Full opinion text
DECISION AND ORDER SKRETNY, District Judge INTRODUCTION This case arises under the Voting Rights Act of 1965 (“the Act”), 42 U.S.C. § 1973. Plaintiffs are challenging the existing method of electing members to the Niagara Falls City Council, pursuant to Section 2 of the Act, 42 U.S.C. § 1973. Plaintiffs include the National Association for the Advancement of Colored People (“NAACP”), its local Political Action Chairperson Renae Kimble, and nineteen other registered African American voters of the City of Niagara Falls, New York. Defendants include the City of Niagara Falls and its Mayor, the present members of the Niagara Falls City Council, and the Niagara Falls City Clerk. Under the present method of electing members to the City Council, a method that was approved by a city-wide referendum in 1985, the seven members of the City Council are elected at-large. Plaintiffs object to this system. They demand that this Court enter a declaratory judgment that the present method violates Section 2 of the Act because it results in a denial or abridgement of their right to vote on account of their race. Plaintiffs believe that denial or abridgement is established by a showing that, under the totality of circumstances, the political processes leading to nomination or election to the City Council are not equally open to participation by African Americans in the City of Niagara Falls in that African Americans have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice. Plaintiffs demand that this Court award them injunctive relief by ordering the implementation of a single-member district method of electing members to the City Council, and by requiring that any dis-tricting plan include an African American majority-minority district. A non-jury trial was held before this Court from October 5, 1993, to November 3, 1993. During their case-in-chief, plaintiffs offered testimony from four lay witnesses (plaintiffs Bond, Palmore, Brown, and Kimble), an expert in the area of demographics (Jerry Wilson), an expert historian (Lillian S. Williams), and an expert who testified regarding racial voting patterns and turnout (Michael McDonald). Defendants offered the testimony of ten lay witnesses, including defendant Quaranto and a number of other city officials, and their own expert, who testified regarding voting behavior, elections, and racial voting patterns (Harold W. Stanley). In rebuttal, plaintiffs offered the testimony of an additional expert witness (James W. Loewen). By Order of this Court, Loewen’s rebuttal testimony was limited to the issues of (1) Stanley’s prediction of plaintiff Kimble’s hypothetical success in the 1987 City Council general election; (2) the impact of staggered elections to the City Council on African American voters’ ability to elect a representative of their choice; and (3) voter “rolloff’ in the 1985 referendum that determined the form of Niagara City government. Lowen’s testimony was limited to these issues in light of the evidence presented by the parties during their eases-in-chief, and the probative value of the testimony that plaintiffs expected to elicit from Loewen on rebuttal. The parties agree that plaintiffs have established the first two of the three prongs necessary to prove a prima facie case under Section 2 of the Act, pursuant to Thornburg v. Gingles, 478 U.S. 30, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986). Therefore, the only issue regarding plaintiffs’ prima facie case is whether plaintiffs have established the third prong, i.e., whether “the white majority votes sufficiently as a bloc to enable it — in the absence of special circumstances ... — usually to defeat the minority’s preferred candidate.” Id. at 50-51, 106 S.Ct. at 2766-67. This matter is now before this Court for a final decision on the merits. On the basis of the findings of fact and conclusions of law set forth below, this Court finds in favor of defendants on plaintiffs’ claim under Section 2 of the Voting Rights Act. FINDINGS OF FACT A. POLITICAL HISTORY OF THE CITY OF NIAGARA FALLS 1. Niagara Falls came into being pursuant to an Act of the New York State Legislature on March 17,1892, with the union of the Villages of Suspension Bridge and Niagara Falls. 2. Originally, the City Charter provided for a mayor/aldermanic form of government, with the aldermen elected from wards within the city. In 1892,- this form provided for eight aldermen, with two elected from each of four wards: The aldermen in each ward served staggered, two-year terms. The Mayor was elected .at-large for a one-year term. 3. In 1914, the New York State Legislature enacted the “Optional City Government Law,” permitting certain New York cities the option of adopting one of seven different forms of municipal government. 4. In 1915, a referendum was held under which the voters opted for a council/manager plan. Under this plan the Mayor and four councilpersons were elected at-large for four year terms. Terms of office for the Council were staggered, with two councilpersons elected every two years. These five officials served part-time, and comprised the City Council. The Mayor voted equally in Council matters. 5. These five officials appointed a City Manager who ran the day-to-day affairs of the city. 6. This plan of government was in effect until January 1, 1988, a period of 73 years. 7. In 1984 the City created a City Charter Revision Commission composed of seven city residents, including one African American, with instructions to study City government and recommend any changes it deemed appropriate. As a result of this study, and after public hearings, a referendum was held at the general election in November 1985, at which time the electorate was given the option of selecting: a. A mayor-council form of government with the mayor as the chief executive officer and a seven member council; b. If a mayor-council form were adopted, whether the seven members of the city council would be elected from seven separate single-member eouneilmanic districts or at-large; c. A manager-council form of government with the city manager as the chief executive officer and a six member city council; d. If a manager-council form were adopted, whether the six members of the city council would be elected from six separate single-member eouneilmanic districts or at-large. 8. On November 5, 1985 the voters of Niagara Falls voted in favor of a “strong mayor” plan with the seven members of the city council to be elected at-large. This Charter amendment became effective on January 1, 1988. The expanded council and new mayor were elected at the November 1987 general election. 9. Under the present “strong mayor” form of government, the Mayor is the chief executive officer of the City, with veto power over the enactments of a separate legislative branch: the seven member City Council. The day-to-day operations of the City are under the direction of an Administrator, who is appointed by the Mayor. Both the Mayor and Council members are elected to four year terms, with the Council members serving staggered terms. Couneilmanic elections are held every two years. B. CANDIDATE SLATING 1. There are five recognized political parties in the State of New York: a. Democrat b. Republican c. Conservative d. Right-to-Life e. Liberal 2. A party must receive at least 50,000 votes statewide in the last gubernatorial election to remain on the ballot as a recognized party. 3. The party receiving the greatest number of votes statewide in a gubernatorial election is awarded the top line on the ballot. The party receiving the next highest number of votes receives the second line, etc. 4. The two top lines determine who is seated as commissioner in the County Boards of Elections, and which party nominates the election inspectors who are in the polling places. 5. The New York State Election Law provides the method by which a person qualifies as a primary candidate. C. ELECTION OF CITY COUNCILPERSONS 1. Under the New York State Election Law, if enrolled in a party, except for judicial positions, a candidate may circulate petitions during times designated by the New York State Legislature. A legislative “political calendar” is voted on and established each year by the New York State Legislature. 2. The City Clerk notifies the County Election Commission which seats on the City Council are up for election. 3. A candidate must obtain the signatures of 5% of enrolled voters in the City, established by the number of voters registered as of April 1st each year. 4. The candidate must file the signatures with the County Board of Elections. 5. If not objected to, the candidate is listed on the ballot. 6. If more than one candidate files on a party line, that party holds a primary election, which is usually held in September. The top vote-getter goes on to the general election, which is usually held in November. The top vote-getters, consistent with the number of seats up for election, are declared winners. 7. A voter may vote only in the primary for the party in which he or she is enrolled. However, in the general election voters may vote for any candidate whose name appears on the ballot, and may refrain from voting for any or all candidates. 8. As provided by New York State law, to get on a minor party line a candidate files an acceptance with the minor party. Then, the party chair files an authorization. The candidate then becomes that party’s candidate. D.POPULATION STATISTICS 1. In 1900, the City’s total population was 19,457 and its total voting age population was 6,476, of whom 153 or 2.26% were African American. From 1900 through 1940, the City’s African American population never exceed 1.25% of the total population. In 1940, 975 or 1.25% of Niagara Falls’ 77,915 citizens were African American. 2. By 1950, the City’s population had grown to 90,872, of whom 3,698 or 4.07% were African American. 3. In the mid 1950’s, the Power Authority of the State of New York began construction of the Niagara Power Project, resulting in a significant, but temporary, increase in population and economic prosperity. By 1960, Niagara Falls’ population had increased to an all-time high of 102,394, of whom 7,664 or 7.48% were African American. 4. In a special census taken in April 1967, the City’s total population had fallen to 88,-286, a decrease of 13.8%. The African American population had remained almost static since I960, at 7,555. A decrease in white population continued in the following years. 5. In 1970, the total population of the City was 85,615, of whom 7,959 or 9.3% were African American. Total voting age population was 50,014, of whom 5,029 or 10.06% were African American. 6. In 1980, the total population had dropped to 70,165, while the African American population continued to increase, to 9,079 or 12.94% of the population. Total voting age population was 53,231, of whom 5,362 or 10.07% were African American. 7. In 1990, the total population further decreased to 61,840, of whom 9,634 or 15.58% were African American. 8. According to the 1990 census, the voting age population of Niagara Falls is 47,073, of whom 6,129 or 13.02% are African American. E. HISTORY OF ELECTIONS 1. At the time plaintiffs filed this lawsuit in September 1989, no African American had ever been elected to the Niagara Falls City Council. 2. In 1991, Andrew “Andy” Walker, an African American, was a successful council-manic candidate in the Democratic primary, and was the leading vote-getter in the general election. At the time of trial, he was Chairman of the Niagara Falls City Council. 3. Plaintiffs have analyzed the following referenda and elections in which African American candidates have run: Year African American Candidate Election and Office 1969 William Abrams City Council Primary 1971 Arthur Ray City Council Primary and General Election 1975 Joseph Profit Mayoral Primary 1977 Joseph Profit City Council Primary 1977 George Matthews City Council Primary 1979 George Matthews City Council Primary 1982 H. Carl McCall Lt. Governor Primary 1985 2 Propositions Referendum 1987 Renae Kimble City Council Primary and General Election 1988 Jesse Jackson Presidential Primary 4.Defendants have analyzed all Council-manic and Mayoral elections from 1975 through 1991, and the vote on four propositions in connection with the 1985 referenda. Defendants have included in their analysis elections in which no African American candidate has run. Defendants contend that this is acceptable because the Voting Rights Act imposes no requirement that the minority-preferred candidate must be a member of the minority group. 5. Between 90% and 97% of the registered African American voters in the City of Niagara Falls are registered Democrats. Therefore, neither plaintiffs nor defendants have analyzed Republican Party primaries, because virtually no African Americans participated in those elections. F. EXPERT METHODOLOGIES 1. Plaintiffs’ expert Michael McDonald analyzed elections primarily by using a method of statistical analysis called “simple” bivariate ecological regression analysis. McDonald also employed a method called “extreme case analysis” (also called “homogeneous precinct analysis”). Defendants’ expert Harold Stanley analyzed elections primarily by using “double” bivariate ecological regression analysis. Stanley used extreme case and simple bivariate ecological regression analyses only as double-checks on his double regression results. To the extent that they analyzed the same elections, the results obtained by McDonald and Stanley did not differ significantly, despite their primary reliance on different statistical methodologies. Moreover, there were no significant differences in results obtained by using the various statistical methods. Because McDonald placed greatest reliance on simple regression, this Court will refer to his results obtained by using that method. Because Stanley placed greatest reliance on double regression, this Court will refer to his results obtained by using that method. 2. Both bivariate ecological regression analysis and extreme case analysis are standard methods of analyzing racially polarized voting. These techniques produce data concerning the voting patterns of white and African American voters, including estimates of the percentages of voters of each race that voted for particular candidates. 3. Simple bivariate ecological regression analysis, on which McDonald relied, measures a candidate’s share of the votes received in a particular election district as a percentage of the number of voters at the polls in that district. That percentage is correlated with the racial composition of the district, measured in terms of percentage of the voting age population in that district. The results for each election district are plotted on a graph. A “correlation coefficient” is generated, demonstrating how consistently voter support for a candidate or group of candidates varies with the racial composition of the election districts. A correlation coefficient of — 1 or +1 indicates a perfect correlation between two variables. Simple regression does not allow for the effects of racial differences in voter turnout; it assumes that turnout rates between racial groups are the same. 4. Double bivariate ecological regression analysis, on which Stanley relied, is a method of ecological regression that does take into account racial differences in turnout rates, and provides an estimate of voter turnout by race. Double regression measures a candidate’s share of the votes received in a particular election district as a percentage of the voting age population of the district. That result is then correlated with the racial composition of the district, and the correlation coefficient is generated, as with simple regression. Then, the level of voter turnout by race is obtained by adding the percentages of whites or African Americans among the total voting age population estimated to have voted for each candidate. 5. Although both McDonald and Stanley acknowledged that “double regression” is a method that is generally preferable to “simple regression”, McDonald testified that he relied primarily on simple regression because, at least in City Council primaries, there was no reliable evidence of racially-related differences in voter turnout. Racially mixed districts generally had low voter turnout; however, districts in which African Americans comprised a majority generally had higher turnout rates than the racially mixed districts. In City Council primary elections, McDonald found no significant differences in voter turnout according to race. In general elections, McDonald found that African Americans turned out in fewer numbers than whites. However, plaintiffs did not undertake to analyze comprehensively the significance of these turnout differences. Furthermore, they did not report the turnout differences in elections that they did not analyze themselves. 6. McDonald relied on statistical results obtained by using simple regression, although these results generally provided a “slightly conservative” estimate of racial polarization rates. McDonald measured the degree of racial polarization in elections by calculating the difference between rates of African American cohesion and white crossover voting. 7. Although more than one seat was available for City Council elections, McDonald identified the minority-preferred candidate as the single candidate who received the highest percentage of African American support at the polls. Stanley identified and ranked more than one minority-preferred candidate in such elections. For example, if four seats were available, Stanley ranked the African American voters’ four most preferred candidates after identifying the candidates who received the four highest levels of African American support at the polls. G. STATISTICAL ANALYSIS OF ELECTIONS 1. 1969 City Council General Election Only plaintiffs analyzed this election, using 1970 census data. Willie Abrams, an African American, ran on a minor party label. In 1969, the City Council was comprised of four members, and two seats were being contested by five candidates. Three of those five candidates had received endorsements from both major and minor parties. This was not a significantly racially polarized election. Abrams received support from only 23.4% of African American voters and 1.4% of white voters. Plaintiffs did not report the levels of voter support enjoyed by the other candidates. 2. 1971 City Council Democratic Primary Only plaintiffs analyzed this election, using 1970 census data. Arthur Ray, an African American, ran in the Democratic primary against three other candidates for two available seats. Ray lost the primary. He received support from 92.1% of African American voters and 26.3% of white voters. McDonald calculated a polarization rate of 66%. Plaintiffs did not report the levels of voter support enjoyed by the other candidates. 3. 1971 City Council General Election Only plaintiffs analyzed this election. Although Arthur Ray lost in the Democratic primary, he ran in the general election on a minor party line. Five candidates ran for two available seats. Three of those five candidates ran on both major and minor party lines. Ray lost the general election. He received support from 78.2% of African American voters and 9.8% of white voters. 4. 1975 City Council Democratic Primary Plaintiffs did not analyze this election because no African American candidate contested. Seven Democrats ran for two available seats. Gallagher and Battaglia won the election. Defendants’ analysis shows that the two preferred candidates of African American voters were Sinclair, who received support from 33.4% of African American voters, and Gallagher, who received support from 31.3% of African American voters. Among white voters, the preferred candidates were Gallagher, who received 42.7% of the white vote, and Battaglia, who received 31.9% of the white vote. Sinclair received support from 24.6% of white voters. Therefore, because both white and African American voters preferred Gallagher, the election was not significantly racially polarized to the extent that Gallagher won. However, because white voters preferred Battag-lia as a second choice and African American voters preferred Sinclair as a first choice, the election was polarized to the extent that Bat-taglia won and Sinclair lost. 5. 1975 City Council General Election Plaintiffs did not analyze this election because no African American candidate contested. Seven candidates contested for two available seats. The election was won by the two endorsed Democrats, Gallagher and Battag-lia. Both of these candidates were preferred by both white and African American voters. Gallagher received support from 50.2% of white voters and 58.7% of African American voters. Battaglia received support from 42.3% of white voters and 52.5% of African American voters. Therefore, this election was not significantly racially polarized. 6. 1977 City Council Democratic Primary This election was analyzed by both plaintiffs and defendants. Six candidates contested for two available seats. Two of those six candidates, Profit and Matthews, were African American. Tangent and Martino won the election. Profit and Matthews were preferred by African American voters, but not by white voters. Plaintiffs’ simple regression analysis shows that Profit received support from 72.7% of African American voters but only 8.3% of white voters. McDonald calculated a polarization rate of 64%. Defendants’ double regression analysis shows that Profit received support from 71.3% of African American voters but only 7.2% of white voters. Plaintiffs’ simple regression analysis shows that Matthews received support from 68.3% of African American voters but only 5.2% of white voters. McDonald calculated a polarization rate of 63%. Defendants’ double regression analysis shows that Matthews received support from 65.8% of African American voters but only 4.4% of white voters. Plaintiffs did not report the levels of voter support enjoyed by any other candidates. Defendants’ double regression analysis further shows that the candidates preferred by white voters were Tangent, who received support from 47.5% of white voters and 13.1% of African American voters, and Marti-no, who received support from 44.0% of white voters but virtually no support from African American voters. Therefore, this election was polarized because Tangent and Martino won the election, whereas Profit and Matthews lost. 7. 1977 City Council General Election Plaintiffs did not analyze this election because no African American candidate contested. Five candidates contested for two available seats. Smith and Tangent won the election. Defendants’ double regression analysis shows that Martino and Tangent were preferred by African American voters; Smith and Pitarre-si were preferred by white voters. Martino received support from 79.5% of African American voters and 39.4% of white voters. Tangent received support from 71.6% of African American voters and 42.3% of white voters. Smith received support from 9.4% of African American voters and 45.4% of white voters. Pitarresi received support from 42.6% of white voters, but virtually no support from African American voters. Therefore, this election was racially polarized to the extent that the most preferred candidate of white voters won, whereas the most preferred candidate of African American voters lost. It is important to note that African American voters were able to elect Tangent, their second most preferred candidate, despite the fact that Tangent was not one of the two most preferred candidates of white voters. 8. 1979 City Council Democratic Primary This election was analyzed by both plaintiffs and defendants. Six candidates contested for two available seats. One of those candidates, Matthews, was an African American. Sottile and Martel won the election; Matthews lost. Matthews and Sottile were preferred by African American voters; Sottile and Martel were preferred by white voters. Plaintiffs’ simple regression analysis shows that Matthews received support from 88% of African American voters but only 7.3% of white voters. McDonald calculated a polarization rate of approximately 81%. Plaintiffs did not report the levels of voter support enjoyed by the other candidates. Defendants’ double regression analysis shows that Matthews received support from 97.2% of African American voters but only 6.9% of white voters. The candidates preferred by white voters were Martel, who received support from 41% of white voters, and Sottile, who received support from 35% of white voters. The candidates preferred by African American voters were Matthews, who received support from 97.2% of African American voters, and Sottile, who received support from 11.2% of African American voters. Martel, who received the greatest support from white voters, received virtually no support from African American voters. Therefore, because both whites and African Americans preferred Sottile, the election was not significantly racially polarized to the extent that Sottile won. However, because whites preferred Martel as a first choice, and African Americans preferred Matthews as a first choice, the election was polarized to the extent that Martel won and Matthews lost. 9.1979 City Council General Election Plaintiffs did not analyze this election because no African American candidate contested. Five candidates contested for two available seats. Bazzoni and Martel won the election. Sottile and Martel were preferred by African American voters; Bazzoni and Martel were preferred by white voters. Defendants’ double regression analysis shows that Bazzoni received support from 57.9% of white voters, but virtually no support from African American voters. Martel received support from 47.8% of white voters, and 25.6% of African American voters. Sottile received virtually unanimous support from African American voters and support from 31.9% of white voters. Therefore, this election was racially polarized to the extent that the most preferred candidate of white voters won, whereas the most preferred candidate of African American voters lost. 10. 1981 City Council Democratic Primary Plaintiffs did not analyze this election because no African American candidate contested. Eight Democrats contested for two available seats. O’Farrell and Rendina won the election. Redding and Tangent were preferred by African American voters; Rendina and O’Farrell were preferred by white voters. Defendants’ double regression analysis shows that Rendina received support from 39.0% of white voters but only 2.6% of African American voters. O’Farrell received support from 36.7% of white voters and 30.6% of African American voters. Redding received support from 33.4% of African American voters and 25.3% of white voters. Tangent received support from 33.3% of African American voters and 24.4% of white voters. Therefore, this election was racially polarized to the extent that white voters were able to elect their preferred candidates (although their first candidate took second place, and vice versa), whereas African American voters were not. 11. 1981 City Council General Election Plaintiffs did not analyze this election because no African American candidate contested. Seven candidates contested for two available seats. Smith and Cook, both Republicans, won the election. Smith and Cook were preferred by white voters; O’Farrel and Rendina were preferred by African American voters. Defendants’ double regression analysis shows that Smith received support from 51.3% of white voters but virtually no support from African American voters. Cook received support from 45.9% of white voters but virtually no support from African American voters. O’Farrell received support from 73.3% of African American voters and 35.6% of white voters. Rendina received support from 37.3% of African American voters and 42.3% of white voters. This election was polarized to the extent that the two candidates preferred by white voters won, whereas the two candidates preferred by African American voters lost. 12. 1988 City Council Democratic Primary Plaintiffs did not analyze this election because no African American candidate contested. Three Democrats contested for two available seats. Redding and Martel won the election. Redding and Martel were preferred by both African American and white voters. Redding received support from 56.3% of African American voters and 61.1% of white voters. Martel received support from 35.1% of African American voters and 60.3% of white voters. Therefore, this election was not significantly racially polarized. 13. 1983 City Council General Election Plaintiffs did not analyze this election because no African American candidate contested. Four candidates contested for two available seats. Redding and Martel won the election. Redding and Martel were preferred by African American voters; Redding and Bazzoni were preferred by white voters. Defendants’ double regression analysis shows that Redding received support from 90.6% of African American voters and 52.3% of white voters. Martel received support from 79.5% of African American voters and 42.1% of white voters. Bazzoni received support from 44.8% of white voters but virtually no support from African American voters. Therefore, this election was polarized to the extent that Martel, the second most preferred candidate of African American voters, won (as did their most preferred candidate), whereas Bazzoni, the second most preferred candidate of white voters, lost. It is important to note that white bloc voting in this election was not substantial enough to defeat the minority’s second most preferred candidate, whereas it was insufficient to elect white voters’ second most preferred candidate. 14. 1985 City Council Democratic Primary Plaintiffs did not analyze this election because no African American candidate contested. Six candidates contested for two available seats. Quaranto and Soda won the election. Cook and Soda were preferred by African American voters; Quaranto and Soda were preferred by white voters. Defendants’ double regression analysis shows that Cook received support from 50.3% of African American voters and 14.6% of white voters. Soda received support from 35.2% of African American voters and 44.4% of white voters. Quaranto received support from 51.2% of white voters but only 21.9% of African American voters. Therefore, this election was polarized to the extent that Quaranto, who was the most preferred candidate of white voters, won, whereas Cook, who was the most preferred candidate of African American voters, lost. 15. 1985 City Council General Election Plaintiffs did not analyze this election because no African American candidate contested. Seven candidates contested for two available seats. Quaranto and Soda won the election. Quaranto and Soda were the preferred candidates of both African American and white voters. Defendants’ double regression analysis shows that Quaranto received support from 50.5% of African American voters and 56.3% of white voters. Soda received support from 46.3% of African American voters and 46.3% of white voters. Therefore, this election was not significantly racially polarized. 16.1987 City Council Democratic Primary Both plaintiffs and defendants analyzed this election. Nine Democrats contested for five available seats on the City Council, which was enlarged to seven members by the 1985 referendum adopting the new City Charter. Renae Kimble, an African American and a plaintiff in this action, was one of the candidates. This election was won by Sottile, Rendina, Agnello, Buehalski, and Aversa. Kimble lost the election, coming in six votes behind Aversa, out of 45,505 votes cast. Kimble was the most preferred candidate of African American voters, but was not one of the five most preferred candidates of white voters. Plaintiffs’ simple regression analysis shows that Kimble received support from 92.2% of African American voters but only 29% of white voters. McDonald calculated a polarization rate of 63.2%. Plaintiffs did not report the levels of voter support enjoyed by the other candidates. Defendants’ double regression analysis shows that Kimble received support from 97.1% of African American voters and 27.5% of white voters. The five most preferred candidates of African American voters were Kimble, Agnello, Rendina, Aversa, and Bu-ehalski; the five most preferred candidates of white voters were Sottile, Rendina, Giove, Buehalski, and Aversa. The support garnered by Kimble has already been discussed. Agnello received support from 29.7% of African American voters and 41.4% of white voters. Rendina received support from 28.7% of African American voters and 49.3% of white voters. Aversa received support from 20.7% of African American voters and 41.8% of white voters. Buehalski received support from 13.8% of African American voters and 43.5% of white voters. Sottile, the candidate most preferred by whites, received support from 52.7% of white voters and 11.7% of African American voters. Therefore, this election was polarized to the extent that Kimble, who was the most preferred candidate of African Americans, lost by six votes. This Court also notes that Giove, who was the third most preferred candidate of white voters, lost the election. 17.1987 City Council General Election Both plaintiffs and defendants analyzed this election. Eleven candidates contested for five available seats. Kimble, who had lost the Democratic primary election, ran on the Right-to-Life and Liberal party labels. Sottile, Ren-dina, Geracitano, Buchalski, and Aversa won the election. Plaintiffs’ simple regression analysis shows that Kimble received the support of 57.0% of African American voters but only 7.9% of white voters. McDonald calculated a polarization rate of 49%. Plaintiffs did not report the levels of voter support enjoyed by the other candidates. Defendants’ double regression analysis shows that Kimble received support from 77.0% of African American voters but only 8.3% of white voters. The five candidates most preferred by African American voters were Kimble, Agnello, Sottile, Rendina, and Buchalski; the five candidates most preferred by white voters were Sottile, Geracita-no, Rendina, Buchalski, and Caso. The support garnered by Kimble has already been discussed. Agnello received support from 38.3% of African American voters and 39.4% of white voters. Sottile received support from 37.9% of African American voters and 51.7% of white voters. Rendina received support from 23.5% of African American voters and 47.1% of white voters. Buchalski received support from 17.2% of African American voters and 45.4% of white voters. Geracitano received the support of 47.8% of white voters but virtually no support from African American voters. Caso received support from 44.3% of white voters but virtually no support from African American voters. Therefore, this election was polarized to the extent that Kimble and Agnello, who were the two most preferred candidates of African American voters, lost, whereas all but the fifth most preferred candidate of white voters won. However, it is important to note here that Kimble testified during trial that she did not actively campaign in the 1987 general election. Furthermore, she suggested that her name remained on the ballot because she had no right to remove it from the Right-to-Life and Liberal Party lines. It is also somewhat significant that African American votes prevented whites from electing their fifth most preferred candidate. 18. 1989 City Council General Election Plaintiffs did not analyze this election because no African American candidate contested. The 1989 City Council Democratic primary was uncontested, there being four Democrats running to fill four available seats. In the general election, eleven candidates contested for four available seats. Quaranto, Gawel, Palillo, and Buchalski won the election. Defendants’ double regression analysis shows that the four most preferred candidates of African American voters were Albond, Quar-anto, Buchalski, and Aversa; the four most preferred candidates of white voters were Quaranto, Gawel, Palillo, and LaRocea. Al-bond received support from 58.5% of African American voters and 26.2% of white voters. Quaranto received support from 47.4% of African American voters and 56.1% of white voters. Buchalski received support from 40.8% of African American voters and 40.0% of white voters. Aversa received support from 40.1% of African American voters and 33.6% of white voters. Gawel received support from 49.8% of white voters but virtually no support from African American voters. Palillo received support from 48.4% of white voters but virtually no support from African American voters. LaRocca received support from 42.8% of white voters but virtually no support from African American voters. Therefore, this election was polarized to the extent that the two most preferred candidates of white voters won, whereas the two most preferred candidates of African American voters lost. However, it is noted that Buchalski was only able to win because of African American voter support. 19. 1991 City Council Democratic Primary This election was analyzed by both plaintiffs and defendants. Five Democrats contested for three available seats. One of those candidates, Andrew Walker, was an African American. At the time of trial, Walker was Chairman of the Niagara Falls City Council, and he is a defendant in this lawsuit. Sottile, Morello, and Walker won the election. Plaintiffs’ simple regression analysis shows that Walker received support from 79.4% of African American voters and 38.6% of white voters. McDonald calculated a polarization rate of just under 41%. Plaintiffs did not report the levels of voter support enjoyed by the other candidates. Defendants’ double regression analysis shows that Walker received support from 85.7% of African American voters and 37.6% of white voters. The three most preferred candidates of African American voters were Walker, Rendina, and Sottile; the three most preferred candidates of white voters were Sottile, Morello, and Aversa. The support garnered by Walker has already been discussed. Rendina received support from 14.9% of African American voters and 41.8% of white voters. Sottile received support from 9.1% of African American voters and 58.5% of white voters. Aversa received support from 43.6% of white voters and 4.6% of African American voters. Therefore, this election was polarized to the extent that the second most preferred candidate of African American voters lost, whereas the second most preferred candidate of white voters won. It is noted, however, that the third most preferred candidate of African American voters won, whereas the third most preferred candidate of white voters lost. 20. 1991 City Council General Election Both plaintiffs and defendants analyzed this election. Nine candidates, including Walker, contested for three available seats. Walker, Sottile, and Geracitano won the election. Plaintiffs’ simple regression analysis shows that Walker received support from 84.8% of African American voters and 42.9% of white voters. McDonald calculated a polarization rate of approximately 41%. Plaintiffs did not report the levels of voter support enjoyed by the other candidates. Defendants’ double regression analysis shows that Walker received virtually unanimous support from African American voters and support from 43.2% of white voters. The three most preferred candidates of African American voters were Walker, Sottile, and Morello; the three most preferred candidates of white voters were Geracitano, Sot-tile, and Walker. The support garnered by Walker has already been discussed. Sottile received support from 45.7% of African American voters and 45.1% of white voters. Morello received support from 35.4% of African American voters and 38.3% of white voters. Geracitano, who was the candidate most preferred by white voters, received support from 45.3% of white voters but virtually no support from African American voters. Therefore, this election was polarized to the extent that the third most preferred candidate of African American voters lost, whereas all three of the most preferred candidates of white voters won. 21. 1975 Mayoral Democratic Primary Both plaintiffs and defendants analyzed this election. Three Democrats contested, including Profit, an African American. O’Laughlin won the election. Profit was the candidate most preferred by African American voters; O’Laughlin was the candidate most preferred by white voters. Plaintiffs’ simple regression analysis shows that Profit received support from 47.4% of African American voters but only 1.4% of white voters. McDonald calculated a polarization rate of 46%. Plaintiffs did not report the levels of voter support enjoyed by the other candidates. Defendants’ double regression analysis shows that Profit received support from 50.5% of African American voters but only 1.7% of white voters. O’Laughlin received support from 65.3% of white voters and 22.9% of African American voters. The third candidate, Sottile, received support from 19.3% of African American voters and 28.1% of white voters. Therefore, this election was racially polarized to the extent that O’Laughlin, the most preferred candidate of white voters, won, whereas Profit, the most preferred candidate of African American voters, lost. However, based on plaintiffs’ analysis, Profit was preferred by African American voters only by a plurality, rather than by a majority. 22. 1975 Mayoral General Election Plaintiffs did not analyze this election because no African American candidate contested. Three candidates contested, and O’Laugh-lin won. O’Laughlin was the most preferred candidate of both African American and white voters. He received support from 78.4% of African American voters and 56.5% of white voters. Therefore, this election was not significantly racially polarized. 23. 1979 Mayoral Democratic Primary Plaintiffs did not analyze this election because no African American candidate contested. Two Democrats contested, and O’Laughlin won the election. O’Laughlin was the candidate preferred by both African American and white voters. He received support from 54.1% of African American voters and 56.2% of white voters. Therefore, this election was not significantly racially polarized. 24. 1979 Mayoral General Election Plaintiffs did not analyze this election because no African American candidate contested. Two candidates contested, and O’Laughlin won the election. O’Laughlin was the candidate preferred by both African American and white voters. He received support from 95.2% of African American voters and 50.5% of white voters. Therefore, this election was not significantly racially polarized. 25. 198S Mayoral General Election Plaintiffs did not analyze this election because no African American candidate contested. No Democratic primary took place in 1983, because O’Laughlin was uncontested for the Democratic nomination. In the general election, two candidates contested, and O’Laugh-lin won the election. O’Laughlin was the candidate preferred by both African American and white voters. He received support from 89.7% of African American voters and 54.0% of white voters. Therefore, this election was not significantly racially polarized. 26. 1987 Mayoral Democratic Primary Plaintiffs did not analyze this election because no African American candidate contested. Four Democrats contested, and O’Laughlin won. Massaro was the candidate most preferred by African American voters; O’Laughlin was the candidate most preferred by white voters. Defendants’ double regression analysis shows that Massaro received support from 33.1% of African American voters and 31.6% of white voters. O’Laughlin received support from 28.6% of African American voters and 35.5% of white voters. Therefore, this election was racially polarized. 27. 1987 Mayoral General Election Plaintiffs did not analyze this election because no African American candidate contested. Four candidates contested, and O’Laughlin won. O’Laughlin was the candidate most preferred by both African American and white voters. O’Laughlin received support from 66.9% of African American voters and 50.3% of white voters. Therefore, this election was not significantly racially polarized. ' 28. 1991 Mayoral Democratic Primary Both plaintiffs and defendants analyzed this election. Three Democrats contested, including plaintiff Renae Kimble. Quaranto won the election. Kimble was the candidate most preferred by African American voters; Quar-anto was the candidate most preferred by white voters. Plaintiffs’ simple regression analysis shows that Kimble received support from 93.9% of African American voters but only 8.1% of white voters. McDonald calculated a polarization rate of approximately 85.8%. Plaintiffs did not report the levels of voter support enjoyed by the other candidates. Defendants’ double regression analysis shows that Kimble received virtually unanimous support from African American voters but support from only 6.1% of white voters. Quaranto received support from 61.0% of white voters but virtually no support from African American voters. Therefore, this election was racially polarized. 29. 1991 Mayoral General Election Plaintiffs did not analyze this election because no African American candidate contested. Four candidates contested, and Palillo won. Quaranto was the candidate most preferred by African American voters; Palillo was the candidate most preferred by white voters. Defendants’ double regression analysis shows that Quaranto received support from 89.9% of African American voters and 44.4% of white voters. Palillo received support from 52.3% of white voters but virtually no support from African American voters. Therefore, this election was racially polarized. 30. 1982 New York State Lieutenant Governor Democratic Primary Only plaintiffs analyzed this election. H. Carl McCall, an African American, was a candidate. Plaintiffs’ simple regression analysis shows that McCall was the candidate most preferred by African American voters. McCall received support from 96.7% of African American voters and 36.8% of white voters. Therefore, this election was racially polarized. 31. 1988 Presidential Primary Only plaintiffs analyzed this election. Jesse Jackson, an African American, was a candidate. Plaintiffs’ simple regression analysis shows that Jackson received virtually unanimous support from African American voters but support from only 18.6% or 10.9% of white voters, depending upon whether total population or voting age population figures are used. Therefore, this election was racially polarized. McDonald calculated a polarization rate of approximately 89%. 32. 1985 Referenda As indicated above, voters were given the opportunity to choose from among four options in revising the Niagara Falls City Charter, or to maintain the status quo. Plaintiffs analyzed only the choice between a mayoral or “city manager” form of government with City Council members chosen from single-member districts. Plaintiffs’ simple regression analysis shows that 67.6% of African American voters and 40.3% of white voters voted in favor of a “strong may- or” form of government with seven council-manic districts. Plaintiffs’ analysis further shows that 47.2% of African American voters and 28.5% of white voters voted in favor of a “strong city manager” form of government with six councilmanic districts. McDonald testified that he could not confidently say that significant racial polarization was evident in this last voting pattern. Defendants analyzed all the propositions on the 1985 referendum. Defendants’ double regression analysis shows that 56.1% of white voters and 4.4% of African American voters voted in favor of a mayor-council form of government with at-large representation, whereas 27.3% of white voters and 14.2% of African American voters voted in favor of a “strong mayor” form of government with seven councilmanic districts. 13.0% of white voters and 4.8% of African American voters voted in favor of the city manager form of government with at-large representation, and 17.5% of white voters and 3.9% of African American voters voted in favor of the city-manager form of government with six eoun-cilmanie districts. H. PRIMA FACIE CASE OF VOTE DILUTION Size and Geographic Compactness of Minority Group 1. The parties do not contest the fact that the African American community in the City of Niagara Falls is sufficiently large and geographically compact to comprise an effective majority-minority district in a hypothetical single-member district system. 2. Plaintiffs introduced the expert witness testimony of Jerry Wilson, a demographer. Using 1990 census data, Wilson showed that the City of Niagara Falls has a total population of 61,856, and a total African American population of 9,634. The City’s total voting age population is 47,073, and African American voting age population is 6,129. The City’s African American voting age population constitutes approximately 13% of the total voting age population. Approximately 30% of the African Americans in the City of Niagara Falls live in census tract 202. Positing a hypothetical seven-member City Council elected from seven single-member districts, Wilson calculated that an ideal election district would be comprised of a total population of 8,834. 3. Using plaintiffs exhibit. 19, Wilson showed that a viable and geographically compact African American majority-minority district could be created, primarily out of census tract 202. The district would contain a total population of 8,394, which would deviate by approximately 4.98% from the population of an ideal district. The district would contain a total voting age population of 5,434. The district would have a total African American population of 5,400, constituting 64.33% of the district’s total population. The district would have a total African American voting age population of 3,291, constituting 60.56% of the district’s total voting age population. 4. Therefore, plaintiffs have met the first prong of Gingles. 5. However, as defendants note, before the size of the City Council was increased from four to seven in 1987, no viable African American majority-minority district could have been designed. Furthermore, using defendants’ exhibit FF, defendants illustrated that, according to 1980 census figures, plaintiffs could create a majority-minority district with only a 50.22% African American voting age population. Furthermore, the hypothetical district would not be geographically compact. These circumstances diminish the probative value of plaintiffs’ analysis of pre-1987 elections. 6. In addition, plaintiffs’ own expert witnesses testified that the at-large electoral system normally cannot be shown to be responsible for a minority group’s inability to elect candidates of its choice until the minority population in the jurisdiction reaches approximately 10-15% of the total population. See Gingles, 478 U.S. at 50, 106 S.Ct. at 2766 (If minority group is not sufficiently large and geographically compact to constitute a majority in a single-member district, “the multi-member form of the district cannot be responsible for minority voters’ inability to elect its candidates.”) (emphasis in original). However, the stipulated facts in this case demonstrate that the census data pertaining to the year 1971 show that African Americans comprised only 9.3% of the population of the City of Niagara Falls in that year. Therefore, by plaintiffs’ own estimation, the 1971 elections analyzed by plaintiffs are of marginal probative value in determining whether African American voters have been unable to elect their representatives of choice as a result of the present form of government. Political Cohesiveness of African Americans in the City of Niagara Falls 1. The parties agree that plaintiffs have shown that African Americans have been politically cohesive in their voting patterns. 2. Political cohesiveness is demonstrated in the present ease regardless of whether this Court considers only elections in which an African American candidate has contested, or whether it considers all the elections analyzed by the parties. 3. Using simple regression analysis, McDonald demonstrated that, with the exceptions of the Willie Abrams candidacy in the 1969 City Council election and the Joseph Profit candidacy in the 1975 Mayoral election, African American voters cast a substantial majority of their votes for African American candidates when African American candidates were on the ballot. 4. Using double regression analysis, Stanley found that “a significant number of minority group members usually vote for the same candidates” or propositions. Gingles, 478 U.S. at 56, 106 S.Ct. at 2769. 5. Therefore, plaintiffs have met the second prong of Gingles. Third Prong of Gingles 1. After reviewing the parties’ analyses of elections in the City of Niagara Falls, this Court concludes that plaintiffs have not met their burden of showing that the white majority in Niagara Falls “votes sufficiently as a bloc to enable it — -in the absence of special circumstances ... — usually to defeat the minority’s preferred candidate.” Gingles, 478 U.S. at 51, 106 S.Ct. at 2766-67. 2. Plaintiffs undertook to meet their burden of proof with regard to this prong by analyzing certain elections in an attempt to demonstrate which single candidate would have won each election in a hypothetical African American majority single-member district. However, assuming the validity of that approach, plaintiffs’ analysis did not extend to elections in which no African American candidate contested. But see Citizens for a Better Gretna v. City of Gretna, 834 F.2d 496, 502 (5th Cir.1987), cert. denied, 492 U.S. 905, 109 S.Ct. 3213, 106 L.Ed.2d 564 (1989) (“In a multiple seat election ... the minority necessarily will have more than one preferred candidate.”); Valladolid v. City of National City, 976 F.2d 1293 (9th Cir.1992); Clarke v. Cincinnati, 1993 WL 761489 (S.D.Ohio 1993). Therefore, plaintiffs failed to analyze the City Council Democratic primary elections for the years 1975,1981,1983, and 1985; and the City Council general elections for the years 1975, 1979, 1981, 1983, 1985, and 1989. Moreover, plaintiffs failed to analyze the mayoral Democratic primaries for the years 1979 and 1987; and the mayoral general elections for the years 1975,1979,1983, 1987, and 1991. 3. By declining to analyze these elections, plaintiffs provided this Court with no indication of which candidates would have won these elections if they had taken place in a hypothetical, African American majority, single-member district. Moreover, this Court cannot determine itself which candidates would have won those elections, because such a determination depends upon complicated mathematical formulae, and requires expert analysis and testimony. Therefore, plaintiffs have simply failed to meet their burden of proof at trial in this regard. 4. When all of the elections analyzed by plaintiffs and defendants are considered, the proof shows that, on the whole, the white majority population in the City of Niagara Falls does not vote in such a way that it usually defeats the African American voters’ candidate of choice. 5. In recent elections, African Americans have been particularly successful in electing their candidates of choice. 6. In reviewing the elections analyzed by the parties, this Court has paid particular attention to recent elections for seats to the Niagara Falls City Council. This is especially appropriate because of the diminished relevance of the 1971 and pre-1987 elections for the reasons discussed above. 7. Particularly relevant to this Court’s analysis is the fact that African American voters’ candidate of choice in the 1991 City Council Democratic primary and general elections, Andrew Walker, won those elections. Although these victories were obtained after this lawsuit was filed, this Court finds insufficient evidence in the record to conclude that their significance should be minimized. Specifically, plaintiffs have failed to persuade this Court that these victories were the result of an attempt by white voters to “humiliate” this lawsuit. Furthermore, African Americans succeeded in electing their “second choice” candidate, Sottile, in the 1991 City Council general election, and their “third choice” candidate, Sottile, in the 1991 City Council Democratic primary. Although minority success in one or a few elections does not per se foreclose a vote dilution claim, see Gingles, 478 U.S. at 55-59 and n. 25, 106 S.Ct. at 2769-70 and n. 25, these elections are highly relevant because they are recent, and because they involve the Niagara Falls City Council. 8. In the 1989 City Council general election, African Americans succeeded in electing two of their four most preferred candidates, Quaranto and Buchalski. They were able to elect Buchalski despite the fact that he was not one of the four most preferred candidates of white voters. 9. In the 1985 City Council general election, African American voters were successful in electing their two preferred candidates, Quaranto and Soda, to the two available seats on the City Council. In the 1985 City Council Democratic primary election, African American voters were able to elect one of their two preferred candidates. 10. It is also somewhat significant that Renae Kimble, who was the African American voters’ candidate of choice in the 1987 City Council Democratic primary, lost by only six votes. Although she lost in the 1987 City Council general election — an election that was analyzed by plaintiffs in support of their claims — the significance of her loss must be minimized by the fact that she chose not to campaign actively in that election. 11. Plaintiffs were unable to prove to this Court’s satisfaction that Renae Kimble would not have won the 1987 City Council general election under the at-large scheme if she had run on the Democratic Party line, due to unresolvable conflicts between the statistical approaches taken by Professors Loewen and Stanley. This diminishes plaintiffs’ proof that the at-large system made a difference in this election because Kimble would have won the election if she had run in an African American majority, single-member district. 12. In the 1987 City Council general election, African American voters were successful in electing three out of five of their most preferred candidates. In the 1987 City Council Democratic primary, they were able to elect four out of five of their most preferred candidates. 13. As noted above, the relevance of the 1971 elections analyzed by plaintiffs are of diminished significance, due to the fact that at that time African Americans constituted only 9.3% of the total population of the City of Niagara Falls. 14. Similarly, as noted above, the relevance of the pre-1987 City Council primaries and general elections analyzed by plaintiffs are of diminished significance, due to the fact that, prior to the City Charter amendment after the 1985 referendum, the City Council was comprised of only four members. Plaintiffs’ expert agreed that it would not have been possible to form an African American majority single-member district at that time. See Gingles, 478 U.S. at 50, 106 S.Ct. at 2766 (If minority group is not sufficiently large and geographically compact to constitute a majority in a single-member district, the multimember or at-large system being challenged “cannot be responsible for minority voters’ inability to elect its candidates.”). Therefore, if this litigation had taken place before 1985, i.e., before the change in the size