, after careful consideration, the Board will be enjoined from failing in the future (a) to take appropriate steps manually to count all ballots validly cast by voters of the City of St. Louis but rejected by the tabulating equipment; (b) to target for voter education on the proper use of the punch card voting system those wards from which the Board receives over a reasonable period of time a relatively high number of uncounted ballots (i.e., five percent of the total ballots cast in any ward); and (c) to offer to voters at the polling places explanations and demonstrations of the proper use of the punch card voting system. While judgment will be entered now, the Court will retain jurisdiction over this matter until June 30, 1988, for the limited purpose of issuing any orders which may be necessary to effectuate the relief granted herein. Plaintiff's request for attorney's fees will remain under consideration pending further briefing and consideration.
ORDER
Findings of fact and conclusions of law dated this day are hereby incorporated into and made a part of this order.
IT IS HEREBY ORDERED, ADJUDGED, AND DECREED that defendants Wamser, Krapf, Wrenn, and Robbins, in their capacities as members of the Board of Election Commissioners of the City of St. Louis (hereinafter Board), their officers, agents, servants, employees, and attorneys, are permanently enjoined from failing
(a) to take appropriate steps manually to count all ballots validly cast by voters of the City of St. Louis but rejected by the tabulating equipment;
(b) to target for voter education on the proper use of the punch card voting system those wards from which the Board receives over a reasonable period of time a relatively high number of uncounted ballots (five percent of the total ballots cast in such wards); and
(c) to offer to voters at the polling places explanations and demonstrations of the *1533 proper use of the punch card voting system.
The Court enters judgment this date, and will retain jurisdiction until June 30, 1988, for the limited purpose of issuing any orders which may be necessary to effectuate the relief granted herein.
IT IS HEREBY FURTHER ORDERED, ADJUDGED, AND DECREED that plaintiff take nothing by his cause of action against intervenor defendant Thomas A. Villa, and the claim against that defendant is dismissed on its merits.
IT IS HEREBY FURTHER ORDERED that on or before January 4, 1988, plaintiff shall file a brief (no more than ten pages in length) citing authority and argument in support of his position(s) on his request for an award of fees and costs. In addition to his brief, plaintiff shall file an itemization of fees and expenses he incurred, along with an explanation of the services rendered, the experience and background of each person rendering such services, the hourly rate of such services, and an explanation of the reasonableness of the rates charged, services rendered, and expenses incurred herein.
IT IS HEREBY FURTHER ORDERED that on or before January 15, 1988, defendants may file a brief (no more than ten pages in length) in response to plaintiff's attorney's fees and costs materials.
IT IS HEREBY FURTHER ORDERED that the parties, in cooperation with the United States Marshall, shall immediately take the steps necessary to return to the Board's custody the March 3, 1987, election ballots and related equipment that are presently secured pursuant to court orders entered in this case. To the extent necessary, any orders securing such materials are vacated solely for the purpose of allowing the return of such materials to the Board's custody.
Plaintiff's request for an award of fees and costs remains under consideration.
*1534 APPENDIX I
HISTORICAL ANALYSIS OF BALLOTS CAST AND VOTES NOT COUNTED IN DEMOCRATIC
PRIMARY ELECTIONS FOR CITYWIDE MUNICIPAL OFFICE
IN ST. LOUIS, MISSOURI
Total Ballots Cast Black Vote Ballots Cast White Vote Ballots Cast Mixed Vote Cumulative
Type Of Democratic in As % Of in As % Of in Racially As % Of Democratic
Election Ballots Cast Black Wards Total Vote White Wards Total Vote Mixed Wards Total Vote Vote Counted
1981 Democratic
Mayoral Primary 108,481 38,777 35.7% 52,090 48.0% 17,614 16.2% 105,427
1983 Democratic
Pres of B/A Primary 62,887 25,879 41.2% 25,536 40.6% 11,472 18.2% 59,656
1985 Democratic
Mayoral Primary 87,840 30,263 34.5% 39,759 45.3% 17,818 20.3% 84,249
1987 Democratic
Pres of B/A Primary 75,516 26,150 34.6% 34,962 46.3%[A] 14,404 19.1%[A] 72,562
Ballots Black Ballots Black Not Counted White Ballots White Not Counted Mixed Ballots Mixed Not Counted
Type of Cast, Votes Cast, Votes As % Of Cast, Votes As % Of Cast, Votes As % Of
Election Not Counted Not Counted Total Not Counted Not Counted Total Not Counted Not Counted Total Not Counted
1981 Democratic
Mayoral Primary 3,054 1,723 56.4% 858 28.1% 473 15.5%
1983 Democratic
Pres of B/A Primary 3,231 2,014 62.3% 616 19.1% 601 18.6%
1985 Democratic
Mayoral Primary 3,591[B] 2,151 59.9% 742[B] 20.7% 698 19.4%
1987 Democratic
Pres of B/A Primary 2,864 1,677 58.6% 657 22.9% 530 18.5%
NOTES
[1] Pursuant to plaintiff's counsel's representations in open court at the start of trial, plaintiff's claims under 42 U.S.C. § 1983 and the federal constitution are dismissed. See also order dated October 6, 1987.
[2] The relevant statutory provisions, Mo.Rev. Stat. §§ 115.225 and 115.229 (1986), provide such certification must occur prior to use of such equipment. The certificate of record, to which plaintiff has objected for lack of foundation, indicates the tabulating equipment used for the March 3, 1987, primary election was certified by the Secretary of State as of April 22, 1987.
[3] Twenty-one of those precincts were either plus or minus one vote; and two of those precincts were plus or minus two votes.
[4] In particular, the record reflects the following undisputed totals for ballots cast, but not counted, in those primaries:
"Black" "White" "Racially
Total Wards Wards Mixed" Wards
1981 Democratic mayoral primary 3054 1723 858 473
1983 Democratic Board President primary 3231 2014 616 601
1985 Democratic mayoral primary 3592A 2151 743A 698
1987 Democratic Board President primary 2864 1677 657 530
A These figures should be 3591 and 742, respectively. See Appendix I attached hereto.
Plaintiff's Ex. 6, p. 1 (footnote added). For purposes of these figures, the "black" wards are wards 1, 3, 4, 5, 18, 19, 20, 21, 22, 26, and 27; the "white" wards are wards 9, 10, 11, 12, 13, 14, 15, 16, 23, 24, and 25; and the "racially mixed" wards are wards 2, 6, 7, 8, 17, and 28. See Plaintiff's Ex. 6.
[5] The available data show the following totals of uncounted ballots in each ward for the City's March 3, 1987, primary election:
Ballots with Ballots with
Ward No-Vote Ward No-Vote
01 142 15 28
02 65 16 69
03 112 17 97
04 269 18 201
05 75 19 145
06 86 20 233
07 86 21 128
08 65 22 141
09 44 23 71
10 75 24 88
11 42 25 39
12 59 26 129
13 48 27 102
14 94 28 131
Plaintiff's Ex. 6, p. 9.
[6] Homogeneous precinct or, here, ward analysis involves the comparison of analyzed conduct within voting units in which the predominant number of potential voters are of one race. The resulting figures "serve as estimates of the behavior of all the respective group members in that political jurisdiction." Engstrom & McDonald, Quantitative Evidence in Vote Dilution Litigation: Political Participation and Polarized Voting, 17 Urban Lawyer 369, 371 (Summer 1985). Unlike homogeneous analysis, which requires the presence of homogeneous voting units and does not use information from non-homogeneous voting units in the relevant area, bivariate regression analysis looks at information about the analyzed voting behavior in all the voting units within the relevant area. See id. at 373-74. This analysis will result in estimates of the degree of relationship between two variables being analyzed. See id. These methods have been found standard for analysis of racially polarized voting. See, e.g., Thornburg v. Gingles, 478 U.S. 30, 106 S. Ct. 2752, 2768, and 2768 n. 20, 92 L. Ed. 2d 25 (1986). While it is not clear that such techniques are as accepted for analysis of votes cast and counted or not counted, no party or expert disputed the propriety of applying these techniques to this issue, cf. n. 9, infra, and the contentions raised in this case are of first impression. Thus, the Court finds it appropriate to permit consideration of such analyses to cast but uncounted ballots.
[7] With respect to the homogeneous ward analysis, Dr. Henderson found by reviewing (1) the reported numbers of ballots not counted in each ward, and (2) each such number expressed as a percentage of total votes cast, see Tables 1 and 2 of Plaintiff's Ex. 8, pp. 1-2, that both the numbers and percentages of votes cast but not counted were much higher in "black" wards than in "white" wards. His reported data show that (a) the actual number of votes cast but not counted in his eight "black" wards totals 1,355 such votes, ranging from a low of 112 such votes to a high of 269 votes cast but not counted; (b) the actual numbers of votes cast but not counted in his eleven "white" wards totals 657 such votes, ranging from a low of 28 such votes to a high of 94 such votes; (c) the percentages of such votes in relation to total votes cast in each of his eight "black" wards ranges from a low of 4.01 to a high of 9.35; and (d) the percentages of such votes in relation to total votes cast in each of his eleven "white" wards ranges from a low of .95 to a high of 3.21.
The 657 total votes for the "white" wards noted above is three votes more than would be obtained by adding the relevant figures reported by Dr. Henderson in his Table 1. This difference reflects a correction in the number of votes not counted in Ward 9 that he reported as 41, when the record reflects this figure should be 44. See, e.g., n. 5, supra. The figures for Ward 9 (44 votes not counted, which is approximately 2.42% of the total vote cast), are in the midrange of the figures for the "white" wards. Therefore, the Court finds the three-vote difference here is not so appreciable as to warrant drastic changes in Dr. Henderson's present analyses.
The Court finds Dr. Henderson's homogeneous analysis of votes cast but not counted holds true even if data for Wards 5, 19, and 27 (three additional wards deemed "black" by Ms. Gabbert and Dr. Warren but considered "racially mixed" by Dr. Henderson) are considered. Those wards have the following reported data:
Number of Ballots Percentage of
Ward Not Counted Total Votes
05 75 5.46
19 145 6.80
27 102 4.38
With these wards, the actual number of votes cast but not counted in the eleven "black" wards totals 1,677 (1,355 from the original eight wards, plus 322 from the three additional wards), ranging from a low of 75 such votes to a high of 269 votes cast but not counted, and all other summaries noted in paragraphs (b) through (d) above remain unaffected. Thus, the lowest number or percentage of uncounted ballots in the "black" wards is still much higher than the lowest in the "white" wards and other comparisons are unchanged.
As a result of the bivariate regression analysis, Dr. Henderson found that 7.22% of the ballots cast by black voters were not counted, whereas only 1.77% of ballots cast by white voters were not counted. While Dr. Henderson found a "very strong" relationship between the amount of ballots not counted and the amount of black voting age population, he variously described that relationship as "positive," Plaintiff's Ex. 8 at p. 3, and as "negative," Trial Transcript, Vol. I, at 177. Despite this apparent inconsistency in that characterization of the relationship, Dr. Henderson's conclusion from the bivariate regression analysis remained the same. Since bivariate regression analysis takes all wards into account, there would be no change in these results due to Dr. Henderson's omission of three wards from his consideration of what are "black" wards.
[8] In his testimony, Dr. Henderson mentioned he analyzed the March 1985 mayoral primary election. That analysis, however, is not available of record.
[9] One of defendants' experts, Dr. Warren, stated he thought regression analysis was not necessary and was inappropriate here because of the availability of what he termed "better" approaches, e.g., homogeneous analysis. While the several experts may dispute which approach is "better" under the circumstances, Dr. Henderson subjected the available data to both analyses, so Dr. Warren's present critique does not support rejection of Dr. Henderson's analyses.
[10] The graphs used by Dr. Warren to depict the estimated population shifts show that in 1980, census figures reflected the residential population of the "white" wards was 196,533 and the voting age population of such wards was 156,342; and in 1987, the figures were estimated by Dr. Warren to be 183,696 in residential population and 148,592 in voting age population. See Defendants' Exs. E and F, respectively. For the "black" wards, (a) census figures showed that in 1980, the residential population was 173,619 and the voting age population was 115,684; and (b) Dr. Warren's calculations showed that in 1987, the residential population was 155,812 and the voting age population was 105,556. See Defendants' Exs. E and F.
[11] The Court notes that, without explanation, tables provided by defendants reporting cast and counted votes in the 1981 City primary and general elections, Defendants' Exs. Z and AA respectively, identify Ward 25 as a "racially mixed" ward rather than as a "white" ward, which is how defendants characterize it for purposes of the 1980 municipal primary, Defendants' Ex. B(1), for the 1985 municipal primary and general elections, Defendants' Exs. BB and BB(1) respectively, and for the 1987 City primary, Defendants' Ex. CC. Defendants did not provide any justification for the apparent population shifts to explain the apparently abrupt changes in classifying Ward 25. In data provided by Ms. Gabbert, one of plaintiff's experts, Ward 25 remained classified as "white" from 1981 through 1987. See Plaintiff's Ex. 6. In the absence of explanation, the Court considers it a "white" ward.
[12] The exhibits on which Dr. Warren relies to support or explain his position(s) regarding the over- and undervoting pertaining to Representative Clay's and Representative Gephardt's primary elections in 1980, 1982, 1984, and 1986, Defendants' Exs. J, K, L, and M, are graphs reportedly depicting over- and undervoting percentages for "black" and "white" wards in those elections. Neither the graphs nor Dr. Warren's testimony explain what the percentages are (e.g., percentage of total ballots cast but not counted in each election, percentage of all ballots cast in each election, or percentage of total ballots cast but not counted for the predominantly "black" or predominantly "white" wards in each election). The Court assumes the graphs depict percentages of uncounted ballots in relation to the total ballots cast, a percentage used by Dr. Henderson, see n. 7, supra.
[13] See discussion in footnote 12, supra.
[14] While neither the exhibit nor Dr. Warren's testimony, see Trial Transcript, Vol. III at 29-32, clarifies that Defendants' Ex. II depicts over- and undervoting in a St. Louis, Missouri, November election, the Court understands from Dr. Warren's reference to the graphically depicted congressional first district race as "Clay's race," Trial Transcript, Vol. III at 30-31, that the exhibit is based on data from the St. Louis, Missouri, November 1986 election. Additionally, the graph depicts percentages which have not been explained. See, e.g., n. 12, supra.
[15] Neither the graph depicting voter turnout in the "black" wards, Defendants' Ex. VV, nor Dr. Warren's testimony thereon, Trial Transcript, Vol. III at 36-37, explicitly clarified that the reported figures are for St. Louis, Missouri, elections. The Court understands those are the elections from which data is reported due to Dr. Warren's testimony referring to "Bosley" and "Roberts," St. Louis area candidates, and the exhibit's mention of "Homer Phil," an apparent reference to an election focusing on Homer G. Phillips Hospital, a hospital located in north St. Louis.
[16] Based on the record, defendant Villa's objections to Plaintiffs Ex. 6 for lack of foundation, relevancy, and hearsay are denied.
[17] These findings are further supported by Defendants' Exs. WW and XX, to which plaintiff objects for lack of foundation. These exhibits purport to depict by graph the percentages in "black" wards and in "white" wards of over- and undervoting in the City's 1982 and 1984 general elections. The graphs show only one race or issue (out of seven races or issues in 1982 and out of five races or issues in 1984) in which over- and undervoting was at a higher percentage in the "white" wards than in the "black" wards. As noted earlier with respect to Defendants' Exs. J, K, L, and M, see footnote 12, supra, it is not clear what the depicted percentages represent (e.g., are they percentages of the total number of ballots cast but not counted in each election, or percentages of total ballots cast but not counted in "black" wards or "white" wards for each election, or are they percentages of the total ballots cast either in all wards or in wards of predominantly one race for each election). The Court does not rely on these exhibits for its findings in this case.
[18] While not necessarily sufficient to change either the results of the homogeneous analysis or Dr. Henderson's opinion, the Court notes that Table 4 of Plaintiff's Ex. 8, the table which sets forth data relevant to this analysis, reports that 35 votes for Villa were cast in Ward 3. The figure relied on by defendants and by plaintiff's other expert, is 55 votes for Villa cast in that ward. See Defendants' Ex. CC and Plaintiff's Ex. 6 at p. 9, respectively.
[19] Defendant Villa's lack of foundation and hearsay objections to what is now Plaintiff's Ex. 20, Byran O. Jackson, "The Effects of Racial Group Consciousness on Political Mobilization in American Cities" (undated) [what was originally listed as Plaintiff's Ex. 17], are sustained. The Court has not considered that study for the findings made here.
[20] As noted in Thornburg, supra, 106 S.Ct. at 1260 n. 5:
Bullet (single-shot) voting has been described as follows:
"`Consider [a] town of 600 whites and 400 blacks with an at-large election to choose four council members. Each voter is able to cast four votes. Suppose there are eight white candidates, with the votes of the whites split among them approximately equally, and one black candidate, with all the blacks voting for him and no one else. The result is that each white candidate receives about 300 votes and the black candidate receives 400 votes. The black has probably won a seat. This technique is called single-shot voting. Single-shot voting enables a minority group to win some at-large seats if it concentrates its vote behind a limited number of candidates and if the vote of the majority is divided among a number of candidates.'" City of Rome v. United States, 446 U.S. 156, 184, n. 19 ... [100 S. Ct. 1548, 1565, 64 L. Ed. 2d 119] (1980), quoting U.S. Commission on Civil Rights, The Voting Rights Act: Ten Years After, pp. 206-207 (1975).
[21] One such appeal was a written statement published by plaintiff's campaign stating that a white candidate other than intervenor defendant Villa was backed by "the Klan," an apparent reference to the Ku Klux Klan, a secret organization advocating white supremacy. Plaintiff urges certain information published by Villa's campaign, see Ex. 14 (a letter signed by local Republicans urging voters to vote for Villa despite political party affiliation so as to "undermine the deal struck by [Congressman] Clay, [candidate] Osborn and [candidate] Roberts") and Ex. 15 (a written media announcement purportedly questioning plaintiff's business capabilities due to plaintiff's support of a local black politician's effort to obtain a city contract the politician had lost), also constituted racial appeals. The Court is not persuaded that the latter were "racial" appeals.
[22] The "Zimmer" factors are factors deemed typical of Voting Rights Act claims in light of Zimmer v. McKeithen, 485 F.2d 1297 (5th Cir. 1973), aff'd sub nom. East Carroll Parish School Bd. v. Marshall, 424 U.S. 636, 96 S. Ct. 1083, 47 L. Ed. 2d 296 (1976) (per curiam). See Thornburg, supra, 106 S.Ct. at 2759-60.
[23] The parties disputed (a) whether or not Congressman Clay in fact "supported" plaintiff during the campaign preceding the March 3, 1987, city primary election, and (b) if so, to what extent the support existed. The record most persuasively indicates plaintiff received as much support from Representative Clay as plaintiff wanted, and that plaintiff did not want further support.
[24] This Court knows that 5:00 p.m. Friday requests for temporary restraining orders and other instant or even ex parte emergency relief come only from unexpected situations and not from a counsel's wish to have a judge not have a fair opportunity to consider the case pro and con. The Court is concerned that members of the press and public, less tempered by the fires of judicial experience, may construe eleventh hour exhortations less charitably.
[25] The trial court having completed its use for and examination of the March 3, 1987, primary election ballots and related equipment, the same should be returned promptly to the Board subject to any subsequent requirements of the United States Court of Appeals for the Eighth Circuit should this matter be appealed.
The Court commends the local United States Marshals for their diligence and for their willingness to carry out the important responsibility of securing the ballots and equipment, and of supervising the court-ordered recounts.
[A] A mathematical check of the figures provided in this table reveals that the correct percentage of "White Vote as % of Total Vote" for the 1987 Democratic primary election for President of the Board of Aldermen should be 46.3% rather than 45.1%; and that the correct percentage of "Mixed Vote as % of Total Vote" for the same election should be 19.1% rather than 18.6%.
[B] Based on a review of the relevant source materials provided with this table and other relevant materials of record, the Court finds that (1) the total of ballots cast but not counted for the 1985 municipal mayoral primary should be 3,591 rather than 3,592; and (2) the total white ballots cast but not counted should be 742 rather than 743. This change is due to an apparent mathematical error in calculating the latter number (39,759 total ballots cast in "white" wards, less 39,017 ballots counted in "white" wards, see, e.g., Defendants' Ex. BB(1), equals 742). The corrected total coincides with the number obtained by subtracting the totals given as source materials for this part of the table (87,840 total ballots cast less 84,249 ballots counted), see Plaintiff's Ex. 6 at 15.