Vecinos De Barrio Uno v. City of Holyoke, 960 F. Supp. 515 (D. Mass. 1997)

Daniel J. Gleason, Julie A. Trachten, Nutter, McClennen & Fish, Ozell Hudson, Jr., Lawyers' Committee for Civil Rights, Alan J. Rom, Law Office of Sherwin L. Kantrovitz, Boston, MA, David P. Hoose, Alan M. Katz, Katz, Sasson & Hoose, Sprigfield, MA, William Newman, Lesser, Newman, Souweine & Nasser, Northampton, MA, for Vecinos Debarrio Uno, Latino Citizens United for Holyoke, Lillian Santiago-Garcia, Maria Santiago, Sonia Rodriguez, Piculin Rolan-Cruz, Ana Ramos, Gloria Caballer Arce, Luis-Orlando Isaza.

Kenneth J. Cote, Jr., Daniel M. Glanville, City of Holyoke Law Dept., Holyoke, MA, Steven P. Perlmutter, Robinson & Cole, Boston, MA, Edward R. Mitnick, Asst. City Solicitor, Sheridan & Sheridan, South Hadley, MA, for City of Holyoke.

Kenneth J. Cote, Jr., City of Holyoke Law Dept., Holyoke, MA, Steven P. Perlmutter, Robinson & Cole, Boston, MA, Edward R. Mitnick, Asst. City Solicitor, Sheridan & Sheridan, South Hadley, MA, for William A. Hamilton, Holyoke Bd. of Registrars, Henry Wheeler, Elba Rueda, Timothy Howes.

PONSOR, District Judge.

This lawsuit was filed in 1992 by Hispanic citizens of the City of Holyoke against the *516 City itself and certain individuals, claiming violations of the Voting Rights Act of 1982, 42 U.S.C. § 1973. Originally, the suit challenged the at-large component of the elections for both Holyoke's School Committee and City Council.^[1]

On March 27, 1995, following a bench trial, this court issued an opinion holding that the at-large component for the City Council elections, by which a majority of the Council is elected, did violate the Voting Rights Act. On April 27, 1995, after giving all parties the opportunity to propose an appropriate remedy, this court issued its remedial order, reducing the at-large slots for the City Council members from eight to two. The court found no violation of the Act with regard to School Committee elections, to which only two members had been elected at-large for many years, and entered judgment for the defendants on that portion of the lawsuit.

Defendants appealed the court's decision regarding the City Council. On December 20, 1995, the First Circuit Court of Appeals vacated the judgment and remedial order and returned the case to this court. Judge Selya noted that while "the district court's opinion is well-considered and in many respects deftly navigates the marshy terrain of voting rights jurisprudence," it nevertheless lacked a clear explanation of the connection between its factual findings and legal conclusions. Vecinos De Barrio Uno v. Holyoke, 72 F.3d 973, 991 (1st Cir. 1995). The First Circuit's decision directed this court to provide further explanation on some specific issues, invited more discussion of others if appropriate, and left the procedure to be followed on remand to the court's discretion. Prominent among the issues to be addressed on remand was the significance of Holyoke's "rapidly changing political environment," Id., at 989.

Upon remand this court issued a scheduling order to govern further proceedings; counsel then requested a stay to permit settlement discussions. When these discussions did not bear fruit, this court issued an order requiring submissions on three questions. First, what legal and factual issues, in counsel's view, did the Court of Appeals require this court to address, or suggest that this court might address? Second, what was counsel's position regarding each of these issues? Third, what future proceedings would be needed in order to present evidence on any disputed point?

On May 23, 1996, after receiving counsel's submissions, the court issued a memorandum detailing the matters to be addressed on remand. Vecinos De Barrio Uno v. Holyoke, 926 F. Supp. 23 (D.Mass.1996). A schedule at the conclusion of the memo established the timeline for supplemental discovery and further evidentiary proceedings.

In December 1996 further testimony over four days included updated analyses from both the parties' experts and testimony from several fact witnesses, including Alejandro Sanchez, Jr., an unsuccessful candidate in the 1995 at-large election for the City Council, and Joseph McGivern, current president of the City Council. Upon conclusion of the evidentiary proceedings the court received extensive supplemental memoranda directed to the issues on remand.

This court has now considered the testimony, documents and memoranda submitted by both sides, particularly the evidence regarding the 1995 City Council election and the testimony of Sanchez and McGivern. In view of what at this time appears to be the significant evolution both of Holyoke's political climate and its electoral practices, this court will enter judgment for defendants on the remaining Voting Rights claim. Recent developments within Holyoke make it impossible to say, at this time, that the current City Council election system provides Hispanic voters less opportunity than other members of the electorate to participate in the political process and elect representatives of their choice. Since this decision obviously represents a departure from the court's earlier holding, it is important to set forth the underlying reasons in detail.

An extensive narration of the essential facts and the basic legal authorities originally underlying this lawsuit may be found in the district court's March 1995 decision and the First Circuit's opinion on appeal. Vecinos De Barrio Uno v. City of Holyoke, 880 F. Supp. 911 (D.Mass.1995), vacated and remanded, 72 F.3d 973 (1st Cir.1995). This memorandum will confine itself to facts and law necessary to address the First Circuit's questions and set out the rationale for this decision.

This court's May 23, 1996 memorandum identified the court's potential tasks on remand as follows: first, a reconsideration of the original decision in light of the First Circuit's memorandum and the parties' further submissions; second, an analysis of whether "racial antagonism" exists in Holyoke and the extent to which it plays a part in the electoral process; third, an appraisal of the question whether the Hispanic community has recently enjoyed a "rise in its political fortunes;" fourth, a closer look at how (if at all) results of ward elections shed light on the at-large election process; fifth, an exploration of whether Ward 4 constitutes an "influence district" as the case law has defined this term; sixth, an evaluation of the 1995 City Council election in light of the Voting Rights Act; seventh, a reassessment of the significance (if any) of election results where Hispanic candidates ran unopposed, and, finally, an analysis of updated evidence regarding the current proportion of Hispanics in Holyoke's general and voting age populations.

As it happened, some of these eight issues receded or dropped out of consideration completely during the subsequent evidentiary proceedings and briefing. No direct evidence of overt "racial antagonism" was submitted, and little discussion of this point was offered, other than defendant's argument that it did not play a role in the 1995 election and plaintiffs' contention that racial antagonism should be inferred from continued white bloc voting against qualified Hispanic candidates. The parties submitted no further evidence and only minimal argument concerning elections where Hispanic candidates ran unopposed, and the court will disregard these elections on remand. Finally, the court turned aside plaintiffs' effort to offer evidence regarding purported increases since the 1990 census in the proportion of Hispanic persons in Holyoke's general and voting age populations, when the expert designated to offer such evidence disclosed that she would need significant additional time to put the data in presentable form.

The court's analysis on remand will begin with a reconsideration of the three threshold criteria for a Voting Rights case, and then take a fresh look at the "totality of the circumstances" bearing on the election system now informed by the parties' supplemental submissions and, most importantly, by the evidence of the 1995 election. This examination will include scrutiny of the significance of ward elections and the possibility that Ward 4, with a 28% Hispanic voting age population, may be deemed an "influence district." Next, the court will examine the evidence of any increase in the political strength of the Hispanic community, once more focusing on the evidence of the 1995 election and the light it throws on developments over the past decade in Holyoke.

As will be seen, the new Sanchez/McGivern testimony has turned the tide for the defendants. Sanchez, an Hispanic candidate who achieved remarkable success with non-Hispanic, as well as Hispanic, voters on his first try for office in 1995, provided compelling evidence of the current accessibility of the election system in Holyoke. McGivern, a seasoned non-Hispanic candidate who has solicited and received significant support from Hispanic voters, also testified persuasively about the open nature of Holyoke's political life at this time. More than the parties' highly qualified expert statisticians, these two real life "experts" in Holyoke's 1997 political scene have demonstrated that the city's at-large election system for the City Council cannot be proved, as of this time, to violate Section 2 of the Voting Rights Act.

The application of the Voting Rights Act to a particular political context is one of the most difficult and intricate responsibilities a *518 district court must shoulder. Viewed from the perspective of the judge as factfinder, the subtle, constantly changing quality of the community environment makes ascertaining the pertinent social and political realities, and describing them accurately at a given point in time, a little like attempting to write on water.

At the same time, observed from the vantage of the judge as construer of the law, the Voting Rights Act and its interpretive decisional authority offer awkward tools for such a delicate job. Applying this cumbersome, almost paradoxical statute in such an important area is analogous to performing heart surgery with a butter knife.

The Supreme Court's seminal decision in Thornburg v. Gingles, 478 U.S. 30, 106 S. Ct. 2752, 92 L. Ed. 2d 25 (1986), established the prerequisites to a successful Voting Rights case. First, the minority group must be "sufficiently large and geographically compact to constitute a majority in a single-member district." Id., at 50-51, 106 S. Ct. at 2766. Second, the minority group must be politically cohesive. Third, the facts must reveal that "the white majority votes sufficiently as a bloc to enable it ... usually to defeat the minority's preferred candidate." Id. If these criteria are satisfied, the court must then examine the totality of the circumstances to determine whether the election system, seen in its social and historical context, causes "an inequality in the opportunities enjoyed by [minority] and white voters to elect their preferred representatives." Id., at 47, 106 S. Ct. at 2764.

The First Circuit's decision on appeal affirmed this court's finding that the Hispanic voters of Holyoke constitute a sufficiently compact group to satisfy the first prong of Gingles. Uno, 72 F.3d at 986. No evidence or argument offered on remand suggests any error in this finding. It is therefore not necessary to address this segment of the tripartite analysis.

The question whether the Hispanic voters in Holyoke are politically cohesive is relatively easy. To determine cohesiveness "the inquiry is essentially whether the minority group has expressed clear political preferences that are distinct from those of the majority." Sanchez v. Colorado, 97 F.3d 1303, 1311 (10th Cir.1996), petition for cert. filed, 65 U.S.L.W. 3648 (U.S. Mar. 13, 1997) (No. 96-1452), quoting Gomez v. City of Watsonville, 863 F.2d 1407, 1415 (9th Cir.1988), cert. denied, 489 U.S. 1080, 109 S. Ct. 1534, 103 L. Ed. 2d 839 (1989). A court will judge political cohesiveness by examining the "voting preferences expressed in actual elections" using the statistical evidence submitted by the parties. Id.

The majority of elections in the past twelve years in Holyoke have witnessed Hispanic voters expressing political preferences clearly distinct from the non-Hispanic majority. This difference is especially pronounced in elections where Hispanic candidates have run at-large, but emerges to some degree in ward elections as well.

In 1983, two Hispanic candidates, Juan Cruz and Joseph Gomez, ran at-large for City Councilor. The undisputed results of the experts' analysis of this election show Cruz ranking first and Gomez third with Hispanic voters, and fifteenth and sixteenth (out of a field of sixteen, running for eight slots) among non-Hispanic whites. Robert Little, an African-American, and Victor Provost, a non-Hispanic white, ranked second and fourth among Hispanics and thirteenth and twelfth among non-Hispanic whites.

In 1985, no Hispanic ran for an at-large seat. However, in Ward 2 an Hispanic, Juan Cruz, ran for City Councilor against a non-Hispanic white, Arthur Beaulieu, and Betty Medina Lichtenstein ran for School Committee against a non-Hispanic white, Elaine Pluta.

This is a useful point to address the First Circuit's question about the relevancy of ward elections to the central question in this case: the existence, vel non, of Hispanic vote dilution in the City Council at large elections. Where a ward is divided into precincts, and where these precincts contain sharply different percentages of Hispanic and non-Hispanic voters, then an analysis of the voting patterns *519 within a ward, compared by precinct, may tell the court something about racial polarization in the voting overall, in much the same way that comparison of differently configured wards will convey something about voting patterns in the at-large context. While the inferences drawn may not be conclusive, they constitute chips in the overall mosaic. Plaintiffs and defendants agree (for different reasons) that consideration of ward elections is appropriate even in a lawsuit that only challenges an at-large system.

In 1985 Ward 2 contained three precincts: 2A with a 10% Hispanic voting age population (VAP); 2B with an 80% Hispanic VAP, and 2C with a 77% VAP. Betty Medina Lichtenstein received 30% of the vote in 2A, 80% in 2B and 66% in 2C. Plaintiffs' expert reasonably concluded, and the court finds, that these numbers demonstrate a significant Hispanic preference, i.e., Hispanic political cohesion behind Medina Lichtenstein, who won her election and became the first Hispanic elected official in the history of Massachusetts. Cruz lost his election but fared better in the Hispanic precincts. Precinct analysis during the decade 1983-93 tended to show, though not always with perfect consistency, that precincts with higher concentrations of Hispanic voters supported Hispanic candidates, while those with lower concentrations did not. In this way, well exemplified by Medina Lichtenstein in 1985, ward results tended more often than not to corroborate evidence of Hispanic political cohesion emerging in at-large elections.

In 1987, Orlando Isaza, an Hispanic, ran at-large for the City Council. Both sides' experts agree that he ranked first in preference among Hispanic voters (out of a field of fourteen), and eleventh among non-Hispanic white voters, and lost the election. Only five other candidates received even measurable support from Hispanic voters, all at levels far lower than Isaza. Other evidence from this election, discussed below, bears on the more difficult question of whether white bloc voting or some other cause explains Isaza's loss. On the straightforward question of Hispanic political cohesion, however, 1987 provides yet another example, in the at-large context, of a clearly expressed Hispanic voter preference distinct from that of the majority.

In 1989, no Hispanic ran at large, and the ward elections tended to show evidence rebutting the claim of Hispanic voter cohesion. Elaine Pluta, a non-Hispanic, defeated an Hispanic candidate in heavily Hispanic Ward 2 in the School Committee race. Similarly, Lillian Santiago-Garcia, an Hispanic, lost her bid to become a School Committee member to a non-Hispanic incumbent in Ward 1, also predominantly Hispanic. Pluta's candidacy, in particular, largely erased the distinction between Hispanic and non-Hispanic voter preferences in her race.

In 1991, Pluta ran at-large for the City Council, ranking first among Hispanics and fifth among non-Hispanic whites and won a seat on the Council, again demonstrating her popularity across the board among voters in Holyoke. An Hispanic at-large candidate, Rodriguez-Ortiz, was ranked second among Hispanic voters, but fifteenth out of sixteen among non-Hispanics. Although the candidacy of Rodriguez-Ortiz failed to inspire significant Hispanic turnout, he was ranked high among Hispanics who did go to the polls. His candidacy offers some evidence that Hispanic voters tended to cohere behind Hispanic candidates, in contrast to non-Hispanic voters.

Hispanic voter cohesion was further reflected in the results of the 1991 voter referendum, with a solid majority of Hispanic voters favoring a tax override to support the schools and a solid majority of non-Hispanic whites opposed.

In 1993, Rodriguez-Ortiz again ran at-large for the City Council, coming in first among Hispanic voters in a field of twelve candidates. The only other candidate to receive measurable support from Hispanics was Arlela Bowie. Among non-Hispanic white voters Bowie ranked twelfth, Rodriguez-Ortiz eleventh. While the low Hispanic turnout caused experts from both sides to treat as less significant the evidence regarding racial polarization from these election results, the rankings do show that the few Hispanics who voted in the 1993 at-large election (2% of the Hispanic VAP) continued to express a preference for a candidate manifestly *520 not favored by the majority. At the same time, however, the returns from Ward 4 reflected a contrary tendency. With a 28% Hispanic VAP, this ward returned Joseph McGivern over an Hispanic candidate, with both Hispanics and non-Hispanics supporting him.

The 1995 election offered the strong evidence of a pattern of Hispanic voter political cohesion. Alejandro Sanchez, Jr. ran at-large for the City Council and ranked first among Hispanic voters, receiving support from virtually every Hispanic who voted, but came in dead last, tenth out of a field of ten, among non-Hispanic voters.

Sanchez was called by defendants to testify at the evidentiary hearing on remand and emphatically rejected any suggestion that his loss stemmed from white bloc voting against him. As will be seen, the court agrees. But on the question of political cohesion it simply cannot be disputed that Hispanics who voted supported Sanchez to a degree markedly greater than non-Hispanics voters.

In sum, Cruz, Gomez, Isaza, Rodriguez-Ortiz, and Sanchez, all Hispanic candidates for at-large City Council seats in the twelve year period, polled at or near the top among Hispanic voters, and all ranked at or near the bottom among non-Hispanic voters. Pluta, the lone significant at-large crossover candidate, was non-Hispanic. Precinct analyses in the ward elections usually (though not always) tended to confirm this pattern of contrast between Hispanic and non-Hispanic voting preferences. The 1991 referendum provided another thread in the same skein.

The Hispanic electorate in Holyoke "has expressed clear political preferences that are distinct from those of the majority." Sanchez v. Colorado, 97 F.3d at 1311. The court therefore finds, on reconsideration, that plaintiffs have carried their burden of showing by a fair preponderance of the evidence that Hispanic voters in Holyoke are politically cohesive.

Defendants offer in essence two reasons to rebut this finding, neither one persuasive. First, they point to the fact that Hispanic voters have supported a few non-Hispanic candidates who went on to be elected to the City Council at-large. The most prominent example, of course, is Elaine Pluta. Other non-Hispanic at-large candidates have on occasion garnered some measurable support from Hispanic voters and won a place on the Council. With the exception of Pluta, however, Hispanic support of winning non-Hispanic at-large candidates has generally been weak. This evidence is simply not sufficient to overcome the more powerful contrary evidence of cohesion.

Second, defendants have attempted, throughout the litigation, to use their "turnout" argument to undermine the plaintiffs' claim of political cohesion. They continue to argue that the much lower turnout rate of Hispanic voters, in comparison to non-Hispanics, signifies a failure on the part of the Hispanic community to cohere behind Hispanic-preferred candidates. This argument received very careful consideration at the first trial and was rejected. Hispanic turnout rates have varied from election to election, as the court noted, but were sufficient when viewed over the course of the decade to demonstrate a clear pattern of political cohesion. Moreover, as the court has noted, Hispanic voters who did go to the polls in this period tended overwhelmingly to support Hispanic candidates. These Hispanic-preferred candidates consistently received support from much smaller percentages of the non-Hispanic majority and as a result ended up losing, usually at or near the bottom of the pack.

Finally, the court found numerous socio-economic factors, and a number of specific obstacles to Hispanic political participation during the 1980's, that offered an explanation, at least in large part, for the depressed Hispanic turnout. On appeal the First Circuit affirmed this court's conclusion, stating that "the evidence of low Hispanic turnout does not undercut the court's ultimate conclusion that the plaintiffs established minority political cohesion." Uno, 72 F.3d at 987.

In sum, and at the risk of repetition, the stronger evidence both at the original trial and on remand has proved by a fair preponderance of the evidence that Hispanic voters tend to prefer candidates usually rejected by most of the non-Hispanic majority. Hispanic *521 voters are sufficiently cohesive politically to satisfy the second Gingles criterion.

A showing of compactness demonstrates that a minority voting group has the numbers to elect its preferred candidate; a showing of political cohesion indicates that the group has the will to do so. Once "the potential to elect a representative of its own choice" is shown, proof of the final factor white bloc votingmust be sufficient to establish that the election system "thwarts a distinctive minority vote by submerging it in a larger white voting population." Growe v. Emison, 507 U.S. 25, 40, 113 S. Ct. 1075, 1084, 122 L. Ed. 2d 388 (1993).

As the First Circuit pointed out, plaintiffs cannot prevail merely by presenting persuasive evidence of significant non-Hispanic white bloc voting against their preferred candidates. That is only the first step in satisfying the third Gingles criterion. They must also show that this pattern of voting in combination with the at-large voting system actually caused the defeat of Hispanic-preferred candidates at the polls. The court must scrutinize carefully "the countervailing evidence of other causative agents" that might explain Hispanic lack of success before concluding that "the record sustains a claim that racial politicsspecifically, the interaction of race and the electoral systemhave resulted in significantly diminished opportunities for participation in elective government." Uno, 72 F.3d at 985.

To make this assessment, the court must turn to the difficult question of the impact of white bloc voting in the at-large City Council elections in the period 1983-95. As will be seen, the evidence is uncontrovertible that such bloc voting has frequently taken place. The proof is much weaker on the question whether this bloc voting has actually been the cause of the lack of success of Hispanic candidates at the polls. In the final analysis, given the results of the 1995 election and the Sanchez/McGivern testimony, the evidence now before the court fails to demonstrate a sufficiently clear causative effect of white bloc voting in the current-day political reality in Holyoke to justify a finding that the at-large component of Holyoke's City Council election scheme violates the Voting Rights Act. An analysis of the elections during the period 1983-95 brings this point home.

In 1983, as noted above, candidates Cruz and Gomez ranked at the top of Hispanic preferences, and at the bottom among non-Hispanics. Cruz, who fared best, was able to achieve only a roughly 20% crossover rate among non-Hispanic whitesthat is, eighty percent of the voters from non-Hispanic areas of Holyoke declined to support him. Gomez did even worse. The lack of non-Hispanic support in 1983 was easily measurable, since four of the city's seven wards were well over 90% non-Hispanic. Thus, white bloc voting occurred.

The actual impact of this pattern is less clear. These two first-time candidates ran in a city with a much smaller Hispanic population than today, and enjoyed a turnout of only around 10-11% of the Hispanic voting age population at that time. Cruz is estimated to have received only 300 votes out of a potential Hispanic VAP of approximately 2835. As a result of the smaller Hispanic population and the low turnout, almost 90% of the voters who did support Cruz were non-Hispanic. Thus, it is difficult to conclude, in this early election, whether white bloc voting actually caused Cruz' loss or if his defeat was simply a result of electoral politics.

In 1985, no Hispanic ran at-large, but Betty Medina Lichtenstein's first-ever victory for an Hispanic candidate pulled out 22.8% of the Hispanic voting age population in Holyoke, the largest percentage in this period. Precinct analysis in Ward 2 confirmed, as the court noted above, that significant white bloc voting against her occurred, but it was obviously not sufficient to doom her candidacy.

This is a good point to address an argument pressed by the defendants regarding the concept of "legally significant racially polarized voting." In defendants' view, racially polarized votingi.e., voting patterns characterized by minority voter cohesion behind preferred candidates and consistent white bloc voting in opposition to them becomes "significant" only when it can be *522 shown to be the sole or predominant cause of the minority-preferred candidate's loss. Where a minority-preferred candidate such as Medina Lichtenstein wins, or where other causes may explain the loss as perhaps in the case of Cruz, then, defendants say, racially polarized voting, if it occurs, is not "significant."

This construction of the phrase "legally significant racially polarized voting" confuses the task of the court in weighing the evidence with the task of the plaintiffs in carrying their ultimate burden of proof. It is self evident that majority bloc voting against minority-preferred candidates in any election, regardless of the outcome or its cause, is "significant" in the broad-ranging analysis the court must perform in a Voting Rights case. Racially polarized voting of this sort has occurred, to some degree, in almost every election year examined, either in at-large or ward elections, and that fact is very significant. Ultimately, of course, the plaintiffs must convince the court that Hispanic voters in Holyoke were "denied meaningful access to the political system on account of race" and not because of some other race neutral factor. Uno, 72 F.3d at 983. (Emphasis supplied.) But this ultimate burden need not be carried with regard to every election, at risk of rendering these individual pieces of evidence "insignificant." See, Sanchez v. Colorado, 97 F.3d at 1312 (legally significant bloc voting is "a sliding scale that varies with the district and a variety of factual circumstances").

This court's unfortunate lack of precision on this point in its earlier memorandum seems to have been a factor in the First Circuit's decision to remand. Judge Selya wrote "the court's finding that so few elections [three or four out of fifteen] exhibited telltale signs of minority cohesion and/or white bloc voting seems to be tantamount to a finding that those characteristics were absent from approximately two thirds of the analyzed elections." Uno, 72 F.3d at 989. In fact, some degree of Hispanic voter cohesion and opposing white bloc voting was present in almost all of the elections analyzed. In other words, there were telltale signs aplenty, and this court should have made that point more clearly. The existence of some degree of racially polarized voting in most elections has been shown in the discussion of political cohesion above and will be amplified with regard to white bloc voting in this portion of the memorandum.

The complicating factor in this analysis, however, is that few elections in Holyoke have offered clear examples of white bloc voting working in combination with the at-large system actually to affect the outcome. The Isaza campaign ten years ago is probably plaintiff's strongest case in point. As noted above, the precarious state of the evidence regarding the causative impact of white bloc voting has allowed the results of the 1995 election and the Sanchez/McGivern testimony to tip the scales on remand.

Returning to a review of white block voting in the 1983-95 period, in 1987 Orlando Isaza, an Hispanic, ran at-large for the City Council. Well qualified and well supported within Holyoke, he was by far the top vote getter among the roughly 11% of the Hispanic VAP that went to the polls (credibly estimated to be 747 out of 6,793 potential voters). Nevertheless he was able to attract only a 26.6% crossover vote from the non-Hispanic majority, which turned out in much greater numbers, and his eleventh place finish overall spelled defeat. Although other factors probably played a part in Isaza's defeathe had lived in Holyoke for only a year, this was his first and only foray into electoral politics and he did not conduct a perfect campaign white bloc voting against him was more likely than not the determinative factor in his loss.^[2] Defendants' attempt to suggest that two other "Hispanic-preferred" candidates, Whelihan and Murphy, won in 1987 is unpersuasive. Whelihan probably received about 88 total Hispanic votes and Murphy 60. No other candidate who succeeded in the at-large election received even measurable Hispanic support. Thus, 1987 offers solid proof both of Hispanic political cohesion and white bloc voting walling out a qualified Hispanic at-large candidate.

*523 The evidence in 1989 tended to rebut racial polarization as a causative factor in election outcomes. No Hispanic ran in the at-large election, and non-Hispanics won races against Hispanics in heavily Hispanic Wards 1 and 2. Precinct returns tended to show a degree of Hispanic cohesion and non-Hispanic white bloc voting, but the outcome of the election obviously did not turn on these forces.

The 1991 election again offered mixed evidence. Elaine Pluta, the first choice of Hispanics in the at-large election, won. Diosdado Lopez, on his second try, won the City Council seat for Ward 2 against a non-Hispanic. Thus, for the first time, Hispanic candidates of choice occupied two of the fifteen seats on the City Council. At the same time, Lopez polled much more strongly in the more heavily Hispanic precinct, and Rodriguez-Ortiz, the second most preferred at-large candidate among Hispanic voters, came in near the bottom overall. As noted in the original decision, the general unpopularity of Rodriguez-Ortiz' political views, among Hispanics as well as non-Hispanics, was the best explanation for his defeat. Thus, while white bloc voting occurred, it was not outcome determinative.

In 1993 Rodriguez-Ortiz again ran at-large, coming in first among the tiny 2% of the Hispanic VAP that got to the polls, and lost. Again, while the Hispanic voters who turned out ranked him high, evidencing political cohesion, white voters rejected him en masse, evidencing white bloc voting. For the reasons stated above, however, it would be unfair to suggest that racially polarized voting caused Rodriguez-Ortiz' defeat. Based on the witnesses' testimony, he might well have lost the Hispanic vote if turnout had been much greater, simply because many Hispanics would not have supported him. In Ward 2 Hispanic candidates for City Council and School Committee were unopposed. Ward 1, though heavily Hispanic, returned non-Hispanics to both seats, with precinct returns again reflecting a degree of Hispanic political cohesion and white bloc voting. Again, racially polarized voting played no causative role in these elections.

1993 saw the first election in the reconfigured Ward 4, now with a 28% Hispanic VAP. The winner, Joseph McGivern, had been endorsed by one of the Hispanic plaintiff organizations in this case, Hispanic Citizens United for Holyoke (CLUH) in his 1991 mayoral campaign. McGivern's wife is bi-lingual in Spanish, a former ESL teacher, and a current primary school teacher. McGivern testified credibly that he won his 1993 campaign in Ward 4 against an opposing Hispanic candidate by actively vying for Hispanic votes and offering to serve the interests of the Hispanic community. He garnered 74.4% of the Hispanic vote in Precinct 4A, which is 45.2% Hispanic, against an Hispanic opponent. To this extent, 1993 saw the emergence of Ward 4 as a possible so-called "influence district." In such a district, no viable candidate can ignore minority interests, because of the minority's strong influence at the ballot box. While the 1993 election, the first for this newly drawn ward, did not provide conclusive proof on this point, evidence of a tendency appeared.

This 1983-93 election-based evidence was before the court at the time of the first decision. It showed consistent political cohesion by Hispanic voters behind certain preferred candidates and equally consistent white bloc voting against them. With only one exception in ten yearsa period that saw eighty winning at-large candidaciesno victorious at-large City Council candidate had ever received significant Hispanic support. Although the court could rarely say with certainty that white bloc voting caused a particular candidate's defeat, it could not be denied that no Hispanic candidate had ever been able to attract more than about one-quarter of the non-Hispanic vote, or even come close to winning at-large. The one election where the hypothesis of racially polarized voting could best be tested, the 1987 Isaza race, saw white bloc voting shut the door on a qualified, well-supported Hispanic candidate. Ward elections tended to corroborate the pattern of racially polarized voting.

Added to this was the evidence of the "totality of the circumstances," which need only be briefly touched on here: the discrimination suffered by Hispanics in housing, employment *524 and health; the significant unresponsiveness during the 1980's by elected officials to the needs of the Hispanic community; the barely veiled racial appeals that colored the 1987 mayoral election; and the fumbling and sometimes obstructionist official attitudes during this decade towards assisting Hispanic citizens to register and vote.

In the final analysis, as this court acknowledged, the balance of all the factors made the original decision a close call. The absence of any significant number of elections where white bloc voting clearly affected the outcome, the paucity of strong Hispanic at-large candidates, and the relatively recent arrival of Hispanic citizens in any numbers, weighed against the plaintiffs. Given that the test, however, is whether the challenged election system results in significantly diminished opportunity for minority votersand not whether the system utterly shuts minority candidates outthe court returned its verdict for the plaintiffs regarding the City Council. But the balance of proof, as will be seen, was sufficiently precarious to tip the other way when the new evidence on remand was placed on the scales.

In 1995, four Hispanics ran for City Council seats: Alejandro Sanchez, at-large; Lopez, unopposed in Ward 2; George Cruz against James McDermott, the incumbent, in Ward 1; and Harry Rodriguez-Ortiz against Richard Welch in Ward 4. Lopez and Cruz won, and Sanchez barely missed.

Sanchez, an Hispanic of Puerto Rican background, moved to Holyoke when he was four years old and has lived in the city 26 years. He is a 1981 graduate of Holyoke High School, received an associate's degree from Holyoke Community College, a B.A. in criminal justice from Westfield State College and an M.A. in the same field from Anna Maria College in 1992. He is married with three children, all of whom attend Holyoke public schools, and is employed as a police officer in nearby Springfield. For six years he has been a D.A.R.E. officer, teaching a fifteen week anti-drug, anti-violence program to fifth grade students.

In 1995 Sanchez had never run for public office, and his previous experience with Holyoke election campaigns was insubstantial. He stated that he wished to run because "being a resident of the city, I feel that the city has done things for me, for my wife, and my family. We still live there and I wanted to give something back, work for the city." Sanchez Tr. at 11.

Sanchez testified that it was the mistakes in his campaign that caused him to lose. As a newcomer he had low name recognition. His decision to announce his candidacy very early meant he was "old news" when the election campaign heated up. Perhaps most importantly, his election committee was inexperienced and failed to canvass Holyoke's many elderly apartment towers, until Sanchez was advised by other, non-Hispanic candidates to do so, by which time it was too late to address this crucial voter group effectively.

Despite his disappointment on his first try for office, Sanchez testified that "any good candidate can win a seat on the council." Tr. at 36. He planned to try again in 1997.

In travelling throughout the city, Sanchez never encountered any racial antagonism towards him or his candidacy and would not have tolerated it if he had. In fact, he testified that in his personal contacts voters were "very receptive." Tr. at 24. Sanchez was assisted by a number of non-Hispanic candidates, including Elaine Pluta, Dan Burns, Jim Jackowski, Kris Kos-Lecca and Joseph McGivern. He attended fundraisers of non-Hispanic candidates and several non-Hispanic candidates attended his. Sanchez was also assisted by successful mayoral candidate Szostkiewicz, who provided him with computer information about registered voters.

Plaintiffs' expert estimated Hispanic voter turnout to be 8.6% of VAP in 1995; defendants' expert put the estimate at 2%. The 8.6% estimate is certainly high, since it assumes that every single Hispanic voter cast a "bullet" ballot for Sanchezthat is, cast a single vote only for Sanchez without using any of the available seven other votes to support any other candidate. Sanchez' testimony *525 was that no Hispanic voters had ever even discussed bullet voting and he never encouraged it. Thus, the court finds that Hispanic voter turnout was probably closer to 2% than 8.6%. Even assuming that 8.6% of the Hispanic VAP turned out, this percentage would total only 584 votes out of a potential of 6,793 voters.

Nevertheless, even with a neophyte's mistakes and low Hispanic turnout, Sanchez missed the winning eighth seat by only 276 votes (5,029 to 5,304 votes), by far the closest race of any Hispanic candidate in the twelve-year period. He trailed Burns, the eighth place finisher, by only 4.4%, about half the margin of any previous Hispanic candidate. In fact, the Sanchez showing is even better than it appears, because he trailed the ninth place finisher, Jackowski, by only 169 votes, and Jackowski eventually took a seat on the Council as next in line when Pluta resigned to take a position with the Szostkiewicz administration.^[3]

Perhaps most significantly, Sanchez enjoyed at least a 42% crossover vote from non-Hispanic voters, much more than any previous Hispanic candidate. Thus, although he was the first choice among Hispanic voters who went to the polls, and the least favored candidate by a small margin among non-Hispanics, he received over 88% of his total votes from non-Hispanics. With plaintiff's estimate of a 42% crossover vote from non-Hispanics, and census figures indicating that, as of 1990, 21.89% of the total Holyoke VAP was Hispanic, the court must conclude that racially polarized voting was not the cause of his loss, and that he or any energetic, qualified Hispanic candidate has a fair opportunity to run in 1997. Cf., Sanchez v. Colorado, 97 F.3d at 1317-18 (consistent "Anglo" crossover vote of only 10-20% and no Hispanic elected in fifty years, despite continuous high proportion of Hispanic VAP).

The court's conclusion regarding the significance of the 1995 at-large race should not be interpreted as suggesting that the election of one Hispanic-preferred candidate at-large, let alone the near miss of a single Hispanic candidate, is sufficient alone to defeat a claim under the Voting Rights Act. The success of a single minority-preferred candidate, standing by itself, "is not dispositive." N.A.A.C.P., Inc. v. City of Niagara Falls, 65 F.3d 1002, 1023 (2nd Cir.1995). It is the evidence revealed by the election, not just the result, that weighs most strongly in the court's analysis. If Sanchez' near miss could be explained by a large Hispanic turnout, thwarted by massive non-Hispanic bloc voting against him, then the same vote totals would have pointed to the opposite conclusion. The 1995 election proved, on the contrary, that very large numbers of non-Hispanic voters in Holyoke are quite willing to support an attractive, energetic Hispanic candidate. This evidence sharply undercuts the court's prior conclusion, based on the evidence presented in 1995, that it was "difficult to imagine" any Hispanic candidate "having a fair opportunity to win in any at-large election in Holyoke during this period (i.e., 1983-93)." 880 F. Supp. at 926.

In addition, George Cruz' 1995 victory in Ward 1, in his first election bid against a non-Hispanic white incumbent, was also a testament to increasing Hispanic power at the polls. With Lopez' seat in Ward 2, two of the fifteen counsel seats may now be said to be "controlled" by the Hispanic electorate.

Contrary to plaintiffs' argument, developments after the election cast no cloud over this progress. The decision by the City Council to appoint Jackowski rather than Sanchez to the seat vacated by Pluta was based on Jackowski's finishing position in the at-large election, not on any discrimination against Sanchez due to his ethnicity. Sanchez acknowledged that Jackowski was a good candidate; the two of them had campaigned together in Hispanic sections of the city, and Jackowski distributed Spanish-language campaign literature. Moreover, the decision to appoint Henry Jennings, an African American, rather than Juan Rivera to the Ward 1 seat vacated by Cruz when Cruz resigned due to a conflict of interest was justified by Jennings' comparatively lengthy residence in Ward 1 and not animated by racial prejudice. In the past the City Council has not hesitated to appoint Hispanics to *526 vacated seats on the Council and School Committee, as it did in 1992 with Wilfredo Echeverrea and in 1994 with Medina Lichtenstein.

The past two years have seen some evolution in the evidence bearing on the mix of factors enumerated in the Senate Report on the Voting Rights Act under the rubric of the "totality of the circumstances." Many of these factors favored the defendants even at the original trial, as the court noted in its decision at that time.

The first consideration requires the court to consider the extent of any history of official discrimination that "touched the right of the members of the minority group to register, to vote, or otherwise to participate in the democratic process." Id., at 915. The 1995 election witnessed the complete absence of election-related problems that plagued elections in the 1980's. The ballot was bilingual in accordance with federal law; there was a sufficient number of bilingual poll workers; there were no problems with Hispanic voter registration efforts, and there was no evidence of Hispanic voters improperly removed from voting rolls.

The second factor is "the extent to which voting in the elections is racially polarized." Id. This issue has been thoroughly discussed already and need not be addressed in detail now. The fact is that minority political cohesion and non-minority bloc voting have occurred in Holyoke in most elections in the past twelve years. The significance of these factors as causes of any substantially reduced opportunity for Hispanic citizens to participate in the electoral process is less clear. Given the most recent election, the court must conclude that, in the context of Holyoke's hybrid ward and at-large voting system for the City Council, racially polarized voting as a cause of significantly diminished opportunity for Hispanic voters has not been adequately proved to exist in the city's current political scene.

The third factor is the extent to which the election system has used "unusually large election districts ... or other voting practices or procedures that may enhance the opportunity for discrimination against the minority group." Id. This factor favors defendants. No such practices or procedures exist in Holyoke. Cf., Goosby v. Town Bd. of the Town of Hempstead, N.Y., 956 F. Supp. 326 (E.D.N.Y.1997) (town with population of 725,639 governed by six-member Town Board, all elected at-large).

Regarding the fourth factor, the slating process, no evidence suggests any prejudice of this sort to the plaintiffs.

The fifth consideration points the court to the effects of discrimination in "such areas as education, employment and health, which hinder [a minority's] ability to participate effectively in the political process." Uno, 880 F. Supp. at 915. This factor favors plaintiffs. Hispanics in Holyoke continue to suffer the effects of discrimination to a significant degree in all these areas. No evidence has been submitted on remand tending to rebut the findings of the court on this point in its original decision.

With the sixth factor the court must direct its attention to the existence of overt or subtle racial appeals in elections. The last election marred by such indirect appeals occurred in 1987. The 1995 election contained no such elements.

The court must also examine, as the seventh factor, the extent to which Hispanics have been elected to office. On this point improvement has occurred.1995 saw the election of George Cruz in Ward 1 and the continued election in Ward 2 of Hispanic candidates. Moreover, the configured Ward 4 realistically permits election only of a candidate sensitive to the interests of Hispanic citizens.

The significant lack of responsiveness of Holyoke officials to the needs of the Hispanic community, the eighth factor, was evidenced in the 1980's and is significantly diminishing in the 1990's, if not disappearing. The administration of the new Mayor has witnessed a greatly increased effort to recruit Hispanic officials, include Hispanic viewpoints, and address the interests of all the citizens of Holyoke, Hispanic and non-Hispanic.

*527 Finally, the policy underlying the at-large component is not tenuous. The presence of at-large representatives on the City Council insures that the perspective of the community as a whole, and not just one neighborhood, is well represented. This fact was recognized even in the court's original remedial order.

Other considerations beyond the nine enumerated also play their part in the court's expansive evaluation. As noted in the original decision, the Hispanic community is relatively new in Holyoke, growing from 6,165 out of a population of 44,678 in 1980, to 13,573 out of 43,704 in 1990. The increasing Hispanic political strength in the city can now be more easily seen as moving in tandem with the growing population numbers.

In addition, in 1995, the Ward 4 election provided further indication that it was emerging as an influence district. The winner, Welch (a former Holyoke School Committee chairperson) did well even in the more heavily Hispanic: precinct, and Rodriguez-Ortiz continued to be unable to attract significant Hispanic voter turnout. Significantly, Sanchez' at-large bid got just under 50% of the Ward 4 votes, though its Hispanic VAP was only 28%. The picture in this ward is still not crystal clear, but it is hard to imagine any viable candidate of whatever ethnicity ignoring the interests of its large Hispanic population.

Joseph McGivern, current President of the Holyoke City Council, active for twenty years in Holyoke politics, gave the court a detailed picture of the city's electoral mechanism. He testified to the power of incumbency in the city's elections, the importance of name recognition, the significance of the elderly towers, and the difficulties facing first-time candidates. This testimony confirmed the court's finding that factors other than race led to Sanchez' near miss in 1995.

Beyond this, McGivern confirmed the atmosphere of increased inclusiveness in Holyoke's election scene: candidates attending Sanchez fundraisers, the distribution by non-Hispanic candidates of Spanish-language campaign leaflets, and the recent appointment of significant numbers of Hispanics to official positions in Holyokeparticularly to the Holyoke Master Plan Committee, set up by he new Mayor to review issues relative to zoning, development and growth within Holyoke.^[4]

No election since 1987, and perhaps none in Holyoke's history, has provided as clear a test of the fairness of the Holyoke election system as the 1995 at-large City Council race. Viewed in the totality of the circumstances, this election demonstrated, and the court finds, that in the current political reality of Holyoke, Hispanic citizens are not being denied "meaningful access to the political system on account of race." Uno, 72 F.3d at 983.

This court has no illusions about what one writer has called "the lumpish, Jabba-the-Hutt immobility of racial prejudice in this country."^[5] Certainly, it exists in Holyoke as elsewhere. Equally certainly, however, it is much less a factor now than in the past, at least in Holyoke's political life. Some percentage of non-Hispanic white voters will continue to oppose a candidate based merely on his or her ethnicity or surname, but many others are receptive to Hispanic candidates and willing to support them. This is good news for our democracy. Viewing the evidence as a whole, this court cannot say that the interaction of race and the electoral system, as of 1997, now results in "significantly diminished opportunities for minority participation in elective government." Uno, 72 F.3d at 983-84. The court will therefore order entry of judgment for the defendants on the remaining claim.^[6]

A separate order will issue.

For the reasons set forth in the accompanying Memorandum, the court hereby orders entry of judgment in favor of defendants on all remaining counts. The clerk will enter an appropriate judgment.

[1] The portion of the lawsuit directed against Holyoke's ward configuration was settled early in the litigation.

[2] It is notable, however, that another loser in that at-large race, Szostkiewicz (who placed tenth), won on his second try in 1989 and is now mayor of Holyoke.

[3] Sanchez trailed the seventh-place finisher Kos-Lecca by less than 300 votes.

[4] Following the election, Alejandro Sanchez was appointed by Mayor Szostkiewicz chairman of Holyoke's Fire Commission, and his brother Melvin, Sanchez' campaign chairman, a member of the City's Planning Board.

[5] Roger Angell, "Box Score: Has Baseball Fulfilled Jackie Robinson's Promise?" The New Yorker, April 14, 1997, at 6.

[6] The court has considered an approach to this case used by the First Circuit in Black Voters v. McDonough, 565 F.2d 1 (1st Cir.1977). There, finding that the underlying decision was "close," the Court of Appeals affirmed the lower court, but "provisionally and without prejudice to plaintiffs' right to reopen their claim in the future in light of circumstances as they may then appear." Id. at 7. Judge Coffin directed the district court to retain jurisdiction and authorized it to hold supplemental hearings after a year on an amended petition, if plaintiffs filed one. In addressing the new petition, he stated that the district judge "should take account of whether the situation in relevant respects has improved or has worsened so to alter the present calculus." Id.

This court has decided not to take this approach for three reasons. First, in the last twenty years other courts have not adopted this rather unusual judicial mechanism. Second, given the important issues of federalism and separation of powers implicated by Voting Rights suits, maintenance of a shadow judicial presence over democratic processes should be reserved for extreme cases. Third, plaintiffs, in any event, have the right to re-file this lawsuit at any time, if more compelling evidence of a violation of the Act appears.