TABLE OF CONTENTS

I. SUMMARY OF FINDINGS 1303 II. FINDINGS OF FACT A. THE PARTIES 1305 B. THE CLAIMS 1306 C. FACTUAL BACKGROUND 1. History of the Governing Body 1306 2. Demographics of Los Angeles County 1307 3. Campaign Financing 1308 4. Prior Redistrictings 1309 (a) 1959 Redistricting 1309 (b) 1963 Redistricting 1310 (c) 1965 Redistricting 1310 (d) 1971 Redistricting 1311 (1) Intent of Past Redistrictings 1312 (e) 1972 Los Angeles City Council Redistricting 1313 (f) 1981 Redistricting 1313 (1) Intent of 1981 Redistricting 1317 D. SIZE AND GEOGRAPHIC COMPACTNESS OF HISPANIC COMMUNITY 1. 1980 Census Data 1318 2. Growth in Hispanic Population Since 1980 1319 3. Accuracy of Post-Census Data 1320 (a) Reliability of PEPS Data 1321 4. Citizen Voting Age Population 1322 5. Voter Registration and Turnout 1323 6. Misreporting of Citizenship 1323 7. Undercount of Hispanics 1325 8. Spanish-Surname/Spanish-Origin 1325 (a) Adjustments for "European Spanish" 1326 9. Deadwood 1327 10. Plaintiffs' Illustrative Plans 1328 (a) The Grofman Plans 1328 (b) The Estrada Plans 1328 E. POLITICAL COHESIVENESS 1. Hispanic Candidacies in Los Angeles County 1978-1989 (a) Contests for Los Angeles County Supervisor 1328 (b) Other Nonpartisan County Contests 1329 II. FINDINGS OF FACT E. POLITICAL COHESIVENESS 1. Hispanic Candidacies in Los Angeles County 1978-1989 (c) Non-Countywide Elections 1330 (d) Countywide Partisan Elections 1331 2. Analysis of Ethnically Polarized Voting (a) Methodology 1331 (b) Results of Analysis 1333 3. Cohesiveness of Hispanic Voters 1335 F. NON-HISPANIC BLOC VOTING 1337 G. OTHER SENATE FACTORS 1. History of Official Discrimination 1339 (a) Repatriation 1340 (b) Education 1340 (c) Public Facilities 1340 (d) Right to Vote 1340 2. Racial Appeals 1341 3. Size of Election Districts 1341 III. CONCLUSIONS OF LAW A. JURISDICTION 1342 B. THE VOTING RIGHTS ACT 1342 1. The Senate Factors 1342 (a) Geographical Compactness 1344 (1) Voting Age Population 1344 (2) Current Population Data 1345 (3) Estimates and Projections 1345 (b) Political Cohesiveness 1345 (1) Ecological Regression Analysis 1346 (c) Racial Bloc Voting 1346 (d) History of Discrimination 1347 (e) Other Discriminatory Voting Practices 1348 (f) Size of Election Districts 1348 (g) Candidate Slating Process 1348 (h) Lingering Effects of Past Discrimination 1348 (i) Election of Minorities 1348 C. DISCRIMINATORY RESULTS V. INTENT 1348 D. INTER-DECENNIAL REDISTRICTING 1350 E. TOTAL POPULATION AS APPORTIONMENT BASE 1350 F. ONE PERSON ONE VOTE RULE 1350 G. REAPPORTIONMENT 1351

FINDINGS OF FACT AND CONCLUSIONS OF LAW

KENYON, District Judge.

I. SUMMARY OF FINDINGS

The Court has spent the past several weeks since the conclusion of this trial on April 10, 1990, immersed in what the Supreme Court in Thornburg v. Gingles referred to as a "searching evaluation of `past and present reality'" and on a "`functional' view of the political process." 478 U.S. 30, 45, 106 S.Ct. 2752, 2764, 92 L.Ed.2d 25 (1986) citing S.Rep. at 30, n. 120, U.S. Code Cong. Admin. News 1982, pp. 177, 208. The conclusion this Court reaches is that, on a fundamental level, the Hispanic community has sadly been denied an equal opportunity to participate in the political process and to elect candidates of their choice to the Board of Supervisors for this burgeoning County.

As the findings below set forth, plaintiffs have adequately demonstrated, based on the totality of the circumstances, that the 1981 redistricting plan adopted by the Board of Supervisors violated Section 2 of the Voting Rights Act and the equal protection clause of the Fourteenth Amendment.

Specifically, the Court finds that the Hispanic community is sufficiently large and geographically compact such that a five district plan can be drawn in which Hispanics comprise a majority of the citizen voting age population in one of the five districts. The post-1980 estimates of citizen voting age population, based upon PEPS data and the special tabulation of voting age citizens by the Census Bureau, are reliable as an alternative means of proof that under current conditions it is possible to create a supervisorial district with an Hispanic citizen voting age population majority.

Further, even if the Court were to use 1980 Census data, plaintiffs have established through illustrative plans that Hispanic voting age citizens had the potential to elect the candidate of their choice absent a clear citizen voting age majority. It would be myopic, on these facts and circumstances, for the Court to apply the bright line 50 percent requirement set forth by the Ninth Circuit in Romero v. City of Pomona, 883 F.2d 1418, 1426 (9th Cir. 1989), as an absolute measure of undiluted minority voting strength. While this Court can imagine a number of circumstances in which the 50 percent figure is dispositive, as Justice O'Connor stated in her concurring opinion in Gingles:

"[T]here is no indication that Congress intended to mandate a single, universally applicable standard for measuring undiluted minority voting strength, regardless of local conditions and regardless of the extent of past discrimination against minority voters in a particular State or political subdivision."

In this case, the explosive and continuous growth of the Los Angeles County Hispanic community was evident at the time of the adoption of the 1981 redistricting plan as was the steady decline of the County's non-Hispanic white population. These facts, coupled with a long and painful history of discrimination against Hispanics in this County weighs heavily in favor of the conclusion that even relying solely on the 1980 Census data, plaintiffs have met their burden under Gingles.

The Court also finds that Hispanics are politically cohesive and that voting behavior is polarized between Hispanics and non-Hispanics. In particular, the Court concludes that Hispanic voters regularly provide overwhelming support for Hispanic candidates while the degree of non-Hispanic cross-over voting is minimal. Given the estimated levels of polarization, including the effects of non-Hispanic bloc voting, an Hispanic candidate is unable to be elected to the Board under the current configuration of supervisorial districts.

During the 1981 redistricting process, the Supervisors' primary objective was protect their incumbencies and that of their allies. This objective, however, was inescapably linked to the continued fragmentation of the Hispanic population core. The Court believes that had the Board found it possible to protect their incumbencies while increasing Hispanic voting strength, they would have acted to satisfy both objectives. As defendants' counsel argued in opening statement:

"It was not, . . . the case of a Republican protecting [his] incumbency against the Hispanic Republican. It was the Republican protecting himself or protecting his philosophical concerns and those of the ones who elected him from a change to a Democratic seat. . . . Now looking again to the motive of minority members on the Board of Supervisors. Again what you find is that it was not an effort by the Anglos to preclude Hispanics from getting elected. . . . It was not because of a desire on anyone's part to dilute or diffuse or to keep the Hispanic community powerless; it was because they could not find the way to do what everyone wanted to do. And that sometimes happens in politics."

It is undeniable, however, that the Los Angeles County Board of Supervisors knew that by adopting the 1981 redistricting plan, they were further impairing the ability of Hispanics to gain representation on the Board. The Court finds no legal justification for this form of discrimination based on the protection of supervisorial incumbencies.

As the court stated in Rybicki v. State Board of Elections, 574 F. Supp. 1082, 1109 (N.D.Ill. 1982), where the requirements of incumbency "are so closely intertwined with the need for racial dilution . . . an intent to maintain a safe, primarily white, district . . . is virtually coterminous with a purpose to practice racial discrimination." The Court finds, on the evidence presented, that the Supervisors acted with the intent to maintain the fragmentation of the Hispanic vote.

Throughout this trial, the Court heard extensive testimony regarding the size of the supervisorial districts. The Court strongly believes, as one Supervisor testified, that the districts are now too large for any one person to adequately represent. The Court believes that expansion may well be in the best interest of all concerned. However, the Court finds that while the size of the districts contributes significantly to the inability of Hispanics to elect a candidate of their choice, plaintiffs have failed to establish a valid legal claim based solely on the size of the supervisorial districts.

Since the task of reapportionment is properly a legislative function, it is appropriate, in this case, to allow the Board of Supervisors a reasonable opportunity to meet constitutional requirements by adopting a substitute measure. Wise v. Lipscomb, 437 U.S. 535, 540, 98 S.Ct. 2493, 2497, 57 L.Ed.2d 411 (1978). It is the sincere hope of this Court that in fashioning a suitable remedy, defendants will carefully reconsider the issue of expansion.

II. FINDINGS OF FACT

A. THE PARTIES

1. The United States of America is the plaintiff in the consolidated case, No. CV 88-5435 KN, United States of America v. County of Los Angeles, et al. The United States was represented by Steven H. Rosenbaum, of the Voting Section, Civil Rights Division of the Justice Department.

2. Hispanic registered voters in Los Angeles County are the plaintiffs in this consolidated class action, No. CV 88-5143 KN, Yolanda Garza, et al. v. County of Los Angeles, et al. The class representatives include: plaintiff Yolanda Garza, a resident of Supervisorial District One; plaintiff Salvador H. Ledezma, a resident of Supervisorial District Two; plaintiff Raymond Palacios, a resident of Supervisorial District Three; plaintiff Guadalupe De La Garza, a resident of Supervisorial District Four; and plaintiff Monica Tovar, a resident of Supervisorial District Five. All are United States citizens of Spanish heritage and registered voters in Los Angeles County, California. (hereinafter the "Garza plaintiffs"). Richard P. Fajardo, of the Mexican American Legal Defense and Educational Fund (MALDEF), and Mark D. Rosenbaum of the American Civil Liberties Union (ACLU) represented the Garza plaintiffs.

3. Defendant Los Angeles County is a political subdivision of the State of California established under the laws of the State and the Charter of the County of Los Angeles. Los Angeles County is subject to the requirements of the Voting Rights Act of 1965, as amended, Pub.L. No. 97-205, § 3, 96 Stat. 134 (1982), codified at 42 U.S.C. § 1973, et seq.

4. Defendants Edmund D. Edelman, Board Chairman; Peter F. Schabarum, Kenneth Hahn, Deane Dana, and Michael D. Antonovich, are duly elected members of the Board of Supervisors of the County. All are white non-Hispanic persons.

5. Defendant Charles Weissburd is the Registrar-Recorder of Los Angeles County responsible for the conduct of elections in the County, including elections for positions on the Board of Supervisors. Mr. Weissburd is a white non-Hispanic person sued in his official capacity.

6. Defendant Richard B. Dixon is the administrative officer of Los Angeles County and has primary responsibility for the conduct of day-to-day County affairs including oversight and implementation of County and State election laws. Mr. Dixon is a white non-Hispanic person sued in his official capacity.

7. Defendant Frank F. Zolin, named as a defendant by the Garza plaintiffs, is the Clerk/Executive Officer for the County responsible for conducting County elections.

8. Defendants were represented by John McDermott, Lee Blackman, and Richard Simon, of McDermott, Will Emery.

B. THE CLAIMS

9. Both the United States and the Garza plaintiffs challenge the 1981 redistricting plan (hereinafter "the 1981 Plan") under the authority of Section 2 of the Voting Rights Act, 42 U.S.C. § 1973 (hereinafter "the Act").

10. The Garza plaintiffs bring this class action pursuant to Federal Rule of Civil Procedure 23(b)(2) on their own behalf and on behalf of all Hispanic citizens whose right to vote has been or will be abridged by the adoption and maintenance of the 1981 Plan.

11. The Garza plaintiffs also challenge the 1981 Plan on the grounds that it was adopted and/or maintained for the purpose of discriminating against Hispanic citizens in violation of Section 2 of the Act and the Fourteenth and Fifteenth Amendments to the United States Constitution.

12. The Garza plaintiffs allege that the presence of only five supervisorial districts results in the dilution of Hispanic voting rights in violation of Section 2 of the Act and the Fourteenth and Fifteenth Amendments.

C. FACTUAL BACKGROUND

1. History of the Governing Body

13. Los Angeles County was one of the original 27 counties formed in 1850 by the California Legislature.

14. The first Board of Supervisors was elected in 1852.

15. Los Angeles County has been governed by a five-member Board of Supervisors at all times except for a two-year period between 1883 and 1885, when the Board consisted of seven members.

16. Under the provisions of a charter adopted in 1912, Los Angeles County was granted home rule power and was divided into five supervisorial districts. The charter became effective in 1913.

17. Since at least 1914, the Supervisors have been elected during even-numbered years in nonpartisan elections. If no candidate receives a majority of the votes cast in a June primary, the two candidates who receive the highest number of votes oppose each other in a general election in November of that year.

18. Supervisors are elected for four-year, staggered terms.

19. Elections for Supervisor in Districts 2, 4 and 5 were held in 1988. Elections for Supervisor in Districts 1 and 3 are scheduled to be conducted in 1990.

20. The Los Angeles County Board of Supervisors has legislative, executive and quasi-judicial powers.

21. The Board of Supervisors has authority pursuant to state law to alter, with voter approval, the size of the governing body.

22. Pursuant to the charter of the County of Los Angeles, the Board of Supervisors has authority, within one year after a general election, to redraw the boundaries of the supervisorial districts. Charter of the County of Los Angeles, Art. II, Sec. 7 (1985).

23. Los Angeles County is responsible for providing certain classes of governmental services to all County residents including health services, courts, elections and welfare.

24. Los Angeles County is also responsible for providing full municipal services to residents of the unincorporated areas of the County, including fire protection, law enforcement, planning, zoning and building inspection.

25. Supervisor Edelman agreed with the finding in "To Serve Seven Million," a 1976 report of the Public Commission on Los Angeles County Government, that "[n]o other local official in the United States is assigned responsibilities of the breadth and scale of those afforded a Los Angeles County Supervisor."

26. Los Angeles County has a contracting program to provide certain services to cities requesting these services. As a result of the contracting program, the County provides a significant portion of local governmental service to all County residents.

27. The Board of Supervisors has the authority to adopt the County's budget, appropriate funds pursuant to the budget and conduct elections in the County.

28. Los Angeles County had a budget of $9,111,147,132 for the fiscal year ending June 30, 1989.

2. Demographics of Los Angeles County

29. The demographic picture of Los Angeles County has changed dramatically since 1950.

30. The 1950 Census of Population, (hereinafter "Census"), reported that the total population of Los Angeles County was 4,015,687, of whom 287,614 (7.2%) were persons with Spanish surnames.

31. The 1960 Census reported that the total population of Los Angeles County was 6,038,771, of whom 576,716 (9.6%) were persons with Spanish surnames.

32. The total population of Los Angeles County increased by 2,023,084 persons (50.4%) between 1950 and 1960 while the County's Spanish-surnamed population increased by 289,102 persons (100.5%) between 1950 and 1960.

33. The 1960 Census data revealed a population concentration of Spanish surnamed persons in the area south and east of downtown Los Angeles.

34. The 1970 Census reported that the total population of Los Angeles County was 7,032,075 persons, of whom 1,289,311 (18.3%) were persons with Spanish surnames.

35. The total population of Los Angeles County increased by 993,304 persons (16.4%) between 1960 and 1970, while the County's Spanish-surnamed population increased by 712,595 persons (60.24%).

36. The 1970 Census revealed several discrete concentrations of Spanish surnamed persons in the center of the County.

37. The 1980 Census reported that the total population of the County of Los Angeles was 7,477,503 persons, of whom 2,066,103 (27.6%) were persons of Spanish origin, 926,361 (12.4%) were black persons and 434,850 (5.8%) were Asians and Pacific Islanders.

38. The total population of Los Angeles County increased by 445,428 persons (6.3%) between 1970 and 1980 while the number of persons of Spanish origin in the County increased by 776,792 persons (60.2%).

39. According to the 1980 Census, the population of Los Angeles County can be summarized as follows:

Los Angeles County — 1980 Census

Total Hispanic White Black Asian Others ----- -------- ----- ----- ----- ------

40. The data from the previous three decennial censuses demonstrate that as Los Angeles County's total population has grown over the past few decades, the County's racial and ethnic composition has also changed. The group that has grown the fastest in recent years is comprised of persons of Spanish origin, as reported by the Census in 1980.

41. The number of persons reported as having Spanish surnames in the 1970 Census of Los Angeles County was 1,051,409.

42. By the time of the 1980 Census, more than 2 million people in Los Angeles County reported that they were of Spanish origin.

43. The County's Hispanic population is concentrated, to a significant extent, in a compact and contiguous area beginning in the eastern part of the City of Los Angeles and extending eastward into the San Gabriel Valley. (hereinafter "Hispanic Core")

44. This Hispanic Core includes Boyle Heights, Lincoln Heights and El Sereno in the City of Los Angeles, the unincorporated East Los Angeles community, and the cities of Rosemead, Pico Rivera, Montebello, La Puente, El Monte, Maywood, Vernon, Bell, Bell Gardens and other cities and unincorporated communities.

45. The Hispanic Core is contained within a set of 229 census tracts. These tracts are contiguous and persons of Spanish origin were the majority of the population in all but three of the tracts according to the full-count data from the 1980 Census.

46. According to the 1980 Census, the Hispanic Core had a total population of 1,204,279, of whom 877,478 (72.8%) were Hispanic and a voting age population that was 67.4 percent Hispanic.

47. Approximately 40 percent of the County's entire Hispanic population lived in one of the 229 core census tracts in 1980, and these tracts comprise 81 percent of all census tracts with Hispanic population majorities in 1980.

48. Data from the Los Angeles County Department of Health Services and Data Processing have been presented to the Court in the form of a series of small-area population estimates and projections known as the Population Estimates and Projections System (hereinafter "PEPS").

49. PEPS data contains estimates of 1985 and 1987 total population and population by race and ethnicity by various age levels for each populated census tract in Los Angeles County.

50. PEPS also generated projections of 1989 and 1990 total population and population by race and ethnicity by various age levels for each populated census tract in the County.

51. The County's population as a whole grew by 12.3 percent between 1980 and 1987. The County's Hispanic population grew by 42.7 percent between 1980 and 1987. By 1990, Hispanics are expected to constitute 35.8 percent of the total population of the County.

52. According to PEPS data, the number of non-Hispanic whites fell by 378,000 between 1980 and 1987. In 1980, non-Hispanic whites made up 53.2 percent of the County's total population. By 1987, non-Hispanics made up only 42.8 percent of the County's total population.

53. In 1990, non-Hispanic whites are projected to constitute 39.8 percent of the population.

54. In the Hispanic Core, the total population was estimated by PEPS to have grown from 1,204,279 persons to 1,519,630 persons between 1980 and 1987.

3. Campaign Financing

55. Since 1960, only three incumbents running for a seat on the Board of Supervisors were defeated in their reelection bids. Supervisor Hahn has served since 1952; Supervisor Schabarum since 1972; Supervisor Edelman since 1974; and Supervisors Dana and Antonovich since 1980.

56. Incumbent Supervisors enjoy a strong campaign fund-raising advantage over their challengers for reelection.

57. In 1987, the combined campaign funds of the five incumbent Supervisors totaled $3 million. Forty-nine percent of this amount belonged to Supervisors Schabarum and Edelman, who would not run for reelection until 1990. The largest amount, $800,000 belonged to Supervisor Dana, the smallest, $210,000, to Supervisor Hahn.

58. Incumbent Supervisors received 91 percent, $8.2 million of $9.1 million, of all campaign money raised from 1981 to 1986 and raised 74 percent of their contribution in non-election years.

59. During the 1982, 1984, 1986, and 1988 elections, each incumbent Supervisor had more campaign funds expended on his behalf than were expended on behalf of his combined opposition.

60. Potential candidates recognize that to be considered a serious candidate for the Board, a person must spend between one and two million dollars on a campaign.

61. Mr. George Pla, who has managed political campaigns for elections in Los Angeles County, testified that it would be difficult for any candidate to raise $1-2 million, but that it would be even more difficult for an Hispanic candidate because of lack of a financial base. Pla also noted the adverse effect the inability to raise funds had on public perception of an Hispanic candidate's likelihood of success.

4. Prior Redistrictings

62. The 1981 Plan cannot be analyzed in a vacuum. As illustrated by the testimony of J. Morgan Kousser, a professor of History at the California Institute of Technology, if the Court examines the changes in District 3 in the context of the demographic changes in the County as a whole, as well as the place where Hispanics lived and moved to during that period of time, the pattern is persuasive evidence that the lines were drawn and maintained with a racially discriminatory design.

63. Dr. Kousser, in particular, concluded that there was ample evidence to be gleaned from the history of prior redistrictings to indicate that the Board kept the Hispanic Core split in order to secure their positions against challengers who would appeal to Hispanic voters.

(a) The 1959 Redistricting

64. Prior to 1959, District 3 included Western Rosemead and did not include any portion of the San Fernando Valley, Beverly Hills, West Hollywood, West Los Angeles, or Eagle Rock.

65. The 1959 redistricting occurred less than six months after the November 1958 general election for the open position of District 3 Supervisor. Ernest Debs, a non-Hispanic, defeated Hispanic candidate Edward Roybal, by a margin of 52.2 percent to 47.8 percent.

66. Debs received 141,011 votes. Roybal received 128,974 votes. There were four recounts before Debs was finally determined to be the winner.

67. In 1959, Debs reported in a Supervisorial hearing that he and District 4 Supervisor Burton Chace agreed to shift Beverly Hills, West Hollywood, and West Los Angeles from District 4 to District 3.

68. The Board's action transferred between 50,000 to 100,000 voters from District 4 into District 3 and had the effect of substantially decreasing the proportion of Hispanic voters in District 3.

69. Dr. Kousser testified it was his opinion that Debs and Chace agreed to the transfer for two reasons. First, Chace was receptive to the agreement because it enabled him to eliminate Los Angeles City Councilwoman Rosalind Wyman as a possible opponent in his upcoming 1960 bid for reelection. Debs welcomed the change because the move west allowed him to make District 3 more easily winnable against Roybal or another candidate who might appeal to Hispanic voters in the next election.

70. Debs was a Democrat and Chace a Republican. The two were not allies on other issues.

71. At the time of this transfer, District 1, which borders on the east of District 3, was much larger than the other four districts.

72. If Debs had taken communities from District 1, the five districts would have been equipopulous. The lack of effective equal population requirements at the time made it possible for the District 3 to be moved deliberately west instead of east which avoided adding communities from the Roybal stronghold in East Los Angeles.

(b) The 1963 Redistricting

73. On December 19, 1961, the Board of Supervisors, acting in accordance with Section 25009 of the California Government Code enacted in 1961, adopted an order establishing the Supervisorial District Boundary Committee, (hereinafter "Boundary Committee"), to study and make recommendations concerning the need for changing Supervisorial district boundaries in Los Angeles County. Each Supervisor appointed one member to the committee.

74. In 1962, voters defeated a referendum to expand the Board of Supervisors from five to seven members.

75. Evidence suggests Debs wanted the referendum issue on the ballot in 1962 because he sought to move his district out of East Los Angeles and concentrate his district in the western area of the district, Beverly Hills, West Los Angeles and West Hollywood, communities with larger proportions of Non-Hispanic whites.

76. The Board of Supervisors adopted ordinance 8407 on May 14, 1963 which enacted the recommendations of the Boundary Committee and established new district boundaries.

77. The boundary changes involved a shift in the boundary between Districts 3 and 5, in which District 3 was extended north across the Santa Monica Mountains, for the first time, to the Ventura Freeway and into the San Fernando Valley. Eagle Rock was also added to District 3.

78. At the time of the 1963 boundary changes, a growing Hispanic population was beginning to emerge in the San Gabriel Valley, directly adjacent to the eastern boundary of District 3. Eagle Rock, in contrast, was about 4 percent Spanish surname and the portion of the San Fernando Valley annexed to District 2 was about 1 percent Spanish surname.

79. Since District 3 was underpopulated in 1963 and District 1 was overpopulated, population equality among the supervisorial districts could have been fostered by moving District 1's growing Hispanic areas in the San Gabriel Valley directly to District 3. This was not done.

(c) The 1965 Redistricting

80. In 1965, the California Supreme Court ruled that no Supervisorial district in California should have more than 23 percent or less than 17 percent of a County's total population. Miller v. Board of Supervisors of Santa Clara County, 63 Cal.2d 343, 46 Cal.Rptr. 617, 405 P.2d 857 (1965).

81. In response to Miller, the Los Angeles County Board of Supervisors reactivated the Supervisorial District Boundary Committee on October 5, 1965.

82. The 1965 Boundary Committee considered a proposal by Russell Quisenberry, the appointee of District 5 Supervisor Warren Dorn, to move 90,000 people in Alhambra and San Gabriel, areas close to the Hispanic Core, from District 1 to Debs' District 3. Dorn proposed that these changes be implemented after the 1966 election, when Debs faced reelection.

83. The Boundary Committee did not follow the Dorn proposal. Instead, Alhambra and San Gabriel were assigned to Dorn's Fifth District and 87,000 predominantly Anglo residents of San Fernando Valley were moved to Debs' District 3 from Dorn's District 5.

84. The Boundary Committee reported that, based on estimates of population, the Supervisorial Districts if revised according to the committee's recommendations, would have the following populations:

District 1 1,492,000 District 2 1,258,000 District 3 1,398,000 District 4 1,253,000 District 5 1,484,000

85. The Boundary Committee plan provided for an average deviation from population equality of 7.06 percent and a maximum deviation of 17.35 percent.

86. On November 30, 1965, the Board of Supervisors, by a 4-1 vote with Supervisor Hahn dissenting, adopted Ordinance 8998, which enacted the plan proposed by the 1965 Boundary Committee.

87. The adoption of the 1965 plan involved such changes as: (1) The inclusion of the City of Long Beach, which previously was split between two districts, wholly in District 4; (2) The boundary between District 3 and District 5 was shifted from the Ventura Freeway to Oxnard Boulevard; (3) Monterey Park and unincorporated South San Gabriel were transferred from District 1 to District 3; and (4) District 5, as previously discussed, was allocated a portion of the eastern part of the County in the San Gabriel Valley which previously had been represented by Supervisor Bonelli from District 1.

88. The Boundary Committee rejected a proposal to move Alhambra and San Gabriel, areas adjacent to growing Hispanic population, from District 1 to District 3. Instead, the committee recommended a complicated two-stage change which moved Alhambra and San Gabriel from Supervisor Bonelli's District 1 to Supervisor Dorn's District 5, moved a section of the San Fernando Valley from District 5 to Supervisor Debs' District 3, and moved Monterey Park and unincorporated South San Gabriel from District 1 to District 3.

89. Dr. Kousser testified that, in his opinion, the Board avoided transferring Alhambra and San Gabriel directly to District 3 because those areas were adjacent to areas of Hispanic population concentration and were becoming more Hispanic. The more complicated two-stage adjustments permitted the addition of heavily Anglo areas from the San Fernando Valley and offset the much more limited addition of Hispanic population gained by moving Monterey Park and the unincorporated area of South San Gabriel to District 3.

90. None of the persons who served on the 1965 Boundary Committee were individuals with Spanish surnames.

(d) The 1971 Redistricting

91. A comparison of the 1960 Census data with 1970 Census data demonstrates the extent to which areas bordering on District 3 were gaining Hispanic population. Spanish surname population increased during that decade in Alhambra from 6 percent to 19 percent and in Monterey Park from 13 percent to 33 percent.

92. The Hispanic population in the County doubled from 1970 to 1980 and, in 1970, the Hispanic Core showed marked and continuous expansion outward and contiguously into the San Gabriel Valley.

93. Efforts were made during this time to expand the Board of Supervisors. Esteban Torres, who was president of the Congress of Mexican American Unity, testified before the Los Angeles County Economy and Efficiency Committee in April 1970, to urge that the Committee recommend expansion of the Board.

94. Concurrent efforts were also made to expand the Los Angeles City Council.

95. The Board failed to obtain the three votes necessary to place the issue on the ballot. The City Council expansion effort failed to pass at the polls.

96. The Board of Supervisors established the Los Angeles County Supervisorial District Boundary Committee on April 20, 1970. (hereinafter "Boundary Committee")

97. The members of the committee and the Supervisors who appointed them were as follows: John D. Lusk by Supervisor Bonelli; Dan Patacchia by Supervisor Hahn; Leslie G. Cramer by Supervisor Debs; LeRoy Center by Supervisor Chace; and Alfred E. Paonessa by Supervisor Dorn.

98. None of the individuals who served on the 1963, 1965 and 1971 Boundary Committees had a Spanish surname.

99. Richard Schoeni, a County employee, served as the secretary to the 1971 Boundary Committee. In this capacity, Mr. Schoeni provided staff support, gathered information, made suggestions, maintained the committee's records, and drafted the report and recommendations that the Committee submitted to the Board of Supervisors.

100. Pursuant to Section 25001 of the California Government Code, the Board, in redistricting, may consider such factors as: topography, geography, cohesiveness, contiguity, integrity, compactness of territory, and community of interests of the districts.

101. The Boundary Committee adopted the following guidelines in addition to the factors delineated in the California Government Code: (1) to preserve historical representation of certain areas closely identified with a particular district; (2) to avoid the division of cities by supervisorial boundaries whenever possible; and (3) to avoid the separation of cities or communities sharing common interests and problems peculiar to a section of the County.

102. Population statistics generated from the 1970 Census demonstrated that the 1965 supervisorial districts had the following populations:

Percentage District Population of Total -------- ---------- --------

1 1,547,407 22.0 2 1,238,454 17.6 3 1,364,312 19.4 4 1,271,186 18.1 5 1,610,716 22.9

103. Among the proposals discussed during the meetings of the Boundary Committee was one presented by Leslie Cramer, representative of Ernest Debs, to extend District 3 further into the San Fernando Valley north of Oxnard Boulevard.

104. The 1971 Boundary Committee never gave any consideration to moving District 3 east to include more of the San Gabriel Valley or moving Pico Rivera from District 1, which was overpopulated, to District 3. Nor did the committee consider adding such areas as San Gabriel, Rosemead or El Monte to District 3.

105. According to the testimony of Dr. Schoeni, moving District 3 east was not considered to avoid splitting the San Gabriel Valley. However, San Gabriel Valley was already split among District 5 which contains Alhambra, San Gabriel, and East San Gabriel; and District 3 which contains South San Gabriel and part of Rosemead.

106. The Los Angeles County Supervisorial District Boundary Committee Report and Recommendations, which included a detailed description of the supervisorial boundaries prepared by the County Engineer, was submitted to the Board of Supervisors on July 22, 1971. The Board adopted the plan proposed by the Boundary Committee.

107. The Boundary Committee recommended the following changes to the existing plan: (1) Artesia, Bellflower, Cerritos and Lakewood were transferred from District 1 to District 4; (2) Rosemead was transferred from District 3 to District 1; and (3) Van Nuys, Sepulveda, Panorama City and Sun Valley were transferred to District 3 from District 5.

108. As a result of the 1971 redistricting, District 3 gained over 205,000 people from other districts and lost more than 163,000 people to other districts.

109. In 1971, District 3 lost some areas with substantial Hispanic population on its eastern border. Western Rosemead was transferred from District 3 to District 1. A census tract in the City of San Gabriel was also transferred from District 3 to District 5.

110. George Marr, head of the Population Research Section of the Department of Regional Planning testified that he was surprised by the proposal to move a substantial portion of the San Fernando Valley from District 5 to District 3. Marr described the portion of the San Fernando Valley ultimately added to District 3 from District 5 as looking like "one of those Easter Island heads." Marr developed the general feeling that Debs' representative on the Boundary Committee had requested the additional area in the San Fernando Valley because the residents of the area were regarded as "our kind of people."

111. None of the persons who served on the 1971 Boundary Commission were individuals with Spanish surnames.

(1) Intent of Past Redistrictings

112. The Court finds that the Board has redrawn the supervisorial boundaries over the period 1959-1971, at least in part, to avoid enhancing Hispanic voting strength in District 3, the district that has historically had the highest proportion of Hispanics and to make it less likely that a viable, well financed Hispanic opponent would seek office in that district. This finding is based on both direct and circumstantial evidence, including the finding that, since the defeat of Edward Roybal in 1959, no well-financed Hispanic or Spanish-surname candidate has run for election in District 3.

113. While Hispanic population was added to District 3 during the 1959-1971 redistrictings, the Court finds that the proportion of Spanish-surname persons added to District 3 has been lower than the Hispanic population proportion in the County as a whole. No individual area added was greater than 15.1 percent Spanish-surname.

114. Dating from the adoption of the County's Charter in 1912 through the 1971 redistricting process, no Los Angeles County redistricting plan has created a supervisorial district in which Hispanic persons constituted a majority or a plurality of the total population.

(e) 1972 Los Angeles City Council Redistricting

115. In 1971, the California Supreme Court ruled that the 1968 voter-based reapportionment plan for the Los Angeles City Council was unconstitutional. Calderon v. City of Los Angeles, 4 Cal.3d 251, 266, 93 Cal.Rptr. 361, 481 P.2d 489 (1971).

116. In 1972, the Charter and Code Committee set out to devise a new redistricting plan.

117. As then Committee Chair Edmund Edelman stated in a 1971 press release: "It is my intention to urge my colleagues on the committee and on the council to create a district where it would be possible for a Mexican-American to be elected."

118. Edelman proposed a plan which increased the Spanish surname proportion in District 14, held by Councilman Art Snyder, from approximately 40 percent to 68 percent Hispanic by unifying Hispanic communities previously split by Districts 13 and 14.

119. Citizenship and voting age data were not used for purposes of devising the city's 1972 redistricting plan.

120. In devising the 1972 plan, Edelman was assisted by Alma Fitch and Jeff Seymour, both of whom played a role in the 1981 supervisorial redistricting.

121. Chicanos for Fair Representation criticized Edelman's plan and questioned the accuracy of the 68 percent estimate of Spanish surname population, believing it to be 57 percent.

122. The City Council adopted the Edelman Plan and overrode the veto of Mayor Sam Yorty.

123. Despite the substantial increase in the Hispanic population in District 14, Councilman Snyder was able to defeat several Hispanic opponents.

124. In 1985, after Snyder's retirement, Richard Alatorre was elected to represent District 14 and became the first Hispanic to serve on the Los Angeles City Council since Edward Roybal.

(f) The 1981 Redistricting

125. The individuals involved in the 1981 redistricting had demographic information available of population changes and trends in Los Angeles County from 1950 to 1980. It was readily apparent in 1980 that the Hispanic population was on the rise and growing rapidly and that the white non-Hispanic population was declining.

126. According to the report of the 1981 Boundary Advisory Committee, the 1980 Census data showed that the districts under the 1971 boundaries had the following population characteristics:

District Population % Black % Hispanic % -------- ---------------- -------------- --------------

1 1,522,347 20.4 47,772 3.1 550,819 36.2 2 1,423,015 19.0 635,751 44.7 354,314 24.9 3 1,577,877 21.1 44,868 2.8 669,246 42.4 4 1,445,286 19.3 140,585 9.7 236,518 16.4 5 1,509,132 20.2 75,003 5.0 254,830 16.9

127. From a political perspective, since Hispanic population growth was most significant in Districts 1 and 3, if the 1971 boundaries were changed in any measurable way to eliminate the existing fragmentation, the incumbency of either Supervisor Schabarum or Supervisor Edelman would be most affected by a potential Hispanic candidate.

128. All of the plans considered by the participants in the redistricting were based on 1980 Census population data.

129. In 1981, citizenship or voting age data was not considered or requested by County staff, Boundary Committee members or the Supervisors and their aides.

130. No suggestion was made in 1981 that citizenship data or voter registration data be used as the apportionment base.

131. On February 27, 1981, Deane Dana sent a lettergram to Supervisors Schabarum and Antonovich recommending that both a public and a private redistricting committee be established. Dana suggested obtaining the services of Joseph Shumate to assist in the redistricting effort noting the experience Shumate had with the Republican reapportionment efforts in 1970, 1971 and 1972.

132. Mr. Shumate was hired to work in a private capacity on behalf of Supervisors Dana, Schabarum and Antonovich. The objective, according to Mr. Shumate's testimony, was "to assist in determining whether a plan would help or hurt the three Supervisors."

133. Allan Hoffenblum, a political advisor to Supervisor Antonovich, testified that the following statement attributed to him was what he believed at the time:

"We would be remiss if we did not have at least one district that was at least 50 percent Hispanic, otherwise it looks like we're sitting here trying to save five white Supervisors."

134. Supervisor Edelman and others involved in the 1981 redistricting effort were not aware of Mr. Shumate or the role he played on behalf of Supervisors Schabarum, Antonovich and Dana.

135. While these three Supervisors were pursuing their redistricting efforts, Supervisor Edelman asked Jeffrey Seymour to assist him in the redistricting process by examining maps produced and by preparing a political analysis of Supervisor Edelman's district.

136. An analysis of the 1978 Supervisor election in District 3 was conducted after the Boundary Committee recommended a plan with an Hispanic population majority in District 3. The actual results of the analysis were never produced. Mr. Seymour did not rule out the possibility that he requested such an analysis and Supervisor Edelman testified that he "most probably" discussed the results of the 1978 election analysis with Mr. Seymour.

137. Peter Bonardi, a programmer with the Urban Research Section of the Data Processing Department in 1981 and a participant in the data analysis requested by Supervisor Edelman, stated that he was directed not to talk about the analysis of voting patterns and that an "atmosphere of `keep it quiet'" pervaded.

138. Supervisors Hahn and Edelman sought to maintain the existing lines. To this end, the Democratic minority agreed to a transfer of population from District 3 to District 2. Supervisor Edelman acknowledged that he and Supervisor Hahn had worked out a transfer of population from the heavily Hispanic Pico-Union area on the southern border of District 3 to the northern end of District 2.

139. Supervisor Edelman knew that if the 1971 boundary lines were kept intact, the Hispanic community was going to remain essentially the same in terms of its division among the districts.

140. The Board departed from its past redistricting practice in 1981 and approved a contract with The Rose Institute for State and Local Government, a private entity, to perform specialized services and produce redistricting data at a cost of $30,000.

141. The facilities at The Rose Institute were used primarily by persons working privately on behalf of Supervisor Dana, Schabarum, and Antonovich, including Joseph Shumate, conducting private redistricting research and analysis.

142. The Board reactivated the Supervisorial District Boundary Committee and charged the Committee with the responsibility for recommending a redistricting plan in accordance with the provisions of Sections 35000-35006 of the Elections Code and one which would ensure that ethnic minorities are equitably represented, and that city boundaries were respected as much as possible.

143. The five initial appointees to the Boundary Committee, Blake Sanborn, Robert Bush, Ron Smith, Alma Fitch, and Allan Hoffenblum, were persons who had close political ties to the appointing Supervisors or were persons who had been trusted employees and advisors to the Supervisors. All five individuals were non-Hispanic.

144. On July 8, 1981, representatives of Californios for Fair Representation, (hereinafter "Californios" or "CFR"), a coalition of Hispanic organizations active in the redistricting process, criticized the all-Anglo composition of the Committee and requested that it be expanded to include minority representatives, including at least one Hispanic and one black.

145. On July 14, 1981, the Board of Supervisors appointed five additional members to the Boundary Committee, Lauro Neri and Jesus Melendez, both Hispanic; Davis Lear and Robert Perkins, both black; and Dr. Frederico Quevedo, a Filipino. These additional appointees played a minor role in the redistricting process. None of the minority representatives or persons appointed to the Boundary Committee on July 14, 1981 had any previous experience in demography or the redistricting of elective bodies.

146. The Los Angeles County Coalition of CFR worked on redistricting plans for state, county and local jurisdictions within Los Angeles County and was permitted to use the facilities of the Rose Institute for the purpose of preparing their proposals.

147. Leticia Quezada was the chair of the Los Angeles County CFR chapter.

148. In considering different redistricting strategies, CFR declined to create a plan which included one district with a substantial Hispanic majority because they did not think that four Supervisors would vote for such a plan. CFR viewed a plan which included an Hispanic district as very threatening to incumbents since it would involve drastic shifts in population.

149. CFR instead opted to propose a plan that reduced the splintering of the Hispanic community and provided for two Hispanic "influence" districts: one with a bare Hispanic population majority in District 3 and an Hispanic growth district in District 1 with 42 percent Hispanic population.

150. Through various conversations with the Supervisors or their representatives, members of CFR developed an understanding of the objectives of the Supervisors for the redistricting process. Supervisor Edelman indicated that he wanted the San Diego Freeway to be the western boundary of District 3 and the Santa Monica Freeway to be the district's southern boundary. Alma Fitch informed CFR that Supervisor Edelman was happy to represent the Hispanic community but that he did not believe that all the Hispanics should be in his district. Mike Lewis indicated in meetings with CFR that Supervisor Schabarum was willing to transfer Hispanic population from District 1 to District 3 to create an Hispanic district. Specifically, the Supervisor was willing to lose Pico Rivera and South Gate, two majority Hispanic cities.

151. Boundary Committee members Hoffenblum and Smith each introduced plans with identical 50.2 percent Hispanic populations in Supervisorial District 3. The plans differed with respect to District 1. While the Californios plan increased the District 1 Hispanic population from 32 percent Hispanic to 42 percent, both the Hoffenblum and Smith plans reduce the Hispanic population in District 1 to 31.3 percent and 31.7 percent Hispanic respectively.

152. By a 5-4 vote, the Committee recommended that the Board adopt the Hoffenblum Plan. The Final Report of the Boundary Committee stated as to the Hoffenblum Plan:

This Plan increases the opportunity of Hispanics and Blacks by recognizing that a special community of interests exists for Hispanics and Blacks. Boundaries were developed to increase the electoral effectiveness of these two groups in the Second and Third Supervisorial districts.

153. Representatives of Supervisors Hahn and Edelman offered no proposals for plans during the course of the Committee's meetings, and opposed all plans enlarging the Hispanic population in District 3 beyond the then-current district lines.

154. Besides increasing the Hispanic population in District 3, the Hoffenblum Plan reduced the Hispanic population in the districts of Supervisors Dana, Schabarum and Antonovich, particularly in Schabarum's Supervisorial District 1, and the black populations in the districts of Dana and Antonovich. Black and Hispanic populations were added to Districts 3 and 2. The following tables show the changes in population statistics as illustrated by the Hoffenblum Plan from the then-existing boundaries.

Current (1981) Boundaries — Pre-Redistricting

Dist. Population % Black % Hispanic % ----- ----------------- -------------- --------------

1 1,522,347 20.4 47,772 3.1 550,819 36.2 2 1,423,015 19.0 635,751 44.7 354,314 24.9 3 1,577,877 21.1 44,868 2.8 669,246 42.4 4 1,445,286 19.3 140,585 9.7 236,518 16.4 5 1,509,132 20.2 75,033 5.0 254,830 16.9

Hoffenblum Plan

Dist. Population % Black % Hispanic % ----- ---------------- --------------- --------------

1 1,496,560 20.0 48,708 3.3 468,661 31.3 2 1,495,727 20.0 691,655 46.2 384,721 25.7 3 1,495,085 20.0 50,863 3.4 750,266 50.2 4 1,495,738 20.0 81,082 5.4 231,268 15.5 5 1,495,547 20.0 71,701 4.8 230,811 15.4

155. The Smith and Hoffenblum plans, while increasing District 3 to a bare majority, proposed a substantial decrease in the Hispanic population percentage in District 1.

156. The CFR plan, the Smith plan and the Hoffenblum plan all proposed shifting the City of Compton from District 4, the "coastal district," to District 2, where most of the County's black population was concentrated.

157. Smith and Hoffenblum opposed the CFR plan because the plan proposed increasing the Hispanic proportion in District 1 from 36 to 42 percent. Both Boundary Committee members perceived the CFR effort as intended to jeopardize the status of Supervisor Schabarum as well as that of the conservative majority.

158. Hoffenblum testified that one of the objectives of the Republican majority was to create an Hispanic seat without altering the ideological makeup of the Board. According to Hoffenblum, it was "self-evident" that if an Hispanic district was created in Supervisor Schabarum's district it would impact on the Republican majority.

159. The proponents of the Smith and Hoffenblum plans sought to gain areas of Republican strength such as La Mirada, Arcadia, Bradbury in Districts 4 and 5, while losing increasing Hispanic areas such as Alhambra or the predominantly black Compton and other liberal areas of Santa Monica and Venice.

160. The Boundary Committee met officially on eight occasions between July 8 and August 12, 1981.

161. No Board member ever publicly advocated any of the plans introduced by members of the Boundary Commission, including the recommended Hoffenblum Plan or the CFR Plan.

162. Supervisor Edelman would not rule out the possibility that ethnic considerations played at least some part in the rejection by the Board majority of the CFR Plan. Moreover, the fact that CFR proposed a plan in which District 1 had a 42 percent Hispanic population was a possible basis for the rejection of the plan by the majority. Supervisor Schabarum would not accept a 45 or 50 percent Hispanic proportion in his district in 1981.

163. The Supervisors proposed no plans raising Hispanic population in any district beyond where it already remained by virtue of the 1971 boundary lines. Although the feasibility of establishing even a 50 percent Hispanic district was never disputed, no Supervisor ever publicly discussed or endorsed the idea.

164. The Court finds that in 1981 a district could have been devised which more fairly and adequately recognized Hispanic voting strength while complying with standard redistricting criteria.

165. On September 24, 1981, prior to the Board's adoption of the challenged plan, Board members met, two at a time in a series of private meetings in the anteroom adjacent to the board room, where they tried to reach agreement on a plan.

166. On at least ten separate occasions, pairs of Supervisors entered the room and negotiated the final redistricting plan.

167. The Board rejected the Boundary Committee's report.

168. According to the deposition of Mr. Schoeni:

"The Boundary Committee report was received; the Board beard testimony; the Board set aside the Boundary Committee report and proceeded from a clean slate, if you will, with Supervisor Edelman mediating and trying to gain as much in terms of population equity as was possible."

169. Using the map which was in the anteroom, Supervisor Schabarum and Antonovich discussed changes in the boundary between Districts 1 and 5, including the transfer of Sierra Madre. Schabarum and Antonovich did not discuss the Hoffenblum, Smith or CFR plans.

170. Mr. Schoeni drew proposed district boundaries on the map in the anteroom, and once a map was developed which purported to reflect a consensus of the Board, an immediate tabulation of the changes was performed and given to Board Chair, Supervisor Edelman.

171. On September 24, 1981, after Board members had reached an agreement on a plan, the Board met publicly and unanimously adopted this recent creation which had never been presented to the public.

172. Supervisor Schabarum testified that he had described the 1981 plan as "ho-hum" because it "just juggled the boundaries around a little bit to get the job done within the law."

173. Supervisor Schabarum also testified that he thinks it "fundamentally un-American and unsound" to fashion district lines with the intent of permitting ethnic groups to be represented.

174. The Court finds that the Supervisors and their aides understood the potential for increasing Hispanic voting strength and sought to avoid the consequences of a redistricting plan designed to eliminate the fragmentation of the Hispanic population.

(1) Intent of 1981 Redistricting

175. The plan adopted in 1981 retained the boundary between the First and the Third Supervisorial Districts, the districts that contain the largest proportions of Hispanics. In doing so, the 1981 Plan continued to split the Hispanic Core almost in half.

176. The Board appeared to ignore the three proposed plans which provided for a bare Hispanic population majority.

177. The Court finds that the Board of Supervisors, in adopting the 1981 redistricting plan, acted primarily with the objective of protecting and preserving the incumbencies of the five Supervisors or their political allies.

178. The Court finds that in 1981 the five members of the Board of Supervisors were aware that the plan which they eventually adopted would continue to fragment the Hispanic population and further impair the ability of Hispanics to gain representation on the Board.

179. The continued fragmentation of the Hispanic vote was a reasonably foreseeable consequence of the adoption of the 1981 Plan.

180. The Court finds that during the 1981 redistricting process, the Supervisors knew that the protection of their five Anglo incumbencies was inextricably linked to the continued fragmentation of the Hispanic Core.

181. The Supervisors appear to have acted primarily on the political instinct of self-preservation. The Court finds, however, that the Supervisors also intended what they knew to be the likely result of their actions and a prerequisite to self-preservation — the continued fragmentation of the Hispanic Core and the dilution of Hispanic voting strength.

D. SIZE AND GEOGRAPHIC COMPACTNESS OF HISPANIC COMMUNITY

1. 1980 Census Data

182. As stated supra, in Part C, 2, the 1980 Census reported that the total population of the County of Los Angeles was 7,477,503 persons of whom 2,066,103 or 27.6 percent were persons of Spanish origin.

183. According to full-count data from the 1980 Census, persons of Spanish origin were the majority of the population in all but three of the 229 contiguous census tracts comprising the Hispanic Core.

184. Dr. William P. O'Hare, a sociologist and demographer who is director of policy studies for the Population Reference Bureau, a non-profit research and educational organization in Washington, D.C., compiled a demographic profile of the 229 Hispanic Core census tracts using full-count tract-level reports of total population and voting age, plus a special tabulation of voting age citizens provided by the Census Bureau. These core census tracts had the following aggregate characteristics:

Hispanic Population Core — 1980

Total Hispanic White Black Asian ----- -------- ----- ----- -----

185. The Court finds that based on 1980 Census data, a supervisorial district can be drawn encompassing the Hispanic Core community so that the percentage of citizen voting age Hispanics in the districts would be such that Hispanics would have the potential to elect a candidate of their choice. While the Court agrees with defendants that plaintiffs' experts, Dr. Grofman and Dr. Estrada, could not devise a plan with a voting age citizen majority on the basis of 1980 Census data that comports with the one man one vote rule, the Court does find it persuasive that the illustrative districts were just shy of the 50 percent mark, in the 44 to 46 percent range.

186. Further, this case presents precisely the situation anticipated by Justice O'Connor, in her concurring opinion in Gingles, 478 U.S. at 94-95, 106 S.Ct. at 2789, in which the unique demographic changes Los Angeles County has undergone and continues to undergo coupled with the lingering effects and history of discrimination in the County against Hispanics, preclude the application of "a single, universally applicable standard for measuring undiluted minority voting strength." Id. The application of the bright line 50 percent requirement set forth by the Ninth Circuit in Romero v. City of Pomona, 883 F.2d 1418, 1427 (9th Cir. 1989), would be inappropriate under these facts and circumstances.

187. Therefore, even if this Court were to agree with defendants' contention that current population data is less reliable than the 1980 Census, the Court would still find that the 1981 Plan violated Section 2 of the Act based on the totality of the circumstances delineated in these findings.

188. The evidence shows that the Board of Supervisors knowingly chose to draw and adopt a plan that minimized the voting potential of the County's Hispanic population. This minimization of Hispanic voting strength was achieved by fragmenting the Hispanic Core.

189. The distribution of tracts among current supervisorial districts reflects this conscious minimization:

Districts Tracts TPOP HPOP % VAP HVAP % --------- ------ --------- ------------- ------- --------------

1 77 433,173 299,648 69% 277,169 174,664 63% 2 50 226,318 155,332 69% 153,645 97,516 63% 3 104 538,093 418,750 78% 348,257 253,564 73% 5 2 6,694 3,749 56% 4,606 2,278 49% --- --------- ------------- ------- -------------- Total 233 1,204,279 877,478 73% 783,677 528,021 67.4%

Districts Tracts TCVAP HCVAP %

--------- ------ ------- -------------

1 77 190,705 95,950 50% 2 50 68,954 22,925 33% 3 104 195,445 117,077 60% 5 2 3,202 1,374 43% --- ------- ------- --- Total 233 458,306 237,326 52%

190. No citizenship data by the Census Bureau with respect to the 1980 decennial census was available in time for the 1981 redistricting process.

2. Growth in Hispanic Population Since 1980

191. The demographics of Los Angeles County have changed dramatically since 1980.

192. The population data for Los Angeles County can be summarized as follows:

1980 Census

Total Hispanic White Black Asian/Other --------- -------- ----- ----- -----------

1985 PEPS Population Estimates

County 8,018,210 30.4% 47.2% 12.0% 10.3% Core 1,359,907 74.4% 13.4% 3.5% 8.8%

1987 PEPS Population Estimates

County 8,418,817 34.4% 42.8% 11.5% 11.3% Core 1,519,630 77.2% 11.0% 3.0% 8.7%

1989 PEPS Population Projections

County 8,718,710 35.8% 40.8% 11.2% 12.2% Core 1,602,484 78.2% 9.7% 2.9% 9.3%

1990 PEPS Population Projections

County 8,880,109 36.6% 39.8% 11.0% 12.6% Core 1,648,827 78.7% 9.0% 2.9% 9.5%

193. According to PEPS data, the number of Hispanic persons in Los Angeles County increased by more than 850,000 between 1980 and 1987 and that the number of Hispanics in Los Angeles County is projected to increase by almost 350,000 more between 1987 and 1990.

194. The population of Los Angeles County grew by 12.3 percent between 1980 and 1987. The Hispanic population grew by 42.7 percent during this period.

195. By 1990, Hispanics are expected to constitute 35.8 percent of the County's total population.

196. Between 1980 and 1987, the number of non-Hispanic whites decreased by 378,000. In 1980, whites comprised 53.2 percent of the total population. In 1987, whites made up 42.8 percent of the population.

197. According to PEPS projections, non-Hispanic whites will constitute 39.8 percent of the County's population by 1990.

198. The Hispanic population growth in Los Angles County during the 1980's has occurred primarily in areas where there was already a significant Hispanic presence in 1980. Almost two-thirds of the Hispanic population growth between 1980 and 1987 has occurred in census tracts where Hispanics made up more than 25 percent of total population in 1980.

199. As the Court stated in is findings on the 1981 Redistricting supra, the post-1980 growth in Los Angeles County's Hispanic population was foreseeable at the time of the 1981 redistricting because it reflects a series of long-term demographic trends that were evident by 1981. The County's Hispanic population had increased significantly in each of the last three decennial censuses and the County's white non-Hispanic population showed a sharp decline between 1970 and 1980.

200. Spanish-surnamed persons made up 55.6 percent of the registered voters in the Hispanic Core in November 1988.

3. Accuracy of Post-Census Data

201. The Court finds that post-census data is a more accurate indicator than the 1980 Census of current demographic conditions in Los Angeles County. Specifically, the Court considers the PEPS estimates and projections to be a unique and reliable source of information for this purpose.

(a) Reliability of PEPS Data

202. PEPS had its genesis in a population research project begun at the University of California at Los Angeles and later transferred to the Los Angeles County government.

203. The process of producing a set of PEPS estimates and projections is referred to as an iteration. There have been two iterations of the PEPS process. The first iteration produced estimates for 1985 and projections for 1990. The second iteration produced estimates for 1987 and projections for 1989 and 1990.

204. The PEPS population estimates are based upon a combination of information from the 1980 Census, the California Department of Finance, the United States Bureau of the Census Current Population Survey, and administrative and vital records.

205. Estimates are based on observations of what occurred in the past according to administrative records. Projections study past and present trends and estimate future situations.

206. The racial and ethnic groups reported by PEPS include total population, Latino population, white non-Latino population, black population, and Asian/Other. The PEPS Latino category is designed to include persons whose origins are from Mexico, Central or South America. The PEPS black and Asian categories include persons who identify themselves as black and Spanish origin or Asian and Spanish origin.

207. PEPS made no attempt to estimate or project the number of citizens in Los Angeles County.

208. PEPS data has been relied on for planning purposes by the Los Angeles County government, including the Department of Health Services, the Department of Mental Health, the Sheriff's Department, the Superior Court system, the Municipal Court system and the Public Library system.

209. Population figures for Hispanics from PEPS data are consistent with data from the Census Bureau estimating that by 1985, the number of Hispanics in Los Angeles County had climbed to 2,742,700, an increase of nearly 700,000 from 1980.

210. Dr. Nancy Minter, who supervised most of the work on the first and second iterations, testified that in her opinion, the 1985 PEPS estimates are a more accurate reflection of current population in Los Angeles County than the 1980 Census at the countywide level and when the tract data is aggregated, to the supervisorial district level.

211. The Court concludes that the "glitch" in the 1987 PEPS estimates and the 1989 projections which consisted of the omission of certain death records for the cities of Long Beach and Pasadena from the calculations for the second PEPS iteration did not affect the 1985 population estimates generated by the first iteration.

212. Dr. Minter testified that in the second PEPS iteration, the white non-Hispanic population Countywide may have been underestimated by approximately 100,000 persons, and that the Asian/Other population may have been overestimated by approximately 100,000 persons.

213. Mr. Jerry Lubin, the director of the PEPS project, testified that after having discovered this "glitch," he never told PEPS users to use the 1985 rather than the 1987 estimates or to ignore the 1987 estimates.

214. Dr. O'Hare performed additional analyses of the 1985 estimates, the 1987 estimates and the 1989 projections after disclosure of the "glitch."

215. The comparison of PEPS estimates and projections with the Department of Finance and Census Bureau city estimates revealed an extremely strong level of consistency among the three sets of data. While the PEPS estimates and projections for Long Beach and Pasadena did show a greater degree of difference from the other two data sets than did the other five cities for 1987 and 1989, the degree of difference was relatively small.

216. The Court concludes that the missing data referred to by Dr. Minter and Mr. Lubin does not appear to have had any significant impact on the reliability of the second PEPS iteration as to total population data or as to the reliability of those estimates for the Hispanic population.

217. The Court concurs with the conclusion reached by Dr. O'Hare, that the 1985, 1987 and 1989 PEPS data are reliable population estimates and projections, and that, even with the "glitch," each set of data provides a more accurate reflection of Los Angeles County's current population than does the 1980 Census.

218. The Court is unwilling, therefore, to reject Dr. Estrada's estimate of Hispanic citizenship proportions in Hispanic Opportunity Districts I and II in 1989 which utilize PEPS projections and the 1987 PEPS estimates. ( See discussion of Estrada Plans infra)

219. As defendants' expert, Dr. W.A.V. Clark, testified, "the 1985, 1987, 1989 and 1990 [PEPS data] are all part of one project, to the extent that you can use the data and make comments about it. . . . So it is all part of one project, and I don't differentiate in my mind particularly between any one of those estimates or projections. I think of them all as having about the same reliability, recognizing that there [sic] all keyed back to a base line census point."

220. Dr. Clark testified that the reliability of PEPS data increases as it is aggregated, and agreed that when PEPS tract data is aggregated to the level of a Supervisorial district, "you would be on quite safe ground."

221. It is this Court's finding that the Los Angeles County's 1985 and 1987 PEPS tract-level estimates of total population and population by race and ethnicity; and the 1989 PEPS tract-level projections of total population and population by race and ethnicity are an acceptable and reliable basis under California law for the intercensal redistricting of Los Angeles County Supervisorial districts.

4. Citizen Voting Age Population

222. In measuring the citizen voting age population, the Court has considered 1980 Census sample data on citizenship, Hispanic voter registration from 1982 to 1988, and post-1980 estimates of citizen voting age population.

223. The Court finds that sample data from the 1980 Census on citizenship by age and ethnicity is the most reliable measure of the Hispanic citizen voting age population as it existed in 1980.

224. The Court, however, agrees with the contention of the United States and finds that, for practical reasons, decennial census counts of voting age citizen population cannot be an exclusive measure of geographic compactness. Total population data and voting age population data are available for redistricting purposes promptly after the decennial census is taken, while citizenship data is not released until several years later. For example, following the 1980 Census, the Bureau of the Census did not release citizenship data until 1983 and does not anticipate being prepared to do so after the 1990 Census until 1992-93.

225. No figures were published by the Census for the number of voting age citizens in major race/ethnicity groups in each census tract in Los Angeles County based upon the 1980 Census.

226. The Census Bureau prepared two special tabulations at the request of the Justice Department providing a breakdown of sample data concerning the number of voting age United States citizens, according to race and ethnicity and a cross-tabulation between self-identified Spanish-origin status and Spanish-surname status, among voting age citizens, by 1980 Census tract.

227. This special tract-level tabulation prepared by the Census found a total of 4,515,232 citizens of voting age in the County and a total of 659,368 Hispanic citizens of voting age in the County. With the sampling error, the number of Hispanic citizens of voting age in 1980 was between 649,172 and 669,564.

228. Many jurisdictions, including Los Angeles County, will be legally required to complete their redistrictings before citizenship data becomes available after the 1990 Census.

5. Voter Registration and Turnout

229. Between 1982 and 1988, Spanish-surname voter registration increased from 10.5 percent to 13.4 percent, and the estimated Hispanic voter registration from 12.3 percent to 14.6 percent in Los Angeles County.

230. Dr. Grofman reviewed data concerning voter registration and turnout and concluded that as a result of differential rates of voter registration and turnout between Hispanics and non-Hispanics in Los Angeles County, the proportion of Hispanic voting age citizens who are registered and who turnout to vote is considerably lower than non-Hispanics. Consequently, when Spanish-surnamed registered voters or Spanish-origin registered voters comprise more than 50 percent of the registered voters in a district, that translates into a situation in which the Hispanic citizen voting age population in the district will in fact be 50 percent or greater.

231. Statistics for Spanish-surnamed and estimated Spanish-origin registered voters were not available for Los Angeles County for any election prior to November 1982.

232. According to the Field Institute's California Opinion Index for January, 1988, among adult citizens eligible to vote in California, Hispanics were registered at lower rates than non-Hispanics in 1987. The official state registration figures were adjusted by the Field Institute to account for estimated "deadwood" and duplication in the voting rolls. The Field Institute also reported lower Hispanic turnout for the 1986 general election.

233. Dr. Minter, testified that voter registration is not used in PEPS because it is too volatile. The Court, however, finds the examination of Spanish-surname and estimated Spanish-origin registered voter data useful in determining a minimum rate for measuring Hispanic citizen voting age population.

6. Misreporting of Citizenship

234. There are no official Census Bureau adjustments to the 1980 data for misreporting of citizenship.

235. Dr. Jeffrey Passel asserts that, two million non-citizens nationwide falsely reported themselves as citizens in the 1980 Census. Passel's method for determining this misreporting was to compare the numbers of alien population based on the census count, which includes legal and undocumented aliens, with INS numbers of the legal resident population derived directly from the alien address registration system, or the I-53 data.

236. Dr. Passel's analysis concluded that census counts of naturalized citizens were higher than the estimate of naturalized citizens based on INS data.

237. Dr. Passel's studies are not considered corrections to the decennial census data as they were performed for research purposes. This research is based on national estimates with an unknown range of error. For example, Dr. Passel's national estimates of naturalization misreporting do not fully account for derived citizenship, that is, the acquisition of citizenship by a foreign-born child upon the naturalization of one or both parents. The greater the number of derived citizens, the more inaccurate are Dr. Passel's citizen corrections.

238. Defendants' experts, Dr. Clark and Professor Siegel testified that, in their opinions, the special tabulation of voting age citizen data from the 1980 Census is not accurate because a significant number of persons in Los Angeles County erroneously reported that they were United States citizens.

239. Dr. Clark did not rely upon the Census Bureau's special tabulation of voting age citizens for his analysis but instead developed a procedure to estimate citizen voting age population independently of the special tabulation. Dr. Clark testified that the adjustment factor was derived from the Warren/Passel methodology and applied to the Hispanic population in the County as a whole.

240. It is inappropriate, in the Court's mind, to substitute the estimates of Dr. Clark in place of the official Census data special tabulation. The procedure utilized by Dr. Clark does not take advantage of census tract-level information for voting age population or other variables more detailed than total population.

241. Dr. Clark applied his citizenship misreporting estimates to 1980 Census Hispanic total population data; to post-1980 estimates of citizen voting age population; and to modify the procedure of Dr. O'Hare for estimating Hispanic voter registration.

242. The difficulty the Court has with the Clark application of the Passel methodology is that the estimates of misreporting of citizenship employed by Dr. Passel relied upon national correction factors applied to local data. These are referred to as synthetic assumptions. Because such a synthetic correction procedure applies a constant factor to all subareas, local variations in the underlying error will necessarily produce inaccurate results. The greater the local variation, the greater the inaccuracy.

243. Professor Siegel testified that he reviewed and approved of Dr. Clark's estimates of voting age citizens, yet he did not know basic facts about how those estimates were performed, the number of self-reported Hispanic voting age citizens in Los Angeles County in 1980, the adjusted number used by Dr. Clark, nor the number of voting age Hispanics who supposedly misreported their citizenship.

244. Professor Siegel, while working for the Bureau of the Census, testified in Fair v. Klutznick, that "we . . . [the Bureau of Census] do not believe that an estimate of unlawful residents can be made which is of a quality sufficient for apportionment purposes." In a later case, Ridge v. Verity, Professor Siegel submitted a declaration stating that there existed an entirely feasible method by which undocumented aliens from the 1990 census could be excluded for purposes of congressional apportionment.

245. As an employee of the Census Bureau, Professor Siegel testified that synthetic adjustments for population undercount often produced "garbage" at the local level. In this litigation, Professor Siegel has used synthetic adjustments to estimate undercount of the Hispanic population in Los Angeles County, and has endorsed Dr. Clark's use of synthetic adjustments for citizenship misreporting in Los Angeles County.

246. In addition, the Court finds the Passel methodology problematic in its estimate of Hispanic citizen voting age population. The adjusted alien population used by Passel was for persons born in "Spanish" countries, regardless of whether the aliens identified themselves as Hispanics. To subtract non-Hispanic aliens born in these countries from the Spanish-origin population erroneously reduces the Spanish-origin citizen population.

247. The Court finds it noteworthy that the Heer/Passel study and Dr. Passel's data reported in the national-level study, demonstrate that even with a control for period of entry, the Mexican-born population in Los Angeles County was on average only half as likely to report being naturalized than was the Mexican-born population in the balance of the United States. As the table below illustrates:

Total Population Total Self-Reported Born in Mexico Naturalized Citizenship ---------------- ----------- -------------

Entered 1975-80 Natiowide 769,000 92,000 12.0% LA County 254,900 17,000 6.7% Outside LA County 514,100 75,000 14.6%

All Periods of Entry Natiowide 2,326,000 580,000 24.9% LA County 697,800 107,500 15.4% Outside LA County 1,628,200 472,500 29.0%

248. Having considered the estimates of Dr. Clark and Professor Siegel as well as the methodology utilized to derive these citizen voting age population estimates, the Court is not convinced that these estimates will produce a more accurate measure of voting age citizens than will the special tabulation of the Census. Moreover, the Court is unable to determine the magnitude of citizenship misreporting in the 1980 Census special tabulation data for the County and finds that substituting Dr. Clark's and Professor Siegel's estimates of citizen voting age population for the official Census data would be inappropriate.

7. Undercount of Hispanics

249. There are no official Census Bureau adjustments to the 1980 data for the undercount of the minority population.

250. According to Census publications, Hispanics were undercounted in the 1980 Census by 2.2 percent to 7.6 percent, Blacks by 5.5 percent and Whites by .7 percent. The Court agrees with the Garza plaintiffs that to arrive at the most realistic figures for the population of Los Angeles County if adjustments are made for overreporting of citizenship then such adjustments must likewise be made for undercount.

8. Spanish-Surname/Spanish-Origin

251. The Census has published a list of Spanish surnames used to identify persons of Spanish surname during the 1980 Census. This Spanish-surname identifier was included in the Los Angeles County Public Use Microdata Sample (PUMS) file as well as the sample detail file for Los Angeles County from which the Census Bureau prepared special tabulations.

252. Individuals with Spanish-surnames are sometimes not of Spanish-origin, while some persons without Spanish-surnames are of Spanish-origin. According to the 1980 Census, in Los Angeles County there were 574,120 voting age citizens with a Spanish surname and 659,375 voting age citizens who identified themselves as being of Spanish origin. Of the voting age citizens with a Spanish-surname, 88.7 percent were of Spanish origin. Of the voting age citizens without a Spanish surname, 3.6 percent were of Spanish origin.

253. The four-step procedure followed by United States' experts is outlined as follows:

(1) The number of Spanish-surnamed registered voters were totalled within voter registration precincts and census tracts, by matching lists of registered voters from the Los Angeles County Registrar-Recorder and the 1980 Census List of Spanish Surnames.

(2) Within the voting age citizen population for each census tract, the proportion of persons who were both Spanish surnamed and of Spanish origin among all persons with Spanish surnames was applied to the number of Spanish-surnamed registered voters whose residences lay within that tract. This produced an estimate of the number of Spanish-surnamed registered voters who were also of Spanish origin.

(3) Within the voting age citizen population for each census tract, the proportion of persons who did not have a Spanish surname but were of Spanish origin among all persons without a Spanish surname was applied to the number of registered voters without Spanish surnames. This produced an estimate of the number of registered voters who did not have Spanish surnames but were of Spanish origin.

(4) The two estimates of Spanish origin voters resulting from steps (2) and (3) were added together to derive the total number of registered voters of Spanish origin.

254. The estimates of Spanish-origin population derived by Dr. O'Hare are considered by this Court to be valid estimates of the number of registered voters of Spanish origin.

255. Dr. Clark modified Dr. O'Hare's methodology by reducing the number of voting age citizens in each census tract that were used to compute the estimation ratios in order to correct for misreporting of citizenship.

256. Since Dr. Clark's voter registration adjustments for misreporting of citizenship were predicated upon the proposed adjustments which the Court declined to adopt supra, the Court declines to adopt the adjustments to voter registration data proposed by Dr. Clark.

(a) Adjustments for "European Spanish"

257. The Court has great difficulty with the adjustments made for "European Spanish." The Court is not convinced that a clear determination can be made that Filipino, Cuban and "European" voters of Spanish origin in Los Angeles County vote differently from other voters of Spanish origin.

258. Moreover, in the 1980 Census, Spanish-origin status was determined from a separate question which asked "Is this person of Hispanic/Spanish Origin?" and then provided five choices: No (not Spanish Origin), Mexican, Puerto Rican, Cuban or Other Spanish Origin. For the ancestry question in the 1980 Census, respondents were required to fill in a blank in response to the question "What is this person's ancestry?" Ancestry Codes 1-99 reflected persons who reported Western European ancestry such as French or German. Codes 200-204 are identifiable with Spain (e.g. Spaniard, Catalonian). Codes 205 and 206 are "Spanish" and "Spanish-American."

259. Under Dr. Clark's definition of "European" Spanish population, anyone with a Spanish surname who was assigned an ancestry code from 1-99, or 200-206, was removed from the Spanish surname population, regardless of whether those persons had identified themselves as Spanish origin or not.

260. For purposes of this analysis, Dr. Clark reasoned that persons who identify themselves directly with Spain do not identify with the larger Hispanic community of persons of Mexican, Central or South American origin. The assumption is that persons who identified themselves as "Spanish" or of "Spanish-American" ancestry traced their decent directly to Spain and would not identify with the larger Hispanic community, which in Los Angeles County is predominantly of Mexican origin.

261. However, as the Census Bureau warns in its instructions regarding use of the Codes, ancestry is not a substitute for ethnicity.

262. An additional problem with Dr. Clark's analysis is that he factored out Spanish-surnamed persons with ancestry codes 1-99 and 200-206 regardless of whether they had identified themselves as Mexican Spanish-Origin, Puerto Rican Spanish-Origin, Cuban Spanish-Origin (whom he also removed separately), Other Spanish Origin, or not Spanish-Origin at all. A sizeable number of voting age citizens with ancestry codes 205 and 206 were of Mexican Spanish Origin ethnicity (14,240). Dr. Clark factored out these individuals because they had Spanish surnames and wrote in the word "Spanish" or had designated "Spanish-American" ancestry.

263. The Court adopts the counts of Spanish surname and estimated Spanish-origin voters presented by the United States, as reasonably and accurately reflecting Hispanic voter registration and turnout in Los Angeles County between 1982 and 1988.

9. Deadwood

264. Defendants contend that the registered voter statistics provided are flawed since they contain the names of many persons who no longer reside within their listed precinct or those who are deceased. Defendants further contend that this "deadwood" is exceedingly Democratic, containing a disproportionate number of Hispanics.

265. Pursuant to the laws of the State of California, the Los Angeles County Registrar-Recorder is responsible for conducting the Registration Confirmation and Outreach Program ("RCOP") designed to identify and remove "deadwood."

266. RCOP is conducted in January of every year and consists of sending a registration confirmation postcard to voters at the residence shown on the voting rolls.

267. In even-numbered years, the confirmation postcard is sent to all registered voters in the County while in odd-numbered years it is sent only to those persons who failed to vote in the general election in the preceding November.

268. The postcard requests that it not be forwarded to another address even if the voter has moved and left a forwarding address. Thus, if the United States Postal Service is unable to deliver the card at the address listed, the card is returned to the Registrar-Recorder's office with a notation as to why that card is undeliverable.

269. If the postcard indicates the voter has moved and left no forwarding address or if their forwarding address reflects that they moved out of the County, the Registrar-Recorder's office cancels the voter registration. If the postcard indicated the person has moved within the County, the voting rolls are changed to reflect the new address.

270. Defendants are correct in their assertion that the presence of "deadwood" on the voting rolls is a problem and the Court is not completely persuaded that RCOP is effective as the sole procedure for removing such "deadwood." However, defendants have not demonstrated that the "deadwood," even if improperly remaining on the voting rolls, is disproportionately Hispanic.

271. The proportion of persons identified as Democratic who were cancelled under the provisions of the RCOP for the years 1984 to 1990 did not constitute a disproportionate share of the total cancellations in those years.

272. In addition, voters' surnames, party identifications, registration precinct numbers and census tract numbers can be retrieved from the computerized lists of cancellations provided by the Registrar-Recorder. These computerized lists can be matched with the Census Bureau's 1980 list of Spanish surnames, to yield accurate counts of Spanish-surnamed voter cancellations.

Analysis by Spanish Surnamed Voters of Residency Confirmation and Outreach Program (RCOP)

Total Spanish Surname ----- ---------------

November 1988 Registration 3,765,368 502,885 (13.4%) January 1989 RCOP Cancellations 132,424 14,522 (11.0%) January 1990 PCOP Cancellations 245,138 27,102 (11.1%)

273. The Court does not find that Hispanic persons constitute a differentially greater proportion of "deadwood."

274. In addition to the registered voter data, plaintiff's expert Dr. Estrada and defendants' experts, Drs. Freeman, Minter and Clark, used differing methodologies to estimate post-1980 citizen voting age population in Los Angeles County based principally on special Census Bureau tabulations of the 1980 citizen voting age population and PEPS estimates and projections.

10. Plaintiffs' Illustrative Plans

(a) The Grofman Plans

276. In devising these plans, Dr. Grofman considered such standard criteria for redistricting as compliance with the one-person, one-vote rule and avoiding minority vote dilution.

277. In each of Dr. Grofman's five plans, the total population deviation is less than 10 percent, using a valid total population base.

278. In the Grofman Plans, the Hispanic total population percentage in the most heavily Hispanic district increased during the 1980's while the white population decreased. Likewise, in each of the Grofman plans, one district as of November 1988 had an estimated Spanish-origin registered voter proportion in excess of 50 percent.

279. In Grofman Plans 3, 4 and 5, one district as of November 1988 had an estimated Spanish-origin voter turnout proportion in excess of 50 percent.

(b) The Estrada Plans

280. Dr. Estrada proposed two illustrative supervisorial districts, Hispanic Opportunity Districts I and II (HOD I and HOD II). Neither HOD I nor HOD II had an Hispanic voting age citizenship population majority in 1980. In HOD I, there were 260,243 Hispanic citizens of voting age, which is 46.9 percent of the total citizen voting age population of HOD I. In HOD II, Hispanic voting age citizens comprise 282,676, or 46.9 percent of the total HOD II population. Dr. Estrada concluded that an Hispanic citizen voting age majority district could have been created in 1985.

281. In arriving at his conclusions and in devising his illustrative plans, Dr. Estrada considered such demographic factors as: Hispanic children ages 10-17 who would turn 18 over the course of the 1980's and the higher citizenship rate associated with this group; the possible effects of mortality, out-migration and in-migration of citizens upon the 1980 base citizen voting age populations; and, the likelihood that both Asian and Hispanic citizenship rates have diminished from the 1980 rates.

282. Based on 1985 PEPS data, Dr. Estrada's HOD I and HOD II, illustrated that an Hispanic majority citizen voting age district can be created in HOD I. Assuming that citizen voting age population rates remain equal to the 1980 rates for all race/ethnic groups, by 1985 51 percent of the citizen voting age population of the HOD I district was Hispanic. Assuming that 55 percent rather than 46 percent of the Asian/Other population were voting age citizens, an Hispanic citizen voting age majority could still be created in HOD I in 1985.

283. The Garza plaintiffs demonstrated that an Hispanic majority citizen voting age district could be created in HOD I and HOD II using 1987 and 1989 PEPS data as well.

284. The Court finds that under a variety of reasonable demographic techniques, demonstrated by both plaintiffs' and defendants' experts, the demographic changes that have occurred since 1980 make it possible to draw a Supervisorial district in which Hispanics constitute a majority of the citizen voting age population.

E. POLITICAL COHESIVENESS

1. Hispanic Candidacies in Los Angeles County 1978-1989

(a) Contests for Los Angeles County Supervisor

286. Since 1978, Hispanic candidates have run in five supervisorial election contests: District 1 (1978 and 1982); District 3 (1978 and 1982); and District 5 (1988).

287. In the 1978 primary election in District 1, one Hispanic candidate, Alfonso Lavin, ran against three non-Hispanics, including incumbent Peter Schabarum. Lavin is not a recognizable Spanish surname. Lavin received 7.3 percent of the vote and Schabarum received 55.7 percent of the vote.

288. In the 1978 primary election in District 3, two Hispanic candidates, Rosalio Munoz and Gonzalo Molina, ran against the incumbent, Edmund Edelman. Munoz received 11.5 percent of the vote, Molina received 14.0 percent of the vote and Edelman received 74.5 percent of the vote.

289. In the 1982 primary election in District 1, Lavin again ran against incumbent Schabarum, as well as another non-Hispanic candidate. Schabarum received 64.5 percent and Lavin received 10.5 percent of the vote.

290. In the 1982 primary in District 3 Rosalio Munoz ran against incumbent, Edelman, and two other non-Hispanic candidates. Munoz received 11.8 percent of the vote and came in second behind Edelman who received 72.1 percent of the vote.

291. In the 1988 primary election for Supervisor in District 5, two Hispanic candidates, M. Enriquez-Marquez and Jose Galvan, ran against incumbent Antonovich, and seven other non-Hispanic candidates. At the time of the election, Spanish-surnamed voters comprised approximately 8 percent of the registered voters in District 5. Enriquez-Marquez placed fourth with 2.3 percent of the vote; Galvan placed last with 0.5 percent of the vote. Antonovich, who received 44.8 percent of the vote, was forced into a run-off with Baxter Ward, who received 22.4 percent of the vote in the primary.

292. All the Hispanic candidates in these supervisorial contests were minor candidates with relatively little campaign financing who had no realistic chance of mounting a serious challenge to the incumbent Supervisor.

293. Dr. Grofman analyzed three of the elections for Supervisor since 1978 which involved Hispanic candidates with Spanish surnames and one election in 1982 involving an Hispanic candidate without a Spanish surname. Dr. Grofman found that even with respect to these very minor candidates, there were substantial differences between the levels of support received from Hispanics and that received from non-Hispanics.

294. The Court further finds that analyses of Supervisorial elections are not dispositive of political cohesiveness. Rather, plaintiffs were entitled to attempt to establish political cohesion through the study and analysis of other elections within the County of Los Angeles.

295. No specific number of elections need be studied in order to determine whether voting is polarized in Los Angeles County along ethnic lines.

(b) Other Nonpartisan County Contests

296. Since 1978 there have been two election contests for Los Angeles County offices other than Supervisor in which Hispanic candidates have run: County Sheriff (1982) and County Assessor (1986).

297. In the 1982 primary election for County Sheriff, two Hispanic candidates, Alex Jacinto and Robert Feliciano, received 6 percent and approximately 20 percent of the total vote, respectively, against candidates Sherman Block, Charles Greene, and three other non-Hispanic candidates. The winner, incumbent Block, received 63 percent of the total vote.

298. In the 1986 primary election for County Assessor, the Hispanic candidate, Sid Delgado, received roughly 12 percent of the total vote against a field of eleven non-Hispanic candidates for an open seat. Delgado's share of the votes cast placed him in fourth place in the race, with the leading candidate, John Lynch, receiving 21 percent of the vote.

299. Based upon the relative vote shares and campaign expenditures of the Hispanic candidates in the 1982 Sheriff's race and the 1986 Assessor's race, the Court agrees with Dr. Grofman's conclusion that these were relatively minor candidates.

300. Dr. Grofman conducted analyses of the 1982 primary race for Sheriff and the 1986 primary race for Assessor. Based upon his analysis, Dr. Grofman found a dramatic divergence between the support levels from Hispanics versus those from non-Hispanics for the two Hispanic candidates in the 1982 Sheriff primary. The Hispanic support level, based on Spanish-origin data, for Jacinto and Feliciano combined was estimated at 80 percent, while the support of non-Hispanic voters for these two candidates was estimated at only 20 percent. According to the estimates of Dr. Grofman, Feliciano was the plurality choice of the Hispanic voters. Similarly in the 1986 Assessor race, the Hispanic candidate, Delgado, was estimated to have been the plurality choice of Hispanic voters in a very crowded field of candidates with 35 percent support among Hispanics compared to only 10 percent support from non-Hispanics.

(c) Non-Countywide Elections

301. Since 1983 there have been seven elections for Los Angeles City Council in which Hispanics have run for office: District 1 (1987 and 1989); District 4 (1983); District 7 (1989); and District 14 (1983, 1985, and 1987).

302. In the 1983 primary election for District 14, two Hispanic candidates, David Sanchez and Steve Rodriguez, received 2.2 percent and 42.6 percent of the vote respectively in a field of six candidates, which included the Anglo incumbent, Art Snyder, and three other non-Hispanic candidates. Snyder won the election with 50.1 percent of the vote.

303. In a subsequent special election in District 14 in December 1985, six Hispanic candidates and one non-Hispanic candidate ran for an open seat created by the resignation of Councilman Snyder. One of the Hispanic candidates, Richard Alatorre, won the election with 59.58 percent of the vote and became only the second Hispanic elected to the Los Angeles City Council since at least 1900. Dorothy Andromidas, the sole non-Hispanic candidate, received about one percent of the votes cast.

304. As a result of the 1985 lawsuit filed against the City of Los Angeles to remedy the fragmentation of the Hispanic population concentrations in the eastern part of the city, United States, et al. v. City of Los Angeles, No. CV 85-7739 JMI (JRx) (C.D.Cal. 1985), the City of Los Angeles chose to redraw the council districts so as to create a second Hispanic majority city council district, Council District 1.

305. As of the 1988 general election, persons of Spanish origin constituted approximately 46 percent of the registered voters in Council District 1.

306. In the special election in Council District 1 on February 3, 1987, two Hispanic candidates ran against two non-Hispanic candidates for the vacant seat. The two Hispanic candidates received 82.5 percent of the vote. One of the Hispanic candidates, Gloria Molina, was elected with 57.0 percent of the vote.

307. The special elections for Los Angeles City Council District 14 in 1985 and Los Angeles City Council District 1 in 1987 both occurred in districts which contained a clear majority Hispanic population.

308. In April 1989, eight candidates, including two Hispanic candidates, Irene Tovar and Richard Yanez, ran in the primary election for Los Angeles City Council District 7. Tovar received 9.5 percent of the total vote; Yanez received 1.3 percent; and the incumbent, Ernai Bernardi, received 41.9 percent.

309. Dr. Grofman analyzed the 1983 primary contest in City Council District 14 and the 1989 primary in City Council District 7. The population of District 14, which is located essentially within the Hispanic Core, is greater than 60 percent Hispanic. In 1983, Spanish-surnamed persons constituted 49.9 percent of the registered voters in District 14. In contrast, District 7, which is located in the San Fernando Valley area, has a much smaller proportion of Hispanics among its population. As of the 1988 general election, only 25.49 percent of the registered voters in Council District 7 were of Spanish origin. Dr. Grofman's analysis found high levels of Hispanic political cohesion in both of these contests.

(d) Countywide Partisan Elections

310. According to the 1980 Census, three of the congressional districts in the County had Hispanic citizen voting age populations of at least 35 percent: Congressional District 25 (42.1 percent), Congressional District 30 (37.3 percent), and Congressional District 34 (35.2 percent).

311. As of the 1982 general election, Congressional Districts 25, 30 and 34 also contained the greatest proportions of Hispanic registered voters of all the congressional districts in the County.

312. In the 1982 Democratic primary elections, an Hispanic candidate prevailed in each of these three congressional districts: Edward Roybal in Congress District 25, Matthew Martinez in Congressional District 30, and Esteban Torres in Congressional District 34. The three nominees went on to victory in the general elections of 1982. Moreover, these Hispanic candidates continued to prevail in these three congressional districts for all subsequent Democratic primary and general elections.

313. These three congressional districts are located within the Hispanic Core area of Los Angeles County.

314. No Hispanic has been elected in any other congressional district wholly within Los Angeles County since at least 1982.

315. As of the 1982 general election, State Senate Districts 24 and 26 also contained the greatest proportions of Hispanic registered voters of all the state senate districts in Los Angeles County.

316. In the 1982 Democratic primary elections, an Hispanic candidate prevailed in both of these state senate districts: Art Torres in Senate District 24 and Joseph Montoya in Senate District 26. In turn these nominees went on to victory in the general elections of 1982. Moreover, these Hispanic candidates have prevailed in these state senate district for all subsequent Democratic primary and general elections.

317. Both of these senate districts are included within the Hispanic Core area of Los Angeles County.

318. No Hispanic has been elected in any other state senate district wholly within Los Angeles County since at least 1982.

319. According to the 1980 Census, four of the state assembly districts in Los Angeles County had Hispanic citizen voting age populations of at least 35 percent. Assembly District 55 (41.4 percent), Assembly District 56 (57.6 percent), Assembly District 59 (43.5 percent) and Assembly District 60 (37.4 percent).

320. As of the 1982 general election, Assembly Districts 55, 56, 59 and 60 also contained the greatest proportions of Hispanic registered voters of all the state assembly districts in Los Angeles County.

321. In three of these four state assembly districts (Assembly District 55, Assembly District 56, and Assembly District 59), Hispanic candidates prevailed in both the Democratic primary and general elections of 1982. In turn, Hispanic candidates prevailed in these three districts in all subsequent Democratic primary and general elections through 1988. In only one of these assembly districts, District 60, did a non-Hispanic candidate prevail in the Democratic primary and general election contests of 1982 and subsequent years. No Hispanic candidate has run in the Democratic primary for Assembly District 60 since 1982.

322. Each of the three assembly districts in which Hispanic candidates have prevailed is located within the Hispanic Core area of Los Angeles County.

2. Analysis of Ethnically Polarized Voting

(a) Methodology

324. Dr. Grofman has been recognized as an expert witness in racial or ethnic vote dilution in numerous federal court cases. His testimony concerning racially polarized voting was adopted by both the District Court and the United States Supreme Court in Gingles v. Edmisten, 590 F. Supp. 345 (E.D.N.C. 1984), aff'd in part, rev'd in part, Thornburg v. Gingles, 478 U.S. 30, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986). In addition, Dr. Grofman was the sole expert witness for the plaintiffs in Cruz Gomez v. City of Watsonville, 863 F.2d 1407 (9th Cir. 1988), cert. denied, 489 U.S. 1080, 109 S.Ct. 1534, 103 L.Ed.2d 839 (1989), in which his opinions were adopted by the Ninth Circuit Court of Appeals.

325. Drs. Lichtman and Grofman used standard methods in the analysis of electoral data to determine whether voting is ethnically polarized in Los Angeles County elections, considering Hispanics versus non-Hispanics, and whether the existing system of supervisorial districts impedes the ability of Hispanic citizens to elect representatives of their choice and fully participate in the political process. Their analyses of ethnically polarized voting follow procedures that are consistent with the standards prescribed by the Supreme Court in Gingles, 478 U.S. at 52-59, 106 S.Ct. at 2767-2770.

326. Plaintiffs' experts determined the voting behavior of Hispanics and non-Hispanics in Los Angeles County by comparing the ethnic composition of precincts to the division of the vote among competing candidates at each precinct. Ecological regression, the standard method for inferring the behavior of population groups from data collected for aggregate units, was used to estimate the voting behavior of non-Hispanics and Hispanics. The regression methodology generates prediction equations that indicate how voting responds to variations in the proportions of Hispanics and non-Hispanics in each precinct. These equations can provide the information needed to estimate the average voting of non-Hispanics and Hispanics, respectively, in the election district under analysis.

327. Ecological regression, therefore, provides estimates of the average voting behavior of the groups in question. It does not purport to determine the voting behavior of individuals. Not does it purport to estimate exactly the voting behavior of non-Hispanics and Hispanics in each precinct.

328. Drs. Lichtman and Grofman also utilized a technique termed "extreme case analysis." This technique examines the actual choices of voters in the most heavily non-Hispanic and the most heavily Hispanic precincts in a jurisdiction. If voting is polarized along ethnic lines, there should be differences in the percentages of votes going to the non-Hispanic and Hispanic candidates in the most heavily non-Hispanic and most heavily Hispanic precincts.

329. Ecological regression and extreme case analysis were also supplemented by the examination of squared correlation coefficients, an indicator of the reliability of a finding of polarized voting. The possible value of the squared correlation coefficient (R2) varies from 0 to 1.0, with values close to 1.0 indicating that the percentage of the vote case for the Hispanic candidates can be nearly perfectly predicted from the Hispanic versus non-Hispanic composition of political units. Although no particular value of R2 arbitrarily defines the distinction between "high" and "low", social scientists often find values in excess of .25 to be indicative of a substantial relationship between variables and generally consider values of .5 or greater as indicative of a very strong relationship.

330. As indicated by the Gingles decision, plaintiffs' experts divided the analysis of polarized voting into two components: the degree to which the Hispanic electorate cohesively supports Hispanic candidates for public office and the degree to which the non-Hispanic electorate bloc votes for non-Hispanic candidates.

331. The analysis of Hispanic cohesion and non-Hispanic bloc voting provides an indication of whether Hispanic voters have an opportunity to elect candidates of their choice in the existing supervisorial districts. In particular, the issue is whether bloc voting by non-Hispanics will normally be sufficient to defeat Hispanic candidates in the existing Supervisorial districts with the greatest Hispanic concentration.

332. Dr. Grofman analyzed eight nonpartisan elections involving non-Hispanic versus Hispanic candidates in Los Angeles County from 1978 to 1989; four were for County Supervisor, one for sheriff, one for assessor and two for Los Angeles City Council. Of the eight contests analyzed, only the assessor's contest was an election for an open seat.

333. In addition, Dr. Lichtman analyzed 12 open-seat partisan elections involving Hispanic versus non-Hispanic candidates for U.S. Congress, state senate, and state assembly from 1982 to 1988. Of the 12 contests analyzed, five were primaries, four Democratic and one Republican, and several were general elections, including one special run-off election.

334. Dr. Lichtman focused on open-seat elections because generally they are the most hotly contested of all races and provide the clearest indication of whether or not Hispanics and non-Hispanics systematically differ in their choices of candidates.

335. There are no substantive differences in results of the ecological regression and extreme case analyses between the analyses based upon Spanish surname and the analyses based upon Spanish origin data.

336. There is no dispute that plaintiffs' experts accurately computed and reported the results of the application of the ecological regression methodology. Dr. Jerome Sacks, a statistician and one of the defendants' experts, replicated plaintiffs' analyses and produced results that were not substantively different.

(b) Results of Analysis

337. The analysis of polarized voting in this case centered on the exposition and critique of ecological regression as a technique for analyzing group voting behavior.

338. Defendants presented the testimony of three statisticians, Dr. Stephen Klein, Dr. Jerome Sacks, and Dr. David Freedman, who individually and collectively criticized the use of the ecological regression methodology to analyze group voting behavior.

339. Defendants' experts do not dispute that as a general matter in the elections analyzed by plaintiffs' experts the proportion of the vote for the Hispanic candidates increases as the proportion of Hispanics in the precinct increases.

340. The ecological regression methodology can produce physically impossible results.

341. In the 20 elections analyzed by Drs. Grofman and Lichtman, physically impossible results were produced for only four of the 40 estimates of Hispanic and non-Hispanic voting. All involved estimates of Hispanic support for the Hispanic candidate that exceeded 100 percent.

342. All four physically impossible estimates were for general elections and three of four were just a few percentage points over the 100 percent mark.

343. Voting analysts often encounter estimates over 100 percent when voting is highly polarized and the slope of the ecological regression line is steeply pitched. Dr. Lichtman testified that in numerous jurisdictions where he encountered such estimates, he was able to verify the accuracy of the method of bounds through examination of nearly homogeneous precincts that included a majority of the population group in question.

344. Extreme case analysis also shows the accuracy of using the method of bounds for the one instance in which an estimate departs substantially from 100 percent, the estimate of the percent of Hispanics voting for the Hispanic candidate in the 1982 general election is Assembly District 52. Although there are no heavily Hispanic precincts in Assembly District 52, examination of Dr. Sacks' scattergram for this election shows that there are a large number of precincts in which the percentage of registrants with a Spanish origin is 10 percent or less. Dr. Lichtman's extreme case analysis shows that in precincts where Spanish-origin registrants are 10 percent or less, 25 percent of the vote was actually cast for the Hispanic candidate. This percentage conforms almost exactly to the 24 percent non-Hispanic crossover vote derived by the methods of bounds employed by Dr. Lichtman.

345. Defendants' experts also contend that ecological regression is unreliable because it depends on the unreasonable assumption that voting behavior is constant across precincts except for random variation. The regression equation assumes that Hispanics give the same level of support to Hispanic candidates in every precinct.

346. The Court agrees with plaintiffs that the so-called constancy assumption does not significantly undermine the reliability of the estimates gained through the ecological regression methodology in this case.

347. The ecological regression technique is designed to provide accurate estimates only of average group voting behavior in a particular jurisdiction. As a result, the technique can yield accurate estimates even in the presence of substantial random variations in voting behavior within the studied jurisdiction. The technique can also produce accurate estimates in the presence non-random variations, so long as such variations are not related to the percentage of Hispanics within a jurisdiction.

348. Defendants' experts testified that an omitted variable or a variable related both to voting behavior and to the percentage of Hispanic registrants in a precinct may be distorting the ecological regression analysis. The degree of bias will depend on the strength of the omitted variable's independent influence on voting behavior and on the strength of its relationship to the percentage of Hispanics in a precinct.

349. Defendants' experts advance the theory that such variables as Democratic affiliation and low socioeconomic status impact on voting behavior by overriding ethnic affiliation.

350. While in theory there exists a possibility that ecological regression could overestimate the results of ecological regression, experts for defendants have failed to demonstrate that there is in fact any substantial bias resulting from the omitted variable problem in Los Angeles County.

351. In a further attempt to discredit the reliability of the ecological regression technique, defendants' experts developed the "neighborhood model" to provide an alternative method of measuring ethnically polarized voting. The neighborhood model posits that all voters within a precinct vote alike irrespective of ethnic diversity within such a precinct.

352. At its logical boundaries, the omitted or contextual variable theory blends into the neighborhood model: both theories posit that non-ethnic factors impact on voting behavior to the extent of overriding ethnic affiliation. Thus, the greater the number of asserted contextual variables related to the ecological composition of the precinct, the more the omitted variable theory begins to resemble the neighborhood model.

353. The Court concludes that the neighborhood model's emphasis on the ecological structure of a precinct as a determinant of voting behavior impedes it from detecting the presence of polarized voting. As such, it is not a reliable method of inferring group voting behavior.

354. Defendants' critique of plaintiffs' squared correlation coefficient (R2) and extreme case analyses reiterates their constancy assumption objection to the ecological regression methodology.

355. The Court finds that the ecological regression and extreme case analysis performed by plaintiffs' experts, as supplemented by the analysis of correlation coefficients are sufficiently reliable to make the requisite determinations about polarized voting between Hispanics and non-Hispanics.

356. The results of ecological regression and extreme case analysis reveal that Hispanic and non-Hispanic voters in Los Angeles County are polarized along ethnic lines in their choices of candidates.

3. Cohesiveness of Hispanic Voters

357. The results of the ecological regression analyses demonstrated that for all elections analyzed, Hispanic voters generally preferred Hispanic candidates over non-Hispanic candidates. In 15 of the 19 elections studied, a majority of voters with Spanish surnames voted for Hispanic candidates. In 14 of these 15 contests, the Hispanic vote for the Hispanic candidates was much higher, equal to or greater than 32 percentage points, than the non-Hispanic vote for the Hispanic candidates. In 14 of the 19 elections, voters with Spanish surnames voted for Hispanic candidates at a level equal to or greater than the 60 percent that is generally considered to be a landslide victory in American political history.

358. Among the 19 contests studied, Dr. Lichtman properly isolated for analysis the eight nonpartisan contests and the four partisan primary contests. The Supervisorial elections are nonpartisan contests in which candidates compete without explicit party identification. Similarly, in partisan primary contests candidates compete under the same label so that the influence of party identification is eliminated. Since a better that two-thirds majority of Hispanics in the County are Democrats, their behavior in general elections might be influenced by party affiliation in the sense that they may be more likely to vote for the Democratic nominee. Thus, as compared to general election contests, nonpartisan and primary races provide the most stringent test of Hispanic cohesion.

359. In eight of the 12 nonpartisan and partisan primary elections, a majority of voters with Spanish surnames voted for Hispanic candidates. In seven of the 12 contests, a 70 percent or greater majority of voters with Spanish surnames united behind Hispanic candidates. On average, about 64 percent of the Spanish-surnamed voters supported Hispanic candidates in these contests.

360. There were special circumstances involved in the five nonpartisan contests in which the cohesion levels were lower than 60 percent. The five contests include the 1986 contest for assessor and the four Supervisorial contests.

361. In the assessor's contest, although the lone Hispanic candidate, Delgado, failed to gain majority support from Hispanic voters, he still finished in first place among Hispanic voters despite a crowded field that included 12 candidates. Delgado garnered the support of 38 percent of voters with a Spanish surname and 35 percent of voters with a Spanish origin. The second-place finisher among Hispanic voters garnered the support of 20 percent of the voters with either a Spanish surname or a Spanish origin. Overall, as a result of Delgado's support among Hispanics, he finished in fourth place among the 12 candidates. Similarly, in the 1988 Supervisorial race in District 5, Enriquez-Marquez, although a much weaker candidate overall than Delgado, finished first among Hispanic voters in a likewise crowded field of ten candidates. Enriquez-Marquez garnered the support of 36 percent of voters with a Spanish surname and 33 percent of voters with a Spanish origin. The second-place finisher among Hispanic voters garnered the support of 21 percent of voters with a Spanish surname.

362. In only one of four Supervisorial contests analyzed by Dr. Grofman did a majority of Spanish-surname or Spanish-origin voters support the Hispanic candidates. The remaining three Supervisorial contests, however, involved relatively marginal Hispanic candidates. In the 1982 primary contests in Supervisorial Districts 1 and 3, and in the 1988 primary contest in Supervisorial District 5, the Hispanic vote for the Hispanic candidates was much higher than their overall percentages. For these three Supervisorial contests, the Hispanic candidates received a mean vote of 37 percent from Spanish-surnamed voters and a mean vote of 34 percent from Spanish-origin voters compared to an overall mean vote of but 8 percent.

363. The 1982 primary contests provide a useful means of analyzing Hispanic cohesion since six of the 19 elections analyzed were held on primary election day in 1982. In contests for U.S. Congress in the 1982 primary in Congressional District 30 and Congressional District 34, the Hispanic candidates received 78 and 90 percent, respectively, of the vote of Spanish-surnamed voters and 78 and 88 percent, respectively, of the vote of Spanish-origin voters. In the contest for state assembly in the 1982 primary for Assembly District 59, the Hispanic candidates received 83 percent of the vote of Spanish-origin voters. In the 1982 primary contest for sheriff, the Hispanic candidates received 85 percent of the vote of Spanish-surnamed voters and 80 percent of the vote of Spanish-origin voters. Only the relatively marginal candidates for Supervisor in Districts 1 and 3 in the 1982 primary received less than majority support from Hispanics. In District 1, the Hispanic candidate received 21 percent of the vote of Spanish-surnamed voters and 19 percent of the vote of Spanish-origin voters. In District 3, the Hispanic candidate received 44 percent of the vote of Spanish-surnamed voters and 41 percent of the vote of Spanish-origin voters. For all six contests, Hispanic candidates garnered a mean vote of 67 percent from Spanish-origin voters.

364. Dr. Lichtman's analysis of partisan elections in Los Angeles County demonstrates strong political cohesion among Hispanics. In the four Democratic Party primary elections he analyzed, Spanish origin voters are estimated to have provided, on average, 85 percent of their vote for Hispanic candidates. Hispanic candidates received an average of 94 percent of the vote of Spanish-origin voters in the eight general elections Dr. Lichtman analyzed.

365. The Court finds that Hispanic political cohesiveness is strong when Hispanic candidates have a realistic chance of winning.

366. For all 19 elections analyzed, the reliability of the findings of polarized voting is corroborated by extremely high values of the squared correlation coefficient (R2). Whether Spanish-surname or Spanish-origin data are used, in all but five contests, the value of the squared correlation coefficient is at least equal to 0.65. For all 19 elections, moreover, the finding of racial polarization attains a level of statistical significance equal to or greater than the conventional standards used in social science. Researchers generally accept as reliable results for which statistical significance equals or exceeds the conventional standards of either .05 (corresponding to a five in one-hundred probability of obtaining results from chance or random factors) or .01 (corresponding to a one is one-hundred probability). For all 19 elections studied, the statistical significance is better than .00001 (corresponding to one in one-hundred thousand probability of obtaining the results from chance or random factors). The likelihood of obtaining any of these given results under the random factors hypothesis is low and the likelihood of obtaining the consistent pattern of these results is virtually zero.

367. The results of ecological regression analysis are corroborated by the findings of extreme case analysis, a technique that examines the actual vote cast in precincts that are heavily Hispanic or heavily non-Hispanic in their ethnic composition. The results of the extreme case analyses in this case were consistent with, and bolstered the reliability of the results of ecological regression.

368. For all 11 partisan contests studied, the Hispanic candidates received a greater than landslide majority, 60 percent or more, of the votes actually cast in more than 80 percent Hispanic precincts, whether Spanish-surname or Spanish-origin data are used. In contrast, only in the 1982 general election for state assembly in Assembly District 56 and in the 1986 special run-off and general elections for state assembly in Assembly District 55, did the Hispanic candidate receive more than 50 percent of the votes actually cast in the more than 90 percent non-Hispanic precincts. For the eight nonpartisan elections studied, the Hispanic candidates received a majority of the vote cast in more than 80 percent or more than 90 percent Hispanic precincts in three elections. In all eleven partisan elections, the Hispanic candidates received a much higher vote in the heavily Hispanic precincts than in the more than 90 percent non-Hispanic precincts.

F. NON-HISPANIC BLOC VOTING

369. Plaintiffs did not present evidence of white bloc voting. For most of their analyses, they combined Anglos, Blacks and Asians into a non-Hispanic bloc. The potential distorting effect of this construct is lessened by the fact that for most of the elections analyzed, the Black and Asian percentage of the electorate was not significant. Moreover, given the demographic reality of the Hispanic Core in Los Angeles County, if 40 percent of the registered voters in a given precinct are Hispanic, the precinct will likely be predominantly Hispanic in its overall population.

370. Where a racial or ethnic group is only a small component of the electorate, its voting behavior would not have a significant effect on the two-group ecological regression estimates of voting behavior.

371. Of the elections analyzed by plaintiffs' experts non-Hispanic voters provided majority support for the Hispanic candidates in only three elections, all partisan general election contests in which party affiliation often influences the behavior of voters (the 1982 general election contests in Senate District 24 and Assembly District 56 and the 1986 general election contest in Assembly District 55). Overall, for all 19 contests studied, the mean crossover vote for Hispanic candidates among non-Spanish-surnamed voters was 27 percent, compared to a bloc vote of 76 percent for non-Hispanic candidates.

372. In the 12 non-partisan or partisan primary elections non-Hispanic voters did not provide a crossover vote of greater than 34 percent for the Hispanic candidates, whether Spanish-surname or Spanish-origin data are used. Overall, for these 12 elections, the mean crossover vote for Hispanic candidates by non-Spanish-surnamed voters is 17 percent.

373. The results of extreme case analysis corroborate the findings of strong bloc voting by non-Hispanics. Of all 19 contests studied, only in the 1982 general election contest for state assembly in Assembly District 56 and in the 1986 special runoff and general election contests for state assembly in Assembly District 55, did Hispanic candidates receive a majority of the vote actually cast in the more than 90 percent non-Hispanic precincts, whether Spanish-surname or Spanish-origin data are used. Considering only the 12 nonpartisan and partisan primary contests, in no instance did Hispanic candidates receive more than 42 percent of the vote cast in the more than 90 percent non-Hispanic precincts. Overall, for these 12 elections, the mean vote for Hispanic candidates in the more than 90 percent non-Hispanic precincts, using Spanish-surname data, is 19 percent.

374. For several of the elections analyzed through ecological regression, extreme case analysis provides especially strong confirmation of the bloc voting results, because a majority or near-majority of non-Hispanic registrants reside in the more than 90 percent non-Hispanic precincts. Dr. Jerome Sacks, an expert for the defendants, testified that if about 35 to 40 percent of a group reside within "homogeneous" precincts, which he defines as precincts in which the group comprises 90 percent or more of the defined population, then the ecological regression analysis will produce reliable results for that group because the homogeneous precincts anchor the regression line.

375. Specifically, Dr. Sacks testified that in Los Angeles County there are sufficient homogenous precincts Countywide and in Supervisorial Districts 3 and 5 to have confidence that the regression estimates for non-Hispanics voting behavior are reliable.

376. Dr. Sacks' analysis provided five test cases of the reliability of the regression analysis for non-Hispanics: the 1978 and 1982 primaries for Supervisor in District 3; the 1982 Countywide primary for sheriff, the 1986 Countywide primary for assessor; and the 1988 primary for Supervisor in District 5. For these contests, the following table compares ecological regression and extreme case results for non-Hispanics using Spanish-surname data.

90% + Non-Hispanic Election Ecological Regression Precincts 1978 Primary SD 3 20 19 1982 Primary Sheriff 21 23 1982 Primary SD 3 5 6 1986 Primary Assessor 10 11 1988 Primary SD 5 1 2

377. These results show an extremely close correspondence between the estimates of non-Hispanic voting for the Hispanic candidate derived by ecological regression and the actual vote for the Hispanic candidate in precincts that are 90 percent or more non-Hispanic. This correspondence holds both for Spanish-surname and Spanish-origin data. In no instance is there a difference of more than two percentage points between the ecological regression results and the results from extreme case analysis. As would be expected from the fact that there are some Hispanics in the more than 90 percent non-Hispanic precincts, the support for the Hispanic candidate in these precincts is generally a point or two higher than the estimate drawn from ecological regression.

378. These results have implications for the estimates of Hispanic as well as non-Hispanic voting. The vote for the Hispanic candidate(s) is simply the sum of the votes cast for that candidate or candidates by non-Hispanic and by Hispanic voters. Thus, if non-Hispanics are not voting for the Hispanic candidate, then the votes for the Hispanic candidate must be coming from Hispanic voters. Therefore, the reliable ecological regression estimates of low non-Hispanic support for Hispanic candidates for each of the five elections studied also provides strong confirmation of the reliability of the ecological regression estimates for the Hispanic support for these candidates.

379. The analysis of both partisan and non-partisan elections also suggests that the degree of crossover voting by non-Hispanics for Hispanic candidates may decrease as the Hispanic component of a district decreases. Seven of the eight nonpartisan elections were held in districts, or Countywide, with an Hispanic component among their registered voters equal to or less than that of the most heavily Hispanic of existing Supervisorial districts, District 1. For these seven elections, a mean of only 9 percent of voters without a Spanish surname crossed over to support an Hispanic candidate. Of the nonpartisan primary elections, only the election for City Council in District 14 was held in a district with a greater Hispanic component than that of existing Supervisorial District 1. According to 1983 data, 49.9 percent of the registrants in Council District 14 has a Spanish surname. In the 1983 primary in this district, 23 percent of voters without a Spanish surname crossed over to opt for the Hispanic candidates, a percentage that is 2.5 times greater than the mean crossover vote of 9 percent for the remaining seven non-partisan contests.

380. Given the current configuration of the supervisorial districts and the existence of non-Hispanic bloc voting, the Hispanic electorate, though politically cohesive, would not normally have an opportunity to elect a candidate of their choice in even the most Hispanic districts, District 1 and District 3.

381. If the estimated polarization levels are applied to plaintiffs' proposed District 3 of Grofman Plan 1, the election prospects of Hispanics improve substantially.

382. This fact is illustrated by the table below which applies the cohesion and crossover estimates from the three 1982 primary elections in Congressional Districts 30 and 34 and Assembly District 59 to a 50.2 percent Spanish-origin district.

I. CD 30 (78% Hispanic Cohesion, 33% Non-Hispanic Crossover) 1. Hisp. Vote for Hisp. Candidate = .78 X 50.2% = 39.2% 2. Non-Hisp. Vote for Hisp. Candidate = .33 X 49.8% = 16.4% 3. Total Vote for Hisp. Candidate = 39.2% + 16.4% = 55.6%

II. CD 34 (88% Hispanic Cohesion, 26% Non-Hispanic Crossover) 1. Hisp. Vote for Hisp. Candidate = .88 X 48.3% = 42.5% 2. Non-Hisp. Vote for Hisp. Candidate = .26 X 51.7% = 13.4% 3. Total Vote for Hisp. Candidate = 42.5% + 13.4% = 55.9%

383. A similar analysis results in a projected vote of over 50 percent for an Hispanic-preferred candidate in District 3 of Grofman Plan 1 based on available 1982 Spanish-origin statistics (44.0 percent Spanish origin).

384. Under similar analyses, an Hispanic-preferred candidate would be the projected winner in a 44.0 percent Spanish-origin district under either the assumption of equal turnout rate or that of turnout differences between Hispanics and non-Hispanics.

G. OTHER SENATE FACTORS

1. History of Official Discrimination

385. The Hispanic community in Los Angeles County has borne the effects of a history of discrimination in the areas of education, housing, employment, and other socioeconomic areas.

386. In Southern California, restrictive real estate covenants have created limited housing opportunities for the Mexican-origin population. Dr. Camarillo testified that the current Hispanic population concentrations correspond to the historical process in which people were not allowed to live, or were restricted to particular areas of the County.

(a) Repatriation

387. From 1929 to 1939, in the aftermath of the Depression, some 200,000 to 300,000 Mexican-Americans returned to their "country of origin" as part of a program instituted by the Justice Department. While the program was theoretically voluntary, many legal resident aliens and American citizens of Mexican descent were forced or coerced out of the country.

(b) Education

388. In eight of the largest counties in California, in 1923, there were 64 schools with 90-100 percent Mexican-origin children. School officials required Mexican children to have separate graduation ceremonies from Anglos attending the same school. In one Los Angeles County school where officials were unable to provide separate buildings for the Mexican children, they were assigned to separate classrooms.

389. California maintained segregated schools for Hispanics in Los Angeles until 1947 when the California Supreme Court struck down such segregation. Mendez v. Westminster School District of Orange County, 64 F. Supp. 544 (S.D.Cal. 1946), aff'd, 161 F.2d 774 (9th Cir. 1947). However, as the United States points out, school desegregation litigation involving districts contained within Los Angeles County continued until 1989.

390. The mean years of school completed by Hispanic voting age citizens in 1980 was only 10.9 years, compared to 13.1 years for white non-Hispanics. The Hispanic mean was lower than that of any other minority group.

391. According to the Census Bureau's Current Population Survey, only 5 percent of Hispanics had completed 16 or more years of school, compared to 29 percent of Anglos.

(c) Public Facilities

392. As examples of discrimination against Hispanics in the use of public facilities, Dr. Camarillo testified that it was common during the first decade of this century, for access to public swimming pools to be restricted for Mexican-Americans and blacks, usually to the day before the pool was to be cleaned. In movie theaters, Mexican-Americans could not sit in the center sections.

(d) Right to Vote

393. In 1962, California was one of only 19 states which made English language literacy a prerequisite for voting.

394. In 1970, the California Supreme Court held that Article II, Section 1 of the Constitution of California violated the equal protection clause of the Fourteenth Amendment by conditioning the right of persons otherwise qualified to vote upon the ability to read the English language. The court found no compelling state interest in "denying the vote to a group of . . . citizens who already face similar problems of discrimination and exclusion in other areas and need a political voice if they are to have any realistic hope of ameliorating the conditions in which they live." Castro v. State of California, 2 Cal.3d 223, 240, 85 Cal.Rptr. 20, 466 P.2d 244 (1970). The court noted that "fear and hatred played a significant role in the passage of the literacy requirement." Id. 85 Cal.Rptr. at 25, 466 P.2d at 249.

395. Pursuant to the 1975 amendments to the Voting Rights Act, the Census Bureau determined that Los Angeles County was covered by the bilingual ballot election requirements of Section 203, 42 U.S.C. § 1973aa-1a, because more than five percent of the County's citizens of voting age were persons of Spanish heritage, a protected language minority group under the Voting Rights Act, and that the illiteracy rate of such persons was higher than the national rate.

396. As initially enacted, the provisions of Section 203 were due to expire on August 6, 1985.

397. In 1982, Congress extended the protections of Section 203 until August 6, 1992, but devised a new formula for coverage. This extension applied only to those jurisdictions in which the Census Bureau determined that members of a single language minority do not speak or understand English adequately enough to participate in the electoral process.

398. In 1984, the pursuant to the 1982 amendments, the Census determined that Los Angeles County was no longer covered by Section 203 of the Act. Although 14.6 percent of the County's voting age citizens were "persons of Spanish heritage" according to the 1980 Census, the Bureau concluded that fewer than five percent could not speak or understand English adequately enough to participate in the electoral process.

399. On August 7, 1984, the Board of Supervisors voted to discontinue providing election materials in Spanish.

2. Racial Appeals

400. The Garza plaintiffs provided the Court with substantial evidence of racial appeals in elections at all levels within the County.

401. For example, Steven Rodriguez, an Hispanic, ran for Councilman in District 14 of the Los Angeles City Council. When Mr. Rodriguez campaigned in Eagle Rock in 1983, he had doors slammed in his face and had his campaign literature destroyed. During his campaign, Mr. Rodriguez encountered such reaction in excess of 100 times.

402. During his campaigns for United States Congress, Esteban Torres encountered racial appeals by his opponents in the form of statements that Mr. Torres catered only to Hispanics and in the use of his photograph in opponents' campaign literature.

403. In the 1971 runoff for the 49th Assembly District, Richard Alatorre ran against William Brophy. Mr. Brophy distributed mailers which included Mr. Alatorre's photograph and alluded that Alatorre was sympathetic to undocumented aliens.

404. The Court finds that Hispanic residents in Los Angeles County have suffered and continue to suffer from the lingering effects of discrimination.

3. Size of Election Districts

405. While the population equality statistics for statewide electoral districts in California under the 1980 Census population figures range from 295,849 persons for state assembly districts to 525,953 persons for congressional districts, to 591,698 for state senate districts, population equality for a Los Angeles County Supervisorial district is 1,495,501 persons or approximately one sixteenth of the 1980 population of the State of California.

406. A Los Angeles County Supervisorial district equal to one-fifth of the County's population is over 2.5 times larger in population than either a congressional or state senatorial district which meet population equality standards and over 5 times as large as a California assembly district which satisfies the equal population standard.

407. The Los Angeles County Supervisorial districts have the largest population of any single-member district for electing a county governing body in the United States.

408. The 1980 population of each Los Angeles County Supervisorial district was larger than the population of 16 states.

409. Los Angeles County encompasses 4,083 square miles. In land area, the County is four times as large as the State of Rhode Island and twice as large as Delaware.

410. The five district structure clearly provides an advantage to incumbents and requires significant financial expenditures to run a successful campaign.

411. Between 1981 and 1986 incumbent Supervisors secured contributions of $8.2 million.

412. Candidates for the Board of Supervisors must raise more money than candidates for Governor in many states to be a serious challenger.

413. In 1962 and in 1976, the Board submitted the issue of revising the structure of County government to the voters. On both occasions, the voters rejected the proposed change.

414. The Garza plaintiffs contend that the size of the Board of Supervisors has a discriminatory impact upon Hispanic participation in the political process and that the size of the districts constitute a disfavored voting procedure that denies Hispanics equal access to the electoral process.

415. Supervisor Hahn testified that it was difficult for one Supervisor to represent more than a million people.

416. The Court finds that the enormous size and population of each supervisorial district and the fragmentation of the Hispanic population core under the 1981 redistricting plan have impeded the ability of Hispanic persons to participate in the political process, deterred viable Hispanic candidates from running for the Board, and impaired the ability of Hispanics to elect Supervisors of their choice.

To the extent that the preceding Findings of Fact may be deemed to be Conclusions of Law, they are hereby incorporated by reference into the Conclusions of Law.

III. CONCLUSIONS OF LAW

A. JURISDICTION

1. The Court has jurisdiction over this voting rights litigation pursuant to 42 U.S.C. § 1973 and 28 U.S.C. § 1331, 1343(a)(3) (4). Venue is proper in the Central District of California pursuant to 28 U.S.C. § 1391(b).

B. THE VOTING RIGHTS ACT

2. Section 2 of the Voting Rights Act, 42 U.S.C. § 1973, as amended, 96 Stat. 134, provides that:

(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 . . . to vote on account of race or color, or in contravention of the guarantees set forth in section 4(f)(2), as provided in subsection (b).

(b) A violation of subsection (a) of this section is established if, based on the totality of the 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) in that its members have less opportunity than other political 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 [but] 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.

3. Section 4(f)(2) of the Act provides

No voting qualification or prerequisite to voting, or standard, practice or procedure shall be imposed or applied by any State or political subdivision . . . to deny or abridge the right of any citizen of the United States to vote because he is a member of a language minority group. (emphasis added)

1. The Senate Factors

4. The Senate Judiciary Committee majority report accompanying the bill that amended § 2, elaborates on the circumstances that might be probative of a § 2 violation, noting the following "typical factors" (hereinafter "Senate Factors"):

1. The extent of any history of official discrimination in the state or political subdivision that touched the right of the members of the minority group to register, to vote, or otherwise to participate in the democratic process;

2. The extent to which voting in the elections of the state or political subdivision is racially polarized;

3. The extent to which the state or political subdivision has used unusually large election districts, majority vote requirements, anti-single shot provisions, or other voting practices or procedures that may enhance the opportunity for discrimination against the minority group;

4. If there is a candidate slating process, whether the members of the minority group have been denied access to that process;

5. The extent to which members of the minority group in the state or political subdivision bear the effects of discrimination in such areas as education, employment and health, which hinder their ability to participate effectively in the political process;

6. Whether political campaigns have been characterized by overt or subtle racial appeals.

7. The extent to which members of the minority group have been elected to public office in the jurisdiction.

5. Additional factors considered probative of a violation included:

Whether there is a significant lack of responsiveness on the part of elected officials to the particularized needs of the members of the minority group.

Whether the policy underlying the state or political subdivision's use of such voting qualification, prerequisite to voting, or standard, practice or procedure is tenuous.

S.Rep. No. 97-417, 97th Cong.2d Sess. 28, 29 (1982), U.S. Code Cong. Admin. News 1982, pp. 206-207 (hereinafter S.Rep.).

6. The impact of the contested structure or practice on minority electoral opportunities must be assessed based on "objective" factors which include but are not limited to the Senate Factors enumerated above. The Senate Committee noted in its report that there is no requirement that any particular number of factors be proved, or that a majority of them point one way or another. S.Rep. at 29, U.S. Code Cong. Admin. News 1982, p. 207.

7. The Senate Committee set forth a flexible, fact-intensive test for determining § 2 violations. "The question whether the political processes are `equally open' depends upon a searching practical evaluation of the `past and present reality'" and on a "functional" view of the political process. Gingles, 478 U.S. at 45, 106 S.Ct. at 2763 citing S.Rep. at 30, n. 120, U.S. Code Cong. Admin. News 1982, p. 208. As Gingles explained, "the essence of a § 2 claim is that a certain electoral law, practice or structure interacts with social or historical conditions to cause an inequality in opportunities enjoyed by black and white voters to elect their preferred representatives." Id. at 47, 106 S.Ct. at 2764. Rights afforded under Section 2 apply equally to Hispanics. Gomez v. City of Watsonville, 863 F.2d 1407 (9th Cir. 1988), cert. denied, 489 U.S. 1080, 109 S.Ct. 1534, 103 L.Ed.2d 839 (1989). The conclusion as to whether Hispanics have an equal opportunity to participate in the political process and to elect candidates of their choice is "peculiarly dependent upon the facts of each case." Id. 478 U.S. at 79, 106 S.Ct. at 2781.

8. The circumstances under which § 2 violations may be proved is limited in three ways:

First, electoral devices, such as at large elections, may not be considered per se violative of Section 2. Plaintiffs have the burden of demonstrating that, under the totality of the circumstances, the devices result in unequal access to the electoral process.

Second, the conjunction of an allegedly dilutive electoral mechanism and the lack of proportional representation does not establish a violation.

Third, the results test does not assume the existence of racial bloc voting; plaintiffs must prove it.

9. The Supreme Court in Gingles, 478 U.S. at 46, 106 S.Ct. at 2764, addressed a claim that multimember districts diluted black voting strength. Plaintiffs alleged and attempted to prove that their ability to elect the representatives of their choice was impaired by the selection of a multimember electoral structure. Id. at 46 n. 12, 106 S.Ct. at 2764 n. 12. The Supreme Court stated that it had no occasion to consider what standards should pertain to a claim brought by a minority group that is not sufficiently large and compact to constitute a majority in a single-member district, alleging that the use of a multimember district impairs its ability to influence elections. Id. (emphasis in the original).

10. The Court also stated that it had no occasion to consider whether the standards applied in Gingles are fully pertinent to other sorts of vote dilution claims, such as claims alleging that the splitting of a large and geographically cohesive minority between two or more multimember or singlemember districts resulted in the dilution of the minority vote. Id. at n. 12.

11. While many or all of the Senate Factors may be relevant to a plaintiff's § 2 claim, "unless there is a conjunction of the following circumstances, the use of multimember districts generally will not impede the ability of minority voters to elect representatives of their choice." Id. at 48, 106 S.Ct. at 2765. Specifically, the Court outlines three preconditions for multimember districts to operate to impair minority voters' ability to elect representatives of their choice:

First, the minority group must be able to demonstrate that it is sufficiently large and geographically compact to constitute a majority in a single member district. (hereinafter "geographical compactness")

Second, the minority group must be able to show that it is politically cohesive. (hereinafter "political cohesiveness")

Third, the minority must be able to demonstrate 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. (hereinafter "racial bloc voting")

(a) Geographical Compactness

12. Unless minority voters possess the potential to elect representatives in the absence of the challenged structure or practice, they cannot claim to have been injured by the structure or practice. Id. at 50 n. 17, 106 S.Ct. at 2766 n. 17 (emphasis in original). For this reason, the Supreme Court determined that a showing of geographic compactness is a threshold matter. Id.

13. Justice O'Connor, in a concurring opinion joined by Chief Justice Burger, Justice Powell, and Justice Rehnquist, preferred to leave open the broader question of whether § 2 requires a showing of maximum feasible minority voting strength:

In my view, we should refrain from deciding in this case whether a court must invariably posit as its measure of "undiluted" minority voting strength single-member districts in which minority group members constitute a majority. There is substantial doubt that Congress intended "undiluted minority voting strength" to mean "maximum feasible minority voting strength." Even if that is the appropriate definition in some circumstances, there is no indication that Congress intended to mandate a single, universally applicable standard for measuring undiluted minority voting strength, regardless of local conditions and regardless of the extent of past discrimination against minority voters in a particular State or political subdivision.

(1) Voting Age Population

14. The eligible minority voter population, rather than the total population is the appropriate measure of geographical compactness. Romero, 883 F.2d at 1426; Gomez, 863 F.2d at 1414; Skorepa v. City of Chula Vista, 723 F. Supp. 1384, 1386 (S.D.Cal. 1989).

(2) Current Population Data

15. Current voting age population data are probative because they indicate the electoral potential of the minority community. City of Rome v. United States, 446 U.S. 156, 186 n. 22, 100 S.Ct. 1548, 1566 n. 22, 64 L.Ed.2d 119 (1980). See, e.g., Gingles, 478 U.S. at 80, 106 S.Ct. at 2781 (results of elections for years 1978, 1980 and 1982 examined to determine if racially polarized voting existed); Gomez, 863 F.2d at 1409-10 n. 1 (election results from 1971 through 1987 considered); Smith v. Clinton, 687 F. Supp. 1310, 1315-16 (E.D. Ark.) (three-judge court), aff'd, 488 U.S. 988, 109 S.Ct. 548, 102 L.Ed.2d 576 (1988) (election results analyzed include 1982, 1985, 1986 and 1988 contests).

16. The census is presumed to be accurate unless proven otherwise. Latino Political Action Committee v. City of Boston, 568 F. Supp. 1012, 1018 (D.Mass. 1983), aff'd, 784 F.2d 409 (1st Cir. 1986). The evidence disproving the census must be clear, cogent and convincing. Dixon v. Hassler, 412 F. Supp. 1036, 1040 (W.D.Tenn. 1976) (three judge panel), aff'd sub nom. Republican Party v. Dixon, 429 U.S. 934, 97 S.Ct. 346, 50 L.Ed.2d 303 (1976) (applying standard that decennial census will be controlling unless there is `clear, cogent and convincing evidence' that such figures are no longer valid and that other figures are valid).

17. In order to overcome the presumption in favor of the 1980 census data, plaintiffs need not demonstrate that the census was inaccurate.

18. It is sufficient to conclude that there has been significant demographic changes since the decennial census and that there exists post-decennial population data that more accurately reflects evidence of the current demographic conditions. Kirkpatrick v. Preisler, 394 U.S. 526, 535, 89 S.Ct. 1225, 1231, 22 L.Ed.2d 519 (1969); cf. Gaffney v. Cummings, 412 U.S. 735, 746, 93 S.Ct. 2321, 2328, 37 L.Ed.2d 298 (1973) (describing federal census as "more of an event than a process" measuring population "at only a single instant in time").

(3) Estimates and Projections

19. Where shifts in population can be predicted with a high degree of accuracy, such "predictions" may be considered by states that are redistricting. Kirkpatrick, 394 U.S. at 534, 89 S.Ct. at 1230. These findings as to population trends must be thoroughly documented and applied throughout the state in a systematic manner. Id. but Cf. McNeil v. Springfield Park District, 851 F.2d 937, 947 (7th Cir. 1988), cert. denied, 490 U.S. 1031, 109 S.Ct. 1769, 104 L.Ed.2d 204 (1989) (refusing to override presumption in favor of census based on meager evidence and noting that estimates based on past trends are generally not sufficient to override "hard" decennial census data); Graves v. Barnes, 446 F. Supp. 560, 568 (W.D.Texas 1977), aff'd sub nom. Briscoe v. Escalante, 435 U.S. 901, 98 S.Ct. 1444, 55 L.Ed.2d 492 (1978) (study's projections did not offer high degree of accuracy required to supplant population figures of prior decennial census).

(b) Political Cohesiveness

20. The inquiry whether a minority group is politically cohesive is not to be made prior to and apart from a study of polarized voting because the central focus is upon voting patterns. Campos v. City of Baytown, 840 F.2d 1240, 1244 (5th Cir. 1988), cert. denied, ___ U.S. ___, 109 S.Ct. 3213, 106 L.Ed.2d 564 (1989). If a minority group votes together it can be deemed politically cohesive. Id.

21. In determining political cohesiveness, the inquiry is essentially whether the minority group has expressed clear political preferences that are distinct from those of the majority. Gomez, 863 F.2d at 1415. Therefore, as the Court noted in Gingles, one way to demonstrate cohesiveness is by showing that a significant number of minority group members usually vote for the same candidates. Gingles, 478 U.S. at 56, 106 S.Ct. at 2769.

22. In Gomez, the Ninth Circuit reversed the district court for applying an incorrect legal standard. The district court had determined that, " with respect to those Hispanics who have actually voted, the evidence favored a finding of political cohesiveness." Id. at 1416 (emphasis in original). The court concluded, however, that because "no significant number of eligible Hispanics have voted in the elections under consideration," the Hispanic community as a whole was too apathetic to be politically cohesive. Id.

23. Political cohesiveness is to be judged primarily on the basis of the voting preferences expressed in actual elections. Gomez, 863 F.2d at 1416. "The district court erred by focusing on low minority voter registration and turnout as evidence that the minority community was not politically cohesive. The court should have looked only to actual voting patterns rather than speculating as to the reasons why many Hispanics were apathetic." Id.

24. Socioeconomic disparities and differences of political opinion within the Hispanic community are "only relevant to the extent that they reflect differences in voting behavior among Hispanics." Id.

25. Statistical analysis of voting data is highly relevant to the issue of political cohesion. Sanchez v. Bond, 875 F.2d 1488, 1493 (10th Cir. 1989).

(1) Ecological Regression Analysis

26. Political cohesion may be established through ecological regression analysis and lay witness testimony. Carrollton Branch of NAACP v. Stallings, 829 F.2d 1547, 1558 (11th Cir. 1987), cert. denied, 485 U.S. 936, 108 S.Ct. 1111, 99 L.Ed.2d 272 (1988). As the Ninth Circuit stated in Romero, 883 F.2d at 1423, "Both before and after Thornburg, plaintiffs, including plaintiffs in this case, utilized exit polls, ecological regression and homogeneous precinct analysis to show the existence of polarized voting."

27. Bivariate ecological regression analysis has been frequently employed in Section 2 cases after Gingles. See, e.g., Campos, 840 F.2d at 1246-48; Citizens for a Better Gretna v. Gretna, 834 F.2d 496, 500-02 (5th Cir. 1987).

28. Crucial to the validity of regression analysis are the values for "R" and "R2", which measure the strength of the correlation and linear relationship of the variables being examined. Overton v. Austin, 871 F.2d 529, 539 (5th Cir. 1989) (stating that "R2" value expresses the percentage of variance in the vote that is explained by the race of the voters).

(c) Racial Bloc Voting

29. "The number of elections that must be studied in order to determine whether voting is polarized will vary according to pertinent circumstances. One important circumstance is the number of elections in which the minority group has sponsored candidates." Campos, 840 F.2d at 1245 (finding district court warranted in its focus on those races that had a minority member as a candidate).

30. If a small number of minority candidacies prevents the compilation of statistical evidence, a court should not deny relief, but should rely on other totality of circumstances factors to determine if the electoral system had a discriminatory effect. See Solomon v. Liberty County, 865 F.2d 1566, 1577-78 (11th Cir. 1988) (holding that plaintiffs should be able to buttress their claims of white bloc voting by pointing to racial voting patterns in elections for offices they do not challenge in their section 2 suit and that district court erred in ignoring regression analyses considered probative of black political cohesiveness).

31. In a plurality portion of the Gingles opinion, the Court stated that "[u]nder § 2, it is the status of the candidate as the chosen representative of a particular racial group, not the race of the candidate, that is important." Gingles, 478 U.S. at 68, 106 S.Ct. at 2775. The race of the voter, not of the candidate is relevant to vote dilution analysis. Id. However, the Court also recognized that since both minority and majority voters often select members of their own race as their preferred representatives, "it will frequently be the case that a black candidate is the choice of blacks, while a white candidate is the choice of whites." Id.

32. The Fifth Circuit interpreted Gingles to hold that the race of the candidate is in general of less significance than the race of the voter — but only within the context of an election that offers voters the choice of supporting a viable minority candidate. Better Gretna, 834 F.2d at 503 (emphasis added).

33. The legal concept of racially polarized voting, as it relates to claims of vote dilution, refers only to the existence of a correlation between the race of the voters and the selection of certain candidates. Id. 478 U.S. at 74, 106 S.Ct. at 2778.

34. In order to prove a prima facie case of racial bloc voting, plaintiffs need not prove causation or intent. Id.

35. A definition or racially polarized voting which holds that racial bloc voting does not exist when voters of a certain race's choice of a certain candidate is most strongly influenced by the fact that the voters have low income and menial jobs — "when the reason most of those voters have menial jobs and low incomes is attributable to past or present racial discrimination — runs counter to the Senate Report's instruction to conduct a searching and practical evaluation of past and present reality." Id. at 65, 106 S.Ct. at 2774 citing S.Rep. at 30. Such an approach, according to the Supreme Court, would interfere with the purpose of the Voting Rights Act to eliminate the negative effects of past discrimination on the electoral opportunities of minorities. Id.

36. The fact that racially polarized voting is not present in one or a few individual elections does not necessarily negate the conclusion that the district experiences legally significant bloc voting. Gingles, 478 U.S. at 51, 106 S.Ct. at 2766.

(d) History of Discrimination

37. Congress intended that the Voting Rights Act eradicate inequalities in political opportunities that exist due to the vestigial effects of past purposeful discrimination. Gingles, 478 U.S. at 69, 106 S.Ct. at 2776. In Solomon, 865 F.2d at 1579, the Eleventh Circuit found that the trial court erred by failing to consider past and present reality as required by Gingles and by refusing to give any weight to the legislature's reason — to discriminate against blacks — for prescribing the at-large system as the method of electing school board members in Florida.

38. Courts have historically recognized that political participation by minorities tends to be depressed where minority groups suffer effects of prior discrimination such as inferior education, poor employment opportunities and low incomes. Gingles, 478 U.S. at 69, 106 S.Ct. at 2776; see, e.g., White v. Regester, 412 U.S. 755, 768-69, 93 S.Ct. 2332, 2340-41, 37 L.Ed.2d 314 (1973) (holding that district court's order requiring disestablishment of multimembers districts in certain Texas counties was warranted in light of history of political discrimination against blacks and Mexican-Americans residing in those counties and the residual effects of such discrimination on those groups); Kirksey v. Board of Supervisors of Hinds County, 554 F.2d 139, 145-46 (5th Cir.) (en banc) cert. denied 434 U.S. 968, 98 S.Ct. 512, 54 L.Ed.2d 454 (1977) (finding that Supervisors' reapportionment plan, though racially neutral, would perpetuate the denial of black minority access to the democratic process).

39. In Kirksey, 554 F.2d at 151, the Fifth Circuit, in reversing the district court's reapportionment plan, concluded that plaintiffs had proved a long history of denial of access to the democratic process and that the structure and residual effects of the past had not been removed and replaced by current access. "By fragmenting a geographically concentrated but substantial black minority in a community where bloc voting has been a way of political life the plan [though racially neutral] will cancel or minimize the voting strength of the black minority and will tend to submerge the interests of the black community." Id. The court concluded that the plan denies rights protected by the Fourteenth and Fifteenth Amendments.

(e) Other Discriminatory Voting Practices

40. A section 2 claim is enhanced by a showing of the existence of large districts, majority voting requirements, anti-single shot voting provisions and the lack of provision for at-large candidates running from particular geographic subdistricts. Zimmer v. McKeithen, 485 F.2d 1297, 1305 (5th Cir. 1973) (en banc), aff'd sub nom. East Carroll Parish School Board v. Marshall, 424 U.S. 636, 96 S.Ct. 1083, 47 L.Ed.2d 296 (1976).

(f) Size of Election Districts

41. Unusually large election districts is a factor typically relevant to a Section 2 claim. Gingles, 478 U.S. at 45, 106 S.Ct. at 2763.

(g) Candidate Slating Process

42. A slating process is a procedure by which a political group determines what candidate they will sponsor for particular offices. The resulting candidacies comprise that group's "slate." Solomon, 865 F.2d at 1581 n. 31 (finding that on remand district court should consider whether white slating process is open to black candidates who seek to represent black interests). Slating could thus operate to control effective access of minorities to the ballot. Overton, 871 F.2d at 534.

(h) Lingering Effects of Past Discrimination

43. The lingering effects of past discrimination are relevant only if they continue to "hinder [the minority group's] ability to participate effectively in the political process." S.Rep. at 29, U.S. Code Cong. Admin. News 1982, p. 206.

(i) Election of Minorities

44. Minority electoral failure is one of the two most probative indications of vote dilution. Solomon, 865 F.2d at 1583 citing Gingles, 478 U.S. at 48 n. 15, 106 S.Ct. at 2765 n. 15.

C. DISCRIMINATORY RESULTS V. INTENT

45. In Mobile v. Bolden, 446 U.S. 55, 66, 100 S.Ct. 1490, 1499, 64 L.Ed.2d 47 (1980), the Supreme Court determined that minority voters, to establish that their votes have been diluted in violation of section 2 of the Voting Rights Act (hereinafter "the Act"), as well as violation of the Fourteenth and Fifteenth Amendments to the Constitution, must prove that the contested electoral practice was adopted or maintained by the governmental officials for a discriminatory purpose.

46. In 1982, section 2 of the Act was amended to add a "results" test to the intent test. As the Supreme Court stated in Gingles, 478 U.S. at 43, 106 S.Ct. at 2762, the intent test was repudiated because it asked the wrong question. The "right" question is whether "as a result of the challenged practice or structure plaintiffs do not have an equal opportunity to participate in the political processes and to elect candidates of their choice." S.Rep. at 28, U.S. Code Cong. Admin. News 1982, p. 206. The Report of the Senate Committee states in pertinent part:

The subsection [new subsection 2(a)] expresses the intent of Congress in amending Section 2 that plaintiffs do not need to prove discriminatory purpose or motive, by either direct or indirect evidence, in order to establish a violation. With this clarification, Section 2 explicitly codifies a standard different from the interpretation of the former language of Section 2 contained in the Supreme Court's Mobile plurality opinion, i.e. the interpretation that the former language of Section 2 prohibits only purposeful discrimination.

Under Section 2, as amended plaintiffs would continue to have the option of establishing a Section 2 violation by proving a discriminatory purpose behind the challenged practice or method. However, if plaintiff chose to establish a violation under the alternative basis now codified in the statute as the "results" standard, then proof of the purpose behind the challenged practice is neither required or relevant. . . . The courts are to look at the totality of the circumstances in order to determine whether the result of the challenged practice is that the political processes are equally open; that is, whether, members of a protected class have the same opportunity as others to participate in the electoral process and to elect candidates of their choice. The courts are to conduct this analysis on the basis of a variety of objective factors concerning the impact of the challenged practice and the social and political context in which it occurs.

The motivation behind the challenged practice or method is not relevant to the determination. The [Senate] Committee expressly disavows any characterization of the results test codified in this statute as including an intent requirement, whether or not such a requirement might be met in a particular case by inferences drawn from the same objective factors offered to establish a discriminatory result. Nor is there any need to establish a purposeful design through inferences from the foreseeable consequences of adopting or maintaining the challenged practice.

S.Rep. at 67, 68, U.S. Code Cong. Admin. News 1982, pp. 245-46. The Court finds that the claims that a challenged electoral system or practice violates Section 2 due to a discriminatory purpose may be determined independently of any analysis of the preconditions set forth in Gingles. See Brown v. Board of Commissioners of City of Chattanooga, 722 F. Supp. 380, 383 (E.D.Tenn. 1989) (stating that in adding the "results" test to Section 2 of the Voting Rights Act, Congress left the "intent" test intact); cf. Overton, 871 F.2d at 540-41 (explaining that the factors pertinent to a determination of discriminatory intent of a regulation that continues to have disparate racial impact include the historical background of the regulation, specific sequence of events leading up to the regulation, departures from the normal procedural sequence, substantive departures, and legislative history, especially where there are contemporary statements by members of the decision-making body).

47. The standard of proof required for determining intent or discriminatory purpose is the same as that used in resolving cases under the Fourteenth Amendment's Equal Protection Clause. Rogers v. Lodge, 458 U.S. 613, 617, 102 S.Ct. 3272, 3275, 73 L.Ed.2d 1012 (1982); Arlington Heights v. Metropolitan Housing Corp., 429 U.S. 252, 265, 97 S.Ct. 555, 563, 50 L.Ed.2d 450 (1977).

48. Discriminatory purpose may be inferred from the totality of the relevant facts, including the fact that the law bears more heavily on one race than another. Washington v. Davis, 426 U.S. 229, 240, 96 S.Ct. 2040, 2047, 48 L.Ed.2d 597 (1976).

49. Determining whether invidious discriminatory purpose was a motivating factor demands a sensitive inquiry into such circumstantial and direct evidence of intent as may be available. Id. at 242, 96 S.Ct. at 2048.

50. Courts traditionally refrain from reviewing the merits of the decisions of legislators and administrators on the grounds that these officials are properly concerned with balancing numerous competing considerations. However, racial discrimination is not just another competing consideration. Arlington Heights, 429 U.S. at 265, 97 S.Ct. at 563. When there is a proof that a discriminatory purpose has been a motivating factor in the decision, this judicial deference is no longer justified. Id.

51. Historical evidence is relevant to a determination of discriminatory purpose. Rogers, 458 U.S. at 625, 102 S.Ct. at 3279. See Brown, 722 F. Supp. at 385 (finding history of Chattanooga's city government and the black franchise "particularly revealing").

52. Factors that may be probative of a discriminatory purpose include: (1) impact of the official action; (2) historical background of the decision, "particularly if it reveals a series of official actions taken for invidious purposes"; (3) specific sequence of events leading up to the challenged decision; (4) departures from normal procedural sequences; (5) substantive departures . . . "particularly if the factors usually considered important by the decision maker strongly favor a decision contrary to the one reached." Arlington Heights, 429 U.S. at 266-67, 97 S.Ct. at 564.

53. In Rybicki v. State Board of Elections, 574 F. Supp. 1082, 1109 (N.D.Ill. 1982), the court found that where the requirements of incumbency "were so closely intertwined with the need for racial dilution that an intent to maintain a safe, primarily white, district for Senator Joyce is virtually coterminous with a purpose to practice racial discrimination," is indicative of an intent to discriminate.

D. INTER-DECENNIAL REDISTRICTING

54. The California Election Code states in pertinent part:

At any time between the decennial adjustments of district boundaries, the board may cause a census of the County to be taken as provided in Section 26203 of the Government Code, and may adjust the boundaries of the Supervisorial districts on the basis of that census, or on the basis of population estimates prepared by the State Department of Finance or the County planning department or planning commission, pursuant to section 35000.

Cal. Elec. Code § 35003, added by Stats. 1979, c. 546, p. 1747, § 1. Pursuant to California Election Code § 35003 (West 1989), the County is authorized to conduct inter-decennial apportionments.

E. TOTAL POPULATION AS APPORTIONMENT BASE

55. The law of the State of California requires that the Board of Supervisors redistrict using total population figures validated by the California Department of Finance. California Election Code § 35000 states in pertinent part:

Following each decennial federal census, and using population figures as validated by the Population Research Unit of the Department of Finance as a basis, the board shall adjust the boundaries of any or all of the Supervisorial districts of the County so that the districts shall be as nearly equal in population as may be.

Cal. Elec. Code § 35000.

56. Neither the Constitution of the State of California nor the United States Constitution requires the use of citizens or citizens of voting age as the apportionment base. Burns v. Richardson, 384 U.S. 73, 92, 86 S.Ct. 1286, 1296, 16 L.Ed.2d 376 (1966). Nor are states required to include "aliens, transients, short-term or temporary residents, or persons denied the vote for conviction of crime, in the apportionment base by which their legislators are distributed and against which compliance with the Equal Protection Clause is to be measured." Id. at 92, 86 S.Ct. at 1296. As the Supreme Court explained, this decision on which groups to include or exclude "involves choices about the nature of representation with which we have been shown no constitutionally founded reason to interfere." Id.

57. In Burns, the Supreme Court found fault with the use of a registered voter or actual voter base since such a basis depends upon the extent of political activity of those eligible to register and vote as well as upon criteria governing state citizenship. Id. "Each is susceptible to improper influences by which those in political power might be able to perpetuate underrepresentation of groups constitutionally entitled to participate in the electoral process, or perpetuate a `ghost of prior malapportionment.'" Id. at 92-93, 86 S.Ct. at 1297, quoting Buckley v. Hoff, 243 F. Supp. 873, 876 (D.C.Vt. 1965).

F. ONE PERSON ONE VOTE RULE

58. The overriding objective of a legislative apportionment scheme must be "substantial equality of population among the various districts, so that the vote of any citizen is approximately equal in weight to that of any other citizen in the State." Reynolds v. Sims, 377 U.S. 533, 579, 84 S.Ct. 1362, 1390, 12 L.Ed.2d 506 (1964).

59. The right of American citizens to participate fully and effectively in the political processes of state legislative bodies applies equally to County bodies. See Avery v. Midland County, 390 U.S. 474, 480, 88 S.Ct. 1114, 1118, 20 L.Ed.2d 45 (1968) (finding that city, town, or County may no more deny equal protection than it may abridge freedom of speech, establish an official religion, arrest without probable cause, or deny due process of law).

60. While an alternative election system must comport with the one person one vote standard, it need not achieve absolute equality. Reynolds, 377 U.S. at 578, 84 S.Ct. at 1390. The Supreme Court has acknowledged that some leeway in the equal protection requirement should be afforded states in devising their legislative apportionment plans. A maximum deviation from population equality of less than ten percent is permissible under the equal protection clause for purposes of apportioning state and local governing bodies. See, e.g., Brown v. Thomson, 462 U.S. 835, 852, 103 S.Ct. 2690, 2701, 77 L.Ed.2d 214 (1983) (stating that deviations below ten percent are ordinarily considered de minimis); Connor v. Finch, 431 U.S. 407, 418, 97 S.Ct. 1828, 1835, 52 L.Ed.2d 465 (1977) (noting that under-10 percent deviations are considered to be of prima facie constitutional validity in context of legislatively enacted apportionments); see also White, 412 U.S. at 765, 93 S.Ct. at 2339 (permitting variance of 9.9 percent); Gaffney, 412 U.S. at 745, 93 S.Ct. at 2327 (permitting deviation of 7.83 percent with no showing of invidious discrimination).

61. The burden is on the district court to "elucidate the reasons necessitating any departure from the goal of population equality, and to articulate clearly the relationship between the variance and the state policy furthered." Chapman v. Meier, 420 U.S. 1, 24, 95 S.Ct. 751, 764, 42 L.Ed.2d 766 (finding that 20 percent variance in plan formulated by federal court is constitutionally impermissible absent significant state policies or other acceptable considerations that require adoption of a plan with so great a variance).

G. REAPPORTIONMENT

62. The task of reapportionment is properly a legislative function. Whenever practicable, the legislature should be afforded a reasonable opportunity to meet constitutional requirements by adopting a substitute measure rather than for the federal court to devise and order into effect its own plan. Wise v. Lipscomb, 437 U.S. 535, 540, 98 S.Ct. 2493, 2497, 57 L.Ed.2d 411 (1978). The County may also provide to this Court an appropriate schedule for the prompt implementation of the plan following the Court's review.

63. Should the County be unable or unwilling to devise and present a fair election plan to this Court, the Court will undertake the "unwelcome obligation" of ordering into effect a plan of its own design. Connor, 431 U.S. at 415, 97 S.Ct. at 1834.

To the extent that the preceding Conclusions of Law may be deemed to be Findings of Fact, they are hereby incorporated by reference into the Findings of Fact.

IT IS SO ORDERED.