United States v. Alamosa County, Colo.

306 F.Supp.2d 1016 (2004)

THE UNITED STATES of America, Plaintiff,
v.
ALAMOSA COUNTY, COLORADO; Alamosa County Board of Commissioners; Robert Zimmerman; Darius Allen; Charlotte Bobicki; and Holly Z. Lowder, Defendants.

No. 01-MK-2275(BNB).

United States District Court, D. Colorado.

February 5, 2004.

Joseph D. Rich, Ralph F. Boyd, Jr., Tricia G. Jefferson, U.S. Department of Justice, Civil Rights Division, Washington, DC, for plaintiff.

James Scott Detamore, William Perry Pendley, Mountain States Legal Foundation, Lakewood, CO, David Arthur Rooney, Lester, Sigmond, Rooney & Schwiesow, Alamosa, CO, for defendants.

CORRECTED MEMORANDUM OPINION AND ORDER

(This Corrected Memorandum Opinion and Order is issued solely to correct a typographical error at the number of counties listed on page 2, III.A., paragraph 1.)

KRIEGER, District Judge.

THIS MATTER is before the Court following a nine-day bench trial. At trial, the Plaintiff, the United States of America, was represented by Thomas Reed, Tricia Jefferson, and Edward Keefe from the United States Department of Justice. The Defendants, Alamosa County, the Alamosa County Board of Commissioners, Robert Zimmerman, Darius Allen, Charlotte Bobicki and Holly Z. Lowder, were represented at trial by J. Scott Detamore and Joseph Becker of the Mountain States Legal Foundation. Having considered the evidence presented and the arguments of counsel, the Court makes the following findings of fact and conclusions of law.

I. JURISDICTION

The Court has subject matter jurisdiction pursuant to 28 U.S.C. § 1331.

II. ISSUES PRESENTED

The United States of America ("the Government") brought this action against Alamosa County, the Alamosa County Board of Commissioners, its members,^[1] and the Alamosa County Clerk^[2] (collectively referred to as "the Defendants"). The Government seeks a determination that the at-large method for electing county commissioners in Alamosa County violates Section 2 of the Voting Rights Act of 1965, as amended, 42 U.S.C. §§ 1973 and 1973j(d), and therefore should be set aside. The Government claims that this election method diminishes the ability of Alamosa County's Hispanic^[3] residents to participate in the electoral process and elect representatives of their choice, thus resulting in impermissible vote dilution. The Defendants respond that either Section 2 is unconstitutional or the at-large method of electing county commissioners does not violate Section 2

III. FINDINGS OF FACT^[4]

A. The Process for Electing County Commissioners

Alamosa County is one of 64 counties in the State of Colorado. Like all counties in Colorado, its Board of County Commissioners ("the Board") governs and administers county affairs. The current method of electing commissioners to the Board is the same as in other Colorado counties and has been in place since 1913, when Alamosa County was formed. The process is specified by state statute, § 30-10-306, C.R.S., which requires that every county be divided into compact residency districts nearly equal in population. Members of the Board serve four-year, staggered terms.

Candidates for county commissioner are first nominated either by petition^[5] or through the caucus process, then selected in a partisan primary election. The top finisher for each residency district in a primary election advances to the general election. In the general election, candidates for each residency district are elected at-large.

Based on its population, Alamosa County has three residency districts,^[6] and accordingly, the Board consists of three members, each of whom qualifies to run from one of the three districts. The benefit of electing county commissioners at-large, particularly in a rural/urban county such as Alamosa, is that every commissioner is accountable to residents of the entire county rather than solely to neighbors in his or her residency district.

Currently, there is no residency district in Alamosa County in which Hispanic residents constitute a majority of the voting age population or of registered voters. However, the parties agree that a district in which Hispanic residents comprise at least 60% of the voting age population could be drawn.

B. Alamosa County Geography, History, and Demography

Located in south central Colorado, Alamosa County is one of six counties in the San Luis Valley. The largest part of the county is rural with farming and ranching as major industries The county has one urban center, the City of Alamosa, where most businesses, government offices and Adams State College are located.

Persons of Hispanic and Anglo descent have resided in the county for more than a century. Indeed, over 70% of Alamosa County's current Hispanic residents were born in Colorado. Many Hispanic and Anglo families have lived either in Alamosa County or elsewhere in the San Luis Valley for numerous generations, intermarrying at a higher rate than the national average.

According to the 2000 Census, the total population of Alamosa County was 14,966. Of this population, 54% (8,089) self-identified as White (but not Latino or Hispanic), 41.1% (6,197) self-identified as Hispanic, and 4.5% (680) self-identified as Black or African-American, American Indian, Alaska Native, Asian, Native Hawaiian, Pacific Islander, or other designation. Of the voting age population (age 18 and older), these groups represent 57.9%, 37.6% and 4.4%, respectively. By comparing the total percentage of the population to the percentage of the voting age population, it is apparent that most self-identified Whites are of voting age, but in the Hispanic group a greater proportion are under age 18.

Historically, discriminatory practices existed in Alamosa County, particularly during the first half of the twentieth century. These practices created or exacerbated socioeconomic disparities between Hispanic and Anglo residents. Official action and social activism, particularly during the last fifty years, however, have reduced the discriminatory climate in the county. This has improved educational and socioeconomic conditions for Hispanic residents and contributed to increasing diversity within the Hispanic population.

Official school segregation was ended by court order in 1914, but unofficial discrimination against Hispanic students continued through the 1950's.^[7] The "Alamosa Plan," adopted in 1980, was intended to address de facto discrimination. Now, the school superintendent and at least 30% of the professional staff in Alamosa County schools are Hispanic. As of 2000, fewer Hispanic residents in Alamosa County had pursued higher education than non-Hispanic residents, but a substantial number of Hispanic residents over the age of 25 had obtained high school diplomas, attended college, or obtained graduate or professional degrees.^[8] There are some disparities in educational accomplishments between Anglo and Hispanic students, especially those in high school. In 2000, Hispanic third graders fell behind non-Hispanic third graders in CSAP scores, but since then, the gap has significantly narrowed.

Historically, Hispanic and Anglo communities were divided by the east/west railroad tracks in the City of Alamosa; most Anglo families lived north of the tracks and Hispanic families lived south of the tracks. Based upon their neighborhood affiliations, Hispanic and Anglo residents did not routinely socialize, attend the same churches or community organizations or do business at the same establishments. Hispanic residents now live, work, and own businesses both north and south of the demarcation line and there is extensive integration and association among Hispanic and Anglo residents both socially and in business.^[9] In addition, Hispanic residents comprise roughly 30% of the officers of the Alamosa County Sheriff's Office, 50% of Alamosa police officers, 25% of the staff at the District Attorney's Office, and 50% of the staff in the Twelfth Judicial District.

Integration, however, does not mean that every instance of discrimination has been eliminated. For example, some Hispanic residents complain of derogatory comments,^[10] but there is a diversity of perception in this regard.^[11]

If one compares Alamosa County Hispanic residents as a group to Anglo residents as a group, Hispanic residents earn less. However, among Hispanic residents, there is substantial diversity in income. According to the 2000 Census, 48.4% of Anglo households had an annual income greater than $35,000; 31.0% of households had an income greater than $50,000. In comparison, 32.1% of Hispanic households had an income greater than $35,000; 17.9% of households had an income greater than $50,000 At the other end of the spectrum, according to 2000 census figures, 15.6% of Anglo residents and 28.5% of Hispanic residents had incomes at or below the poverty level. In 2000, home ownership by Hispanic residents was slightly less than that of Anglo residents but 15 to 20% higher than that of Hispanic residents, nationally. Among Hispanic homeowners, 34.1% lived in high income areas, 38.0% lived in average income areas, and 16.8% lived in low income areas.^[12]

C. Election Results

Much of the evidence concerned election results during the past twenty-plus years. These elections are divided into two types, endogenous and exogenous. Endogenous elections are Alamosa county commissioner elections, both primary and general. An exogenous election is any other election in which Alamosa County residents voted.

1. Endogenous Primary Elections

With one exception, in every primary election since 1978 when an Hispanic candidate^[13] sought nomination by the Democratic party, he or she was successful.^[14] Hispanic candidates were unopposed in primary elections in 1972 (Pasqual Lee), 1978 (Walt Gomez), 1984 (Jorge Amaya and Lee Patrick Herrera), 1988 (Walt Gomez), and 1992 (Marguerite Salazar). In 1980, Ricky Sanchez defeated James Craft, and in 2002, Ernie Roybal defeated Al Greer.

2. Endogenous General Elections^[15]

Prior to 1984, no Hispanic candidate prevailed in a general election for county commissioner. Since then, however, Hispanic candidates have been elected three times in five elections. Jorge Amaya and Lee Patrick Herrera were elected in 1984; Frank Mestas was elected in 2002. In addition, Walt Gomez was appointed to the Board in 1987. Incumbents usually have been re-elected. After 1982, the only challenger who prevailed over an incumbent was Hispanic candidate Jorge Amaya. Since then, an Hispanic candidate has been elected every time there has been an open seat on the Board.

3. Exogenous Elections^[16]

County-wide offices such as sheriff, assessor, clerk and recorder, coroner, treasurer, or surveyor are also elected at-large, but unlike in the county commissioner elections, no Hispanic has been elected or appointed to a county-wide office since 1970.

State-wide elections have yielded mixed results. Hispanic Carlos Lucero was unsuccessful in his 1984 and 1990 bids for the United States Senate, and Hispanic Gil Romero was unsuccessful in a 1998 bid for the same office. However, Hispanic Ken Salazar was successfully elected in 1998 and 2002 as the State of Colorado Attorney General. Interestingly, all three candidates came from Alamosa County.

D. Statistical Analysis

Each party presented expert testimony analyzing election data of endogenous and exogenous elections since 1980. Dr. Richard Engstrom testified for the Government and Dr. Ronald Weber testified for the Defendants.

Both experts relied upon Spanish surname analysis to identify Hispanic voters.^[17] Spanish surname identification is an accepted means of identifying which voters are likely Hispanic, but it is an imperfect method, particularly in demographically mixed Alamosa County. Ordinarily, surname data analysis has an omission error rate (Hispanic persons misidentified as non-Hispanic) of 16.5% and a commission error rate (non-Hispanic persons misidentified as Hispanic) of 10.5%. The two error rates do not necessarily offset each other. Without consideration of Alamosa County history and demographics, reliance on Spanish surnames can result in a statistical error rate of between 5% and 27%. The degree of error is further compounded due to the long and extensive history of intermarriage among Hispanic and Anglo residents in Alamosa County. Self-identification is a more reliable means of determining ethnicity than surname analysis Thus, the experts' statistical conclusions (though probative) have reduced weight due to "surname bias".

1. Dr. Engstrom's Opinions^[18]

Dr. Richard Engstrom studied 22 elections and used three methods for measuring voter cohesiveness (the degree to which Hispanic voters voted for the same candidate) — homogenous precinct analysis, bivariate ecological regression analysis, and the King methodology. Homogeneous precinct analysis looks at votes cast in a precinct and generates a scatter plot chart showing the relationship, if any, between the ethnicity of the voter and the votes cast for a particular candidate. Bivariate ecological regression analysis plots a straight line through the points on the homogeneous precinct analysis scatter plot. The King methodology modifies the bivariate ecological regression analysis by not assuming linearity and does not allow estimates to exceed 100 percent or fall below zero. Of the three methods, Dr. Engstrom relied most heavily on the first two and made little reference to the King method.

Dr. Engstrom used Spanish surname counts found in election returns, turnout/sign-in data for people receiving ballots on election day and absentee ballot data. His analysis was limited to 22 contested elections involving an Hispanic candidate. He did not consider elections involving only Anglo candidates or uncontested elections.

In 7 of the 22 elections, he found that Hispanic support for the Hispanic candidate exceeded 80 percent. He labeled 19 of the 22 elections as "decisive," finding that in 16 elections, the Hispanic candidate was defeated. As a result, he opined that 21 out of 22 elections were ethnically polarized.^[19] Dr. Engstrom also opined that there has been a decrease in voting cohesion and ethnic polarization since 1982.

2. Dr. Weber's Opinions^[20]

Like Dr. Engstrom, Dr. Weber relied upon homogeneous precinct analysis, bivariate ecological regression analysis, and the King methodology, in his review of 140 elections. He, too, relied primarily on the first two methods. Dr. Weber initially considered 1990 and 2000 census data, but he later relied upon Hispanic surnames as reflected in registration data, election returns, and voter sign-in data. He posited a threshold for determining voter cohesiveness — cohesiveness existed within a group when a candidate received 60% of the group's vote. In elections involving more than two candidates, cohesiveness was demonstrated by a percentage less than 60% based upon the potential splitting of votes by additional candidates.

Dr. Weber examined all of the contested county commissioner primary and general elections in Alamosa County since 1982. He also examined 120 exogenous elections for countywide, statewide, national, and other offices. Dr. Weber opined that the 22 elections examined by Dr. Engstrom were an unrepresentative sample. Based on his analysis of 140 elections, Dr. Weber opined that Hispanic voters voted cohesively in both endogenous and exogenous races and that in the exogenous elections, Hispanic voter cohesion was "quite high."

Dr. Weber also analyzed several elections using a multivariate (multiple variables) analysis. This analysis was offered to assist in determining "why" voters cast their votes in a particular manner by considering a multitude of factors such as incumbency and partisanship. In 18 out of 25 comparisons, Dr. Weber concluded that partisanship and incumbency bore a higher correlation to the outcome of county commissioner general elections than did ethnicity.

E. Findings Combining Statistical Evidence, the Parties' Stipulations and Anecdotal Evidence

Cohesiveness. The parties stipulated that Hispanic voters tend to vote cohesively in both primary and general elections for the Board. The Court defers to this stipulation but, as discussed infra, gives the statistical evidence and conclusions drawn therefrom little weight with regard to other issues. The anecdotal evidence of cohesive voting was mixed and limited to subjective commentary about particular elections. It suggests that although some Hispanic voters preferred Hispanic candidates, others voted based upon party affiliation, particular issues, familiarity with the candidate or other factors.^[21]

Success of Hispanic Candidates. Since 1980, Hispanic candidates have been overwhelmingly successful in primary elections and successful 60% of the time in county commissioner general elections. Hispanic candidates were always successful in exogenous Democratic primary elections between 1982 and 2002, but they rarely prevailed in general elections.

Voter Registration and Participation. The voter registration process in Alamosa County is consistent with statewide procedure and can be accomplished at a variety of locations.^[22] Accommodations are made for Spanish speakers both during the registration process^[23] and at the polls.^[24] Periodically, there have been voter registration drives that target Hispanic residents.

The Hispanic voter registration rate is lower than that of Anglo residents; in 2000, the differential was 30%. This may be explained by a combination of factors. First, the figures are subject to the surname bias. In addition, the Hispanic population is disproportionately younger than the Anglo population and many are not of voting age. By comparison, older Anglo residents are more likely to register.

Even though lower registration rates may suggest that Hispanic residents do not participate in the electoral process, other statistics with regard to political participation tell a different story. Hispanic residents in Alamosa County register to vote at a higher rate than Hispanic residents throughout Colorado and the nation — 62.8% of voting-age Alamosa County Hispanic residents are registered to vote compared with 41.5% of voting-age Hispanic persons in Colorado and 34.9% of voting-age Hispanic persons throughout the United States. Participation of Hispanic voters has increased from 25% of voters in 1982 to 29% of voters in 2002. Although Hispanic voter turnout has varied from election to election,^[25] the turnout rates for Hispanic and Anglo voters in a given election are similar. Indeed, Hispanic voter turnout rates have exceeded those of Anglo voters in contested primary and general county commissioner elections.

There are few, if any, issues facing the Board which can be identified solely as "Hispanic issues." Hispanic and Anglo residents share common concerns^[26] which have varied from time to time^[27] — jobs, health and human services, housing, taxes, land use, and seniors. In addition, rural residents often focus upon subdivision requirements, zoning, maintaining a good relationship with the city and roads. Issues such as housing and social services are not tied to ethnicity but instead reflect the socioeconomic needs of impoverished Alamosa County residents. There is no evidence that the Board has failed to address residents' needs based upon ethnicity.

Ethnic Appeals in Elections. The results of many endogenous and exogenous elections reflect a fundamental electoral truth — that to be elected in Alamosa County, a candidate must appeal to both Anglo and Hispanic voters. Anglo Al Kelly was asked to serve as the campaign manager for Hispanic Patricio Gonzales, a state representative candidate, in order to get the "Anglo vote." To expand electoral support, Hispanic candidate Anna Mae Real-Lindsay chose to emphasize her credentials rather than her ethnicity. Successful Hispanic candidates Jorge Amaya, Lee Patrick Herrera, and Frank Mestas all attribute their success to the fact that they focused upon political issues and qualifications rather than ethnicity.^[28]

Subtle or overt ethnic appeals have been made in some elections. However in each instance, the appeal was made by the Hispanic rather than the Anglo candidate and resulted in failure at the polls. A subtle ethnic appeal was made in Marguerite Salazar's campaign for county commissioner in 1992. Although she did not intend to wave an "ethnic flag," she ran as a designated Hispanic role model immediately after joining the Hispanic Leadership Institute and shortly after her husband, Arnold Salazar, had advocated the adoption of a resolution against Anglo bloc voting. Consistent with her view that Hispanic voters can only be represented by Hispanic candidates,^[29] she campaigned primarily in Hispanic neighborhoods. Overt ethnic appeals were made by Walt Gomez in his unsuccessful campaigns in 1978 and 1988, and his focus upon ethnic identification was divisive among members of the Democratic party.^[30]

IV. CONCLUSIONS OF LAW

A. Is Section 2 of the Voting Rights Act constitutional?

The Defendants argue that Section 2 of the Voting Rights Act is unconstitutional. They contend it does not comply with the Fourteenth and Fifteenth Amendments to the United States Constitution because: (1) only purposeful discrimination may be remedied under these amendments; (2) there is no evidence in the Congressional Record of widespread discrimination through the use of at-large elections; (3) its remedy is overbroad; and (4) it targets private discrimination.

Both the Supreme Court and the Tenth Circuit Court of Appeals have affirmed the constitutionality of Section 2. See South Carolina v. Katzenbach, 383 U.S. 301, 308, 86 S.Ct. 803, 15 L.Ed.2d 769 (1966); Mississippi Republican Exec. Comm. v. Brooks, 469 U.S. 1002, 1003, 105 S.Ct. 416, 83 L.Ed.2d 343 (1984) (summarily affirming circuit court opinion despite challenge to the constitutionality of Section 2) (Stevens, J., concurring); Sanchez v. Colorado, 97 F.3d 1303, 1314 (10th Cir.1996). In light of binding precedent, this Court declines to reexplore the issue.

B. Does the at-large election of county commissioners violate Section 2 of the Voting Rights Act?

The Government contends that the at-large election of county commissioners in Alamosa County violates Section 2 of the Voting Rights Act because it allows Anglo residents to vote as a bloc to defeat the candidates preferred by Hispanic voters. The Defendants counter that the method of electing county commissioners has not prevented Hispanic voters in Alamosa County from participating in the electoral process nor from electing the candidates they prefer.

Section 2 of the Voting Rights Act^[31] prohibits any election practice "which operate[s], designedly or otherwise" to deny members of ethnic or racial minority groups the same opportunity to participate in the political process that other citizens enjoy. Section 2 focuses upon a practice's effect rather than the intent behind the practice.

The Supreme Court prescribed a two-step framework for analyzing Section 2 challenges in Thornburg v. Gingles, 478 U.S. 30, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986). First, a court must determine whether there is sufficient evidence to establish three preconditions: (1) that the minority group is sufficiently large and geographically compact to constitute a majority in a hypothetical single-member district; (2) that the minority group is politically cohesive, i.e., members of the group vote the same way; and (3) that the members of the majority have voted as a bloc to defeat the preferred candidates of the minority group.^[32]Gingles, 478 U.S. at 50-51, 106 S.Ct. 2752. Satisfaction of these three preconditions creates an inference that the challenged practice is discriminatory, but the inference, alone, is insufficient to invalidate an election practice. Sanchez v. Colorado, 97 F.3d at 1310.

If the Gingles preconditions are satisfied, a court then must exercise common sense and thoroughly canvass all material facts to determine whether the practice is discriminatory in light of the "totality of the circumstances." Sanchez v. Bond, 875 F.2d at 1492. A court may consider the nonexclusive list of factors known as the "senate factors,"^[33] as well as any other material facts. The senate factors include: whether there has been official discrimination against a minority group which has affected the ability of its members to vote, register to vote, or participate in the political process; whether elections are ethnically/racially polarized; whether there are voting procedures which enhance the opportunity for discrimination; whether members of the minority group have been denied access to the candidate slating process; whether historical discrimination against the minority group has hindered its ability to participate in the political process; whether overt or subtle ethnic appeals were made by the candidates in their campaigns; whether members of the minority group have been elected to public office; whether elected officials are responsive to the identified needs of the minority group; and whether the justification for the challenged electoral practice is tenuous.

1. First Gingles Precondition: Are Hispanic residents in Alamosa County sufficiently numerous and geographically compact to constitute a majority in a single-member district?

The first Gingles precondition is a practical requirement. The Government must demonstrate that if a Section 2 violation is found, a remedy can be fashioned. Unless the minority group is large and compact enough for formation of a single, Hispanic-majority district there is no feasible remedy. Sanchez v. Colorado, 97 F.3d at 1311.

The Court concludes that the Hispanic population in Alamosa County is sufficiently large and geographically compact enough to form a district with an Hispanic majority. The 2000 Census data and trial testimony show that Hispanic residents constitute 41% of the county population and 38% of the voting age population. The Hispanic population is geographically concentrated in the City of Alamosa, somewhat more densely south of the railroad tracks that bisect the city from west to east. Although the current precinct lines cannot be used to create a single-member district with a majority of Hispanic registered voters, the Government has proposed three hypothetical districts in which Hispanic residents would comprise at least 60% of the voting age population.^[34] In each proposed district, there would be an Hispanic majority of registered voters. None of the districts would be unusual or irregular in shape or configuration.

The Defendants concede that illustrative districts could be created, but argue that the hypothetical districts unnecessarily split precincts and therefore improperly subordinate traditional districting principles to ethnic considerations. As a result, they contend that any district with such a high Hispanic percentage raises constitutional concerns.^[35]

The criticisms by the Defendants are misplaced. No evidence was presented as to the "traditional districting principles" of the State of Colorado or how the formation of the hypothetical districts would be in conflict. More importantly, the current precincts in Alamosa County are very large and have been in place for only ten years. In 1993, Alamosa County redrew its precincts, reducing the number from 16 to 8. In the redrawing of precinct lines, almost every pre-existing precinct was split. The illustrative districts borrow, in part, upon current precinct lines but also incorporate traditional county demarcations such as county roads, state highways and municipal boundaries. Absent apparent conflict with Colorado's traditional districting principles, any one of the proposed districts satisfies the first Gingles precondition.

2. Second Gingles Precondition: Are Hispanic residents in Alamosa County politically cohesive?

A minority group is considered politically cohesive if a significant number of the group's members prefer the same candidate. See Sanchez v. Bond, 875 F.2d at 1493. The customary and primary source for such evidence is expert testimony based upon statistical analyses of actual elections. Id.; Sanchez v. Colorado, 97 F.3d at 1312.

Both Dr. Engstrom and Dr. Weber estimated group voting patterns using bivariate ecological regression analysis, extreme case analysis and the ecological inference technique developed by Gary King. Both relied upon Spanish surnames in voter registration and turnout data to measure the number of Hispanic voters in each of Alamosa County's precincts. They then compared the number of Hispanic voters to the votes cast for each candidate in particular elections. Despite slight differences in methodology, decision rules and the elections studied, they agreed that between 1982 and 2002 Hispanic voters in Alamosa County voted in a politically cohesive manner in both endogenous and exogenous elections. As a result, the parties stipulated that Hispanic voters vote cohesively.

Based solely on this stipulation, without weighing the statistical evidence or considering the anecdotal evidence to the contrary, the Court concludes that the second Gingles precondition is satisfied.

3. Third Gingles Precondition: Does the Anglo majority usually vote as a bloc to defeat candidates preferred by Hispanic voters?

The third Gingles precondition requires that a court determine 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." Sanchez v. Colorado, 97 F.3d at 1312 (quoting Gingles, 478 U.S. at 51, 106 S.Ct. 2752) (emphasis added). Legally significant "white bloc voting" occurs when the Anglo vote normally defeats the combined strength of cohesive minority support plus white crossover votes. Gingles, 478 U.S. at 56, 106 S.Ct. 2752. The focus is upon past election results — how voters voted, rather than why they voted the way they did. A court must determine what usually occurs in a jurisdiction, as distinguished from aberrational election results produced by special circumstances. See id. at 57 n. 26, 106 S.Ct. 2752; Sanchez v. Colorado, 97 F.3d at 1310.

In Sanchez v. Colorado, the Tenth Circuit endorsed the analytical approach taken by the Fourth Circuit in Collins v. City of Norfolk, Va., 883 F.2d 1232, 1237 (4th Cir.1989). This inquiry begins by identifying the "preferred candidate" of the minority group. After identifying the candidate preferred by the minority group, a court examines whether Anglo voters have voted as a bloc (cohesively) against such candidate. If so, the remaining issue is whether such bloc voting resulted in the defeat of the minority group's preferred candidate.^[36] In applying the Collins test, the Tenth Circuit cautions that a court should not assume that a candidate was preferred by a minority group solely by virtue of the candidate's party or racial/ethnic status. It is the burden of the party seeking to establish a Section 2 violation to produce sufficient anecdotal or statistical evidence to establish the minority's preference in each election. Sanchez v. Colorado, 97 F.3d at 1319.

a. What candidates were preferred by Hispanic voters in Alamosa County elections?

Both experts relied upon the statistical information previously identified in conjunction with the second Gingles precondition. Each expert compared the number of votes that a particular candidate received in each precinct with the percentage of Hispanic voters in that precinct. The results were graphed on a scatter plot. The horizontal axis on the scatter plot measured the percentage of Hispanic voters in a precinct, from 0 to 100 percent — the independent variable. The vertical axis measured the votes a candidate received from 0 to 100 percent — the dependent variable. A regression line was drawn through the graphed points from 0 to 100 percent along the horizontal axis. The estimated percentage of Hispanic votes for a given candidate was the point at which the regression line coincided with a precinct of 100 percent Hispanic voters.^[37]

The sample of elections studied by Dr. Engstrom was smaller (8 endogenous and 14 exogenous elections occurring between 1980 and 2002) than the sample considered by Dr. Weber (20 endogenous and 120 exogenous elections during that same time period). In addition, Dr. Engstrom did not analyze elections involving only Anglo candidates or unopposed elections, whereas Dr. Weber included Hispanic/Anglo, Hispanic/Hispanic, Anglo/Anglo and unopposed races in his study.

Unfortunately, neither expert explained how he identified the candidate that was preferred by Hispanic voters. In Hispanic/Anglo elections, both experts assumed and concluded that Hispanic voters preferred the Hispanic candidate.^[38] Because he did not address Anglo/Anglo races, Dr. Engstrom offered no opinion as to the candidates preferred by Hispanic voters in that context,^[39] and no evidence was presented to establish the Hispanic-preferred candidate in Hispanic/Hispanic races. Thus, the Court is left with the parties' stipulated assumption and conclusion that Hispanic voters prefer Hispanic candidates, but only in Hispanic/Anglo races

b. Have Anglo voters voted as a bloc against candidates preferred by Hispanic voters? If so, has this resulted in the defeat of the candidates preferred by Hispanic voters?

Unlike Sanchez v. Colorado, in which the Tenth Circuit specifically noted that no Hispanic candidate had ever been elected from House District 60, and unlike other county-wide elections in which no Hispanic candidate has ever been elected, Hispanic candidates have been elected as county commissioners not once, but three times in the last twenty years. Hispanic candidates have been elected in three out of the five races in which they participated (Mr. Amaya and Mr. Herrera in 1984, and Mr. Mestas in 2002). Interestingly, in the 2002 election, two Hispanic candidates ran in opposition. As a result, Hispanic candidates have prevailed in 60% of the elections in which they competed.

The Government argues that each election of an Hispanic candidate to the Board is anomalous and should either be disregarded or discounted. The Government relies upon language in Gingles which recognizes that in special circumstances — i.e., unopposed elections, elections involving an incumbent, or elections where bullet voting was used — a minority candidate's electoral success may not be dispositive.^[40] The Government argues that the election of Mr. Amaya was unusual because it occurred in a three-way race won by plurality vote, and it implies that a two way race would have had a different outcome. The Government further contends that Mr. Amaya and Mr. Herrera were successful in 1984 only because they did not emphasize their ethnicity. With regard to the election of Mr. Mestas in 2002, the Government argues that the outcome was influenced by the filing of this action early that year and that Mr. Mestas is not truly representative of Hispanic voters because he is a Republican.

The Government's arguments are not persuasive. In Mr. Amaya's three-way 1984 race, Mr. Carpenter drew votes which otherwise might have been cast for either Mr. Relyea or Mr. Amaya, but there is no evidence (only Dr. Engstrom's unsupported speculation) that had Mr. Carpenter not run, votes cast for him would have benefitted Mr. Relyea as compared to Mr. Amaya. The election strategies used by Mr. Amaya and Mr. Herrera in 1984 do not justify discounting their success. To the contrary, their success demonstrates their political savvy. Candidates with an attribute which makes them appear "different" from the majority, be it sex, ethnic background, religion, education, social privilege or the like, often de-emphasize the attribute in order to appeal to the broadest swath of the electorate.^[41] As for the 2002 election, there is no evidence that the pendency of this lawsuit was known to the voters or affected their choice. Indeed, the unrebutted testimony suggests that this lawsuit was unimportant in the election outcome. Mr. Mestas decided to run for the Board and began his campaign before the lawsuit was filed. Had he lost, his Hispanic opponent, Mr. Roybal, would have won.^[42] Thus, the Court concludes that the 1984 and 2002 elections of Hispanic candidates to the Board are not anomalous.

Turning to the statistical evidence, one can only determine whether there has been successful Anglo bloc voting against candidates preferred by Hispanic voters if one can identify whom Hispanic voters prefer. As noted earlier, the evidence does not establish which candidates were preferred by Hispanic voters in Anglo/Anglo or Hispanic/Hispanic races, thus the Court is limited to consideration of Hispanic/Anglo elections and the parties' assumption that in such races Hispanic voters preferred an Hispanic candidate.

Because the electoral practice challenged in this action is the at-large election of county commissioners, those elections are more probative than any others. Exogenous race results might be probative for purposes of comparison if Hispanic candidates were routinely successful in exogenous races but never in county commissioner races, or vice versa, but neither situation is evident here. As to endogenous elections, the Court finds general elections to be more probative because they involve all voters in the county.^[43]

Thus, for the purposes of this precondition, the Court is left with general election races between Hispanic and Anglo candidates for county commissioner seats. Since 1980, there have been four such elections: 1980 (Sanchez(H)/Gallagher); 1984 (Herrera(H)/Miles); 1984 (Amaya(H)/Relyea/Carpenter); and 1992 (Salazar(H)/Gallagher). The experts agree that the 1984 elections do not show evidence of Anglo bloc voting resulting in the defeat of the Hispanic-preferred candidate because both Mr. Amaya and Mr. Herrera won With regard to the 1980 election, Dr. Engstrom concluded that Anglo bloc voting successfully defeated Mr. Sanchez; Dr. Weber did not analyze this election. Both experts agreed that in the 1992 election, Anglo bloc voting defeated Ms. Salazar, the Hispanic-preferred candidate. Construing the evidence most favorably to the Government, this tally shows that, at most, ethnic bloc voting defeated the Hispanic candidate in 50% of the elections. From this, the Court cannot conclude that ethnic bloc voting "usually" results in the defeat of the minority-preferred candidate, but the evidence is sufficient to warrant exploration of the totality of the circumstances.

4. The totality of the circumstances

As noted earlier, a finding that the three Gingles preconditions are satisfied creates only an inference but does not establish a violation of Section 2. That is especially true in this case, where the stipulations of the parties compel the conclusion that preconditions 1 and 2 are satisfied and the evidence is only close as to precondition 3. Here, it is particularly important to canvass all relevant facts, with an eye to past and present reality. See DeGrandy, 512 U.S. at 1011, 114 S.Ct. 2647. The Court considers the nonexclusive list of senate factors, as well as several others.

a. Official electoral practices in Alamosa County

A number of the senate factors focus upon whether official electoral practices discourage or impede electoral participation by members of the minority group.^[44] The practices with which the Senate was concerned are those that might exacerbate the discriminatory effect of the challenged procedure. Examples include the stuffing of ballot boxes, violence against minority voters, property ownership requirements, poll taxes, and all-white primaries. See DeGrandy, 512 U.S. at 1018, 114 S.Ct. 2647. Innocuous election practices (such as anti-single shot provisions) imbued with discriminatory goals can also be relevant, as are practices that exclude minority voters from participation in the selection of candidates such as the denial of access to a candidate slating process. Id.

The Government presented no evidence of an official discriminatory practice in Alamosa County which in combination with the at-large election of county commissioners has adversely affected the ability of Hispanic voters to vote, register to vote, select candidates or participate in the political process. The Government's only argument in this regard is that Hispanic individuals have not had "equal access to or support from" the county Democratic Central Committee. There is no official role for the Democratic Central Committee in the candidate slating process, but the Government contends that the Democratic Central Committee plays a functional role in selecting candidates because it provides leadership in fund-raising, voter registration, delegate selection and campaign support for local Democratic candidates. It argues that the Democratic Central Committee has been united in support of Anglo candidates, but has been divided along ethnic lines when the candidate was Hispanic.

Assuming arguendo that the Democratic Central Committee plays a functional role in the selection of county commissioner candidates, the evidence offered on this point consisted entirely of anecdotal testimony regarding ethnically biased comments and boorish behavior by members of the Democratic party,^[45] some of whom may have been serving on the Democratic Central Committee at the operative time. The comments and behaviors are undoubtedly offensive, but they are comments and actions of individuals and there is no evidence linking such comments or behavior to a policy or practice of the Democratic Central Committee. It is therefore impossible to conclude that the Democratic Central Committee acted as an entity in espousing ethnic bias.

There is also insufficient evidence to conclude that disagreements within the Democratic party or the objectionable statements of party members have adversely impacted the ability of Hispanic individuals to participate in the electoral process. The evidence suggests just the contrary — that Hispanic individuals have been actively involved in Democratic party politics and have successfully nominated many Hispanic candidates. Hispanic voters have served on the Democratic Central Committee for an extended period; Pedro Gomez served as chairperson from 1985 through 1990, was succeeded by Carlos Lucero, who in turn was succeeded by Art Ortiz. Mr. Gomez is currently the chairperson of the committee. In addition, Dr. Weber opined, without contradiction, that the majority of voters in over half of the contested Democratic primaries between 1982 and 2002 were Hispanic. Since 1980 five Hispanic surnamed candidates have been nominated to run in general elections — four by the Democratic party. Of these four, two were elected. Other than individuals who changed party affiliation, there is no evidence that objectionable comments or behavior within the Democratic party affected the participation of Hispanic voters either in primary or general elections.

The Court therefore concludes that there is no official practice which in concert with the at-large election of county commissioners impairs, impedes or discourages participation by Hispanic individuals in the electoral process.

b. The impact of historical experience upon Hispanic voter participation

In many Voting Rights Act cases, historical discrimination is a critical component in evaluating the challenged voting practice because a lingering effect of such discrimination is depression of the educational and socioeconomic circumstances of minority group members, which can contribute to group insularity and can discourage voter participation.^[46] In this case, the Government argues that historical ethnic discrimination in Alamosa County has perpetuated socioeconomic and educational inequities between Anglo and Hispanic residents which, in turn, have resulted in lower voter participation by Hispanic residents. The Defendants acknowledge past discrimination and current socioeconomic and educational disparities between the Anglo and Hispanic populations in Alamosa County, but contend that such differences have not adversely affected voter participation in recent years. They argue that the voter participation of both groups is similar, that the lower rate of voter registration among Hispanic individuals is attributable to the age of the population, and that the Hispanic population has grown increasingly diverse, now reflecting a variety of political and social views. Resolution of this conflict requires an appreciation of the role and history of Hispanic residents in Alamosa County.

Alamosa County is unique. Situated in the San Luis Valley of south central Colorado, it consists of 723 square miles, bordered by Saguache, Huerfano, Costilla, Conejos, and Rio Grande counties. The county has had both Hispanic and Anglo residents for more than a century. Many Hispanic and Anglo families have lived in the county or in the valley for numerous generations, and have intermarried at a rate higher than the national average. Although Hispanic residents are more concentrated in the city than in rural areas, both ethnic groups live in all areas of the county. Thus, Hispanic residents of Alamosa County are not as geographically and socially isolated as many minority groups.

Although discriminatory practices were common during the first half of the 20th century, in response to social change and legal mandate, both official and social practices have improved. Formal school segregation of Hispanic students ended almost ninety years ago; lingering informal de facto desegregation in the schools was addressed with the adoption of the "Alamosa Plan" in 1980, twenty-three years ago. Now, the school superintendent is Hispanic,^[47] as are one-third of all teachers, principals and administrative staff in the school system.

The Government contends that discrimination against Hispanic students in the school system has not been fully addressed or at least continued well into the 1990's, pointing to the 1993 investigation by the Office of Civil Rights into educational opportunities for students who did not speak English. Such evidence lacks persuasive impact. Although the investigation concerned programs for students who had limited English proficiency, these students were not Spanish-speaking Hispanic residents, but rather recent immigrants from central American countries, who speak native languages such as Kanjobal.^[48]

Interestingly, notwithstanding claims of historical discrimination, almost all of the Hispanic witnesses who were educated in Alamosa County and graduated from high school during or prior to the 1970's hold college as well as post-graduate degrees.^[49] Their academic accomplishment may be due, in part, to the proximity of Adams State College, but their educational, economic and professional successes belie complaints of lingering disadvantage. All are or have been politically active and they can take credit for change in the county's official practices and the socioeconomic conditions during the past twenty years.

Census studies point to improving educational accomplishments by current Hispanic students. In 1990, 61.3% of self-identified Hispanic individuals graduated from high school, 34.5% attended college and 5.1% obtained a graduate school or professional degree. By 2000, the numbers increased in all categories to 68.9%, 39.4% and 5.9%, respectively. According to Superintendent Herrera, on third grade CSAP tests, there has been a 35% improvement among Hispanic students since 2000, bringing the performance of Hispanic students within a percentage point or two of non-Hispanic students.

Similarly, the railroad tracks no longer create a rigid social and economic line of demarcation in the City of Alamosa. Hispanic residents such as Marguerite and Arnold Salazar live in the northern part of the city. Hispanic individuals own and operate businesses throughout the city and there is little evidence of de facto social segregation in businesses, restaurants, banks, clubs, or churches.

The 2000 Census figures also reveal socioeconomic progress. The Government argued that countywide averages show that Hispanic residents have been less successful than Anglo residents. The Defendants argue that the same data show that socioeconomic conditions of Hispanic residents have improved, resulting in increasing diversity among the Hispanic population. Both observations have merit.

Hispanic residents have a slightly higher rate of unemployment^[50] and fewer have annual incomes above $50,000.^[51] However, Hispanic residents in Alamosa County are substantially better off economically than Hispanic individuals nationally and census statistics show improvement in all areas between 1990 and 2000.^[52] In 2000, 28.5% of Hispanic households had income levels at or below the poverty level, 32.1% had incomes above $35,000, and 17.9% had incomes above $50,000. This represents an improvement from 1990 when 31.8% of Hispanic households had incomes at or below the poverty level and only 19.8% had incomes above $35,000. As for home ownership, in 1990, 58.0% of Hispanic households owned homes (43% of homeowners lived in high income areas and 39% lived in low income areas). By 2000, the number of Hispanic homeowners increased to 60.9% (30% living in high income areas and 23% living in low income areas). These statistics suggest overall economic progress and the broadening of an Hispanic middle class.

Thus, in both the educational and economic arenas, the Court concludes that: (1) although Anglo residents as a group surpass Hispanic residents as a group, the gap is closing; and (2) the Hispanic population is both educationally and socioeconomically diverse. Thus, Hispanic residents of Alamosa County are unlike insular minority groups in other circumstances who share uniformly depressed socioeconomic conditions and educational experiences.

The educational, socioeconomic and geographic diversity within the Hispanic population is also reflected in social and political views. There is an apparent rift within the Hispanic community, particularly among the well-educated societal leaders who testified in this case. Those who testified on behalf of the Government are, to a large degree, related by blood or marriage,^[53] and have been actively involved in political and social efforts to improve conditions for Hispanic residents since the 1970's.^[54] By virtue of their efforts much social improvement has occurred, but they uniformly see the cup as half-empty. This group complains of continuing discrimination and social polarization demonstrated by occasional ethnic slurs and the failure of institutions to implement their political agenda. They are frustrated and feel that due to their ethnic background they and others have not been treated with appropriate respect. As noted earlier, some of these witnesses regard Hispanic neighbors who do not share their views, particularly Republicans, to be unrepresentative of the Hispanic community.

On the other hand, the Hispanic witnesses who testified for the Defendants see the cup as half-full. They speak to social, educational and economic progress that has occurred during the past twenty years or more and to increasingly greater opportunities. These witnesses point to numerous Hispanic role-models in positions of authority in government, business, social support agencies, the school system and the court system. Some are Republicans; some are Democrats, but these witnesses are satisfied that the current system for electing county commissioners gives Hispanic residents equal access to the electoral process.

The Government also contends that Hispanic voter registration has been depressed due to past discrimination. It notes that only 62.8% of Hispanic residents registered to vote in the 2002 general election as compared to 92.8% of Anglo residents. Assuming the statistics to be accurate,^[55] the Court finds the argument unpersuasive because there are ethnic-neutral explanations for lower voter registration rates among Hispanic residents and because the evidence of Hispanic voter participation is high.

These statistics are subject to surname bias. In addition, the experts agree that voter registration rates are lower when the population is younger. In Alamosa County, the Hispanic population has a disproportionate share of the youngest age groups. Voter registration has fluctuated by election and in accordance with voter registration drives, but Alamosa County's Hispanic voter registration rate of 62.8% far exceeds Hispanic registration in Colorado (41.5%) and nationally (34.9%), and compares favorably with statewide registration for all voters (64.1%). In addition, no expert found any appreciable difference in rates of participation in general elections as between registered Hispanic and Anglo voters.^[56]

Thus, the Court concludes that notwithstanding historical ethnicity-based discrimination, there is insufficient evidence to conclude that socioeconomic or educational conditions currently hinder Hispanic residents in Alamosa County from participating in the electoral process. To the contrary, substantial efforts to address discrimination during the last 50 years have resulted in improving socioeconomic and educational conditions leading to an increasingly diverse and politically involved Hispanic population.

c. Ethnic polarization in the election process

The ultimate issue to be determined in this case is whether ethnic polarization has usually resulted in the defeat of candidates preferred by Hispanic voters in county commissioner elections. The Gingles preconditions require a minimal showing of ethnic polarization based on past election results. The focus of the inquiry at that juncture was limited to "how" voters voted. Now the Court considers "why" voters voted as they did.

There are many factors to consider. In addition to those discussed previously, a court may consider the justification for the challenged electoral practice, whether the Hispanic population in Alamosa County has unique needs, and if so, to what degree elected officials have responded to those needs.^[57] A court may also consider statistical evidence to determine the degree of polarization in pertinent elections.^[58] The weight to be given to this evidence depends upon its statistical validity, its correlation with the anecdotal evidence, whether minority members have been elected to public office,^[59] whether ethnic appeals have been made by candidates,^[60] and whether some factor, such as incumbency or political party, provides an explanation for the election result. Ultimately, however, a court must return to the purpose of Section 2 of the Voting Rights Act — to ensure that "the political processes leading to nomination or election in the State or political subdivision are equally open to participation by members of" a minority group. 42 U.S.C. § 1973(b).

It is helpful at this juncture to remember the origin and purpose of the election method at issue. The at-large election of county commissioners is prescribed by Colorado statute, § 30-10-306, C.R.S., and is applicable to all Colorado counties. The at-large election process makes each elected commissioner, regardless of his or her residency district, individually accountable to every voter in the county.^[61]

The evidence does not suggest that the Board has acted with ethnic bias. The current needs of Hispanic residents appear consonant with those of the population at large. The most notable divergence in interest is tied to socioeconomic need, with the interests of impoverished Hispanic residents being similar to those of impoverished Anglo residents.

The question then becomes whether the at-large election of Board members usually is so polarized as to defeat Hispanic-preferred candidates. If so, this can result in dilution of the votes of members of a minority group if members of the majority vote as a bloc to defeat the minority's preferred candidate.

The Government offered two types of evidence, statistical and anecdotal, to prove that county commissioner races are polarized and the votes of Hispanic voters have been diluted. In the Gingles analysis, the Court found that the statistical evidence barely passed muster to establish the three preconditions. In this analytical phase it is even less persuasive. The statistical methodologies employed by the experts are best suited to a geographically insular minority group in which members share similar incomes, employment status, educational attainment, type and location of housing and living conditions, as well as language and ethnic identity. In such situations, ethnic identity becomes the talisman of similar values and political interests.

The Alamosa Hispanic population, however, is not an insular, monolithic, minority group. It is, instead, diverse according to all measures — socioeconomic circumstances, educational achievement, geographic location and political views. Put simply, Hispanic residents in Alamosa County create a rich tapestry of experience and views not accounted for in the statistical analysis. The statistical analysis is flawed in other respects, as well. It is limited by a small survey size (16 or 8 precincts), the unavailability of absentee data and the effect of surname bias. Even without discount for statistical inaccuracy, the evidence of ethnic polarity in county commissioner elections is limited to 50% of the Hispanic/Anglo races. This hardly makes polarized elections usual, and if one accepts the experts' opinions as to polarity, one must also accept their conclusion that the frequency and degree of polarization is steadily decreasing.

The anecdotal evidence does nothing to buttress the statistical conclusions as to polarization. No witness testified as to how he or she voted or why; not a single witness testified that he or she did not participate in the electoral process due to a perception of futility based upon ethnic discrimination. To the contrary and without exception, Hispanic witnesses demonstrated their extensive knowledge about and active participation in the political process in Alamosa County. In discussing election outcomes, they assessed the effect of incumbency, candidate qualifications and election strategy.

Despite historical discrimination and frustration in implementing particular political or social agendas, Hispanic residents in Alamosa County freely and fully exercise their right of suffrage in the county commissioner elections. The proof of their participation is in the pudding — the successful nomination of Hispanic candidates through both the Democratic and Republican parties and the election of Hispanic candidates to the Board in three out of the five races in which Hispanic candidates competed.

V. CONCLUSION

The Court declines to reexplore whether Section 2 is constitutional. Although the evidence presented at trial is arguably facially sufficient to satisfy the three Gingles preconditions, upon consideration of the totality of the circumstances, it does not prove that the at-large method of electing county commissioners in Alamosa County dilutes the vote of Hispanic residents. The ability of Hispanic voters to participate or elect candidates of their choice in county commissioner elections has neither been abridged nor curtailed in violation of Section 2 of the Voting Rights Act. As a consequence, the current at-large system of electing county commissioners in Alamosa County should not be disturbed.

IT IS THEREFORE ORDERED that the Clerk of Court enter judgment in favor of Defendants Alamosa County, Alamosa County Board of Commissioners, Robert Zimmerman, Darius Allen, Charlotte Bobicki and Holly Z. Lowder, and against the United States of America declaring that the at-large system of electing county commissioners in Alamosa County does not violate Section 2 of the Voting Rights Act.

NOTES

[1] Robert Zimmerman, Darius Allen, and Charlotte Bobicki.

[2] Holly Z. Lowder.

[3] Throughout the trial the parties distinguish between "Anglo" and "Hispanic" persons. Although admittedly imprecise, this opinion uses the term "Anglo" interchangeably with "non-Hispanic." "Hispanic" is used to identify persons of Spanish descent, or in some cases persons with Spanish surnames.

[4] The findings of fact are based upon the testimony of witnesses, both in person and by deposition, and admitted exhibits. The following exhibits were admitted: Government's Exhibits 1, 13, 26, 31, 34, 37, 50, 53, 62, 63A, 65, 95, 96, 98, 99, 101, 105, 106, 108, 120, 121, 123, and 137 through 141; and Defendants' Exhibits A1, A2 attachment B, A24, and A26. The following exhibits were received solely for demonstrative purposes: Government's Exhibits 8, 10, 11, 12 p. 7, 14 through 24, 27 through 27u, 32 through 32w, 38-40, and 143; and Defendants' Exhibits A2-D, A2-F, A2-H, A2-I, A5 through A23, A25 pp. 11-15, A30, A34 through A36, A39, A41, A43, and A51.

[5] § 1-4-801, C.R.S.

[6] Alamosa County is also comprised of eight precincts. Prior to May 26, 1993, it had sixteen precincts.

[7] According to Quentin Garcia, students were punished for speaking Spanish and ethnic first names were anglicized.

[8] Anecdotally, the Court notes that almost all Hispanic witnesses in this case (most over the age of 40) hold a college degree, and many have graduate degrees, for example: Arnold Salazar (undergraduate degree and a graduate degree in social work administration); Lee Patrick Herrera (bachelor's and master's degrees, and an ordained minister); Pedro Gomez (undergraduate degrees in business and economics, and a Master of Education); Henry Herrera (undergraduate degree and postgraduate course work in school administration); Lillian Gomez (bachelor's and master's degrees); Walt Gomez (undergraduate degrees in music education and athletics, and a master's degree); Frank Mestas (bachelor's degrees in education, industrial arts, and math); Daisy Ortega (bachelor's degree); Anna Mae Real-Lindsay (bachelor's degree in education); Toney Cantu (undergraduate degree); Benjamin Duarte (undergraduate degrees in sociology and business); Jorge Amaya (bachelor's degree); Marguerite Salazar (undergraduate degrees in psychology and sociology, and a master's degree); and Charles Griego (bachelor's degree in business).

[9] Several Hispanic witnesses testified that neither they nor other Hispanic residents suffer discrimination based on ethnicity. Bank vice president Toney Cantu testified that Hispanic residents do not suffer discrimination in the financial services industry. He also testified that he has never been denied access to, nor been discouraged from participating in, the political process because he is Hispanic. Likewise, Anna Mae Real-Lindsay, the director of the adult education program and learning center at the Trinidad State Junior College, testified that she has never been denied access to or participation in social, business or political activities because she is Hispanic. Daisy Ortega, an independent sales representative for Avon, has not been denied access to the political process because she is Hispanic, nor has she experienced any impediment to voting or registering to vote. She estimates that less than 10% of citizens in Alamosa County hold ethnically bigoted views. Benjamin Duarte, the court administrator for the Twelfth Judicial District, testified that Hispanic residents do not suffer ethnic discrimination in the state justice system although they are over-represented as defendants in criminal cases.

[10] In 1988, discriminatory comments against Hispanic voters werereported in the precinct # 9 Democratic caucus. Charles Griego testified that when he ran for county commissioner in 1996 he overheard a group of Anglo men say, "We kicked those Mexicans' ass." Pedro Gomez testified that at a reception in 1990, Commissioner Robert Zimmerman told him that his people, "the white people," made Gomez' people, "the Mexican people," eat "humble pie."

[11] Some witnesses complained of the term "wetback" being used as an epithet. It is difficult to measure its discriminatory impact because the term has several meanings and is used in different contexts. According to Lewis Entz, in the San Luis Valley the term "wetback" refers to illegal Mexican immigrants rather than to local Hispanic residents. Other witnesses testified that the term is used most frequently within the Hispanic community and rarely used as a slur by Anglo residents.

[12] There was no evidence concerning the type of areas in which the remaining 11.1% of Hispanic homeowners lived. No comparable evidence was provided for Anglo homeowners.

[13] The parties have identified the ethnicity of candidates primarily by surname. As discussed infra in note 38, the Court has significant reservations about reliance on surnames to determine ethnicity. However, at least with respect to the candidates, the parties also had an opportunity to validate the ethnic label by the individual's self-identification.

[14] The only primary election lost by an Hispanic contender was in 1996 when Charles Griego lost to Charlotte Bobicki. Mr. Griego acknowledged that both candidates campaigned on the same issues, but that Ms. Bobicki was better known in the community. The evidence also suggests that Ms. Bobicki's campaign may have been more vigorous and broad-based.

[15] The following are results of county commissioner general elections involving Hispanic candidates since the 1970's. 1970's. Hispanic Pasqual Lee lost to Anglo incumbent William Wiescamp (1972). Hispanic Walt Gomez lost to Anglo incumbent Phillip Lorton (1978). 1980's. Hispanic Ricky Sanchez lost to Anglo Tim Gallagher(1980). Jorge Amaya won against appointed incumbent Abe Relyea (who had finished out the term of Lewis Entz), and Dennis Carpenter (1984). Hispanic Lee Patrick Herrera won against Guy Miles (1984). Walt Gomez was appointed to complete Mr. Herrera's term (1987) and withdrew from the 1998 general election. 1990's. Hispanic Marguerite Salazar lost to incumbent Tim Gallagher(1992). Since 2000. Hispanic Ernie Roybal lost to Hispanic Frank Mestas (2002).

[16] The following are results of elections involving Hispanic candidates since the 1970's. 1970's. Hispanic Larry Romero ran unsuccessfully for Coroner (1978). 1980's. Hispanic Larry Romero again ran unsuccessfully for Coroner (1982). Hispanic Lawrence Zamora ran unsuccessfully for Sheriff against an incumbent (1986). 1990's. Hispanic Lawrence Zamora once again ran unsuccessfully for Sheriff against an incumbent (1990). Hispanic Anna Mae Real-Lindsay ran unsuccessfully for Clerk and Recorder against incumbent Holly Lowder (1990). Hispanic Leonard Romero ran unsuccessfully for Coroner against an incumbent (1994). Write-in Hispanic candidate Moses Vargas ran unsuccessfully for Sheriff (1998).

[17] Reliance on Spanish surnames is less exact than census data based upon self-identification and candidate identification based on surname combined with self-identification.

[18] The following is a summary of the elections analyzed by Dr. Engstrom and his conclusions. 1970's. No elections were evaluated. 1980's. Endogenous Elections. Dr. Engstrom concluded that the two races in 1984 in which both Hispanic candidates won were polarized, though the candidate of choice for Hispanic voters was not vetoed. Exogenous Elections. Dr. Engstrom concluded that the 1982 coroner election involving Hispanic candidate Larry Romero, the 1986 sheriff race involving Hispanic candidate Lawrence Zamora, and the 1982 state representative election in House District 60 involving Hispanic candidate Alex Marquez were polarized. 1990's. Endogenous Elections. Dr. Engstrom examined only one general election, the 1992 county commissioner race involving candidate Marguerite Salazar, which he found to be polarized. He also found polarization in the 1996 primary election between Charlotte Bobicki and Charles Griego. Exogenous Elections. Dr. Engstrom concluded that the 1990 sheriff race involving Hispanic candidate Lawrence Zamora, the 1990 county clerk race involving Hispanic candidate Anna Mae Real-Lindsay, and the 1994 coroner election involving Hispanic candidate Leonard Romero were polarized. He also evaluated the 1992 district attorney election involving Hispanic candidate Castelar Garcia and found that it was not polarized, but that the 1992 House District 60 race involving Hispanic candidate Silver Jaramillo was polarized, as was the 1994 State Senate District 5 race involving Hispanic candidate John Ribal. Dr. Engstrom also found polarization in the 1998 House District 60 election involving Hispanic candidate Al Gagliardi. Dr. Engstrom did not examine Carlos Lucero's campaigns for the United States Senate, nor did he consider Ken Salazar's successful campaigns for Colorado Attorney General. Since 2000. Endogenous Elections. Dr. Engstrom evaluated the 2002 primary election between Ernest Roybal and Al Greer, but drew no conclusions. He also considered the 2002 general election between Ernest Roybal and Frank Mestas and found polarization because he believed that Mr. Roybal was the candidate of choice for Hispanic voters. Exogenous Elections. Dr. Engstrom found polarized the 2000 House District 60 general election involving candidate Al Gagliardi and the 2002 general election in new House District 62 in which candidate John Salazar won. He also found polarization in the 2002 Senate District 5 election involving candidate Rafael Gallegos.

[19] Dr. Engstrom identified 13 out of 14 exogenous and all eight endogenous races as ethnically polarized.

[20] The following is a summary of elections considered by Dr. Weber. 1970's. No elections were analyzed. 1980's. Endogenous Elections. Dr. Weber analyzed all primary and general elections during the 1980's except for the 1980 elections. Of these twelve elections, Dr. Weber identified polarization in a single general election — the 1984 Amaya general election — which was moderately polarized. Exogenous Elections. Dr. Weber found no polarization in two Democratic primaries in gubernatorial elections in 1982 and 1986. 1990's. Endogenous Elections. Dr. Weber analyzed all primary and general elections during the 1990's. Of these fourteen elections, he found polarization in only two involving Hispanic candidates — the 1996 primary between Mr. Griego and Ms. Bobicki and the 1992 general election involving Ms. Salazar. Exogenous Elections. Dr. Weber found no polarization in the 1992 district attorney race. Since 2000. Endogenous Elections. Dr. Weber analyzed all primary and general elections and found no polarization. Exogenous Elections. Dr. Weber found that the 2002 state house election in which John Salazar won was polarized.

[21] Hispanic residents belong to and are active in both the Democratic and Republican parties. The number of Hispanic residents in Alamosa County who identify themselves as somewhat or very liberal is similar to the number who identify themselves as somewhat or very conservative.

[22] In order to register, a county resident and U.S. citizen who is at least 18 years of age must complete a form which requires the registrant's legal address, mailing address, sex, party affiliation, and date of birth. Persons may register to vote at the county clerk's office, the driver's license office, the city hall, and at service based agencies or by mailing in forms obtainable at military installations, post offices and some schools.

[23] The registration form is available in Spanish at all locations except the driver's license office. Spanish-speaking employees are available to assist at the Clerk and Recorder's Office.

[24] With the exception of absentee ballots, all ballots are available in Spanish and English. There are bilingual election judges and signs at every polling place.

[25] In 1980, 86% of Hispanic voters participated in the District 3 commissioner election, and 87 to 89% of Hispanic voters cast votes in the District 60 state representative election. In 1982, 65 to 72% of Hispanic voters participated in the District 60 state representative election. In 1984, 71 to 74% of Hispanic voters voted in the District 3 commissioner election. In 1998, 44 to 54% of Hispanic voters cast votes in the District 60 state representative election.

[26] Walt Gomez, who ran unsuccessfully for the Board in 1978, observed that "every issue" the county commissioners resolve has an impact on Hispanic residents in Alamosa County.

[27] According to Walt Gomez, the issues of importance to Hispanic residents in 1978 were housing, health, and human services. Jorge Amaya testified that in 1984, the important issue for Hispanic residents was jobs. Marguerite Salazar testified that in 1992, the four issues of importance to Hispanic residents were welfare reform, recycling, jobs, and housing. Charles Griego testified that the current issues of importance to Hispanic residents are housing, upper-level jobs with the county, and bilingual assistance for older citizens.

[28] Jorge Amaya chose to "de-Latinize" his campaign in order to appeal to the largest group of voters. He had previously witnessed that Hispanic candidates who emphasized ethnicity were not successful. Lee Patrick Herrera recognized that it would be "political suicide" to emphasize ethnicity.

[29] Ms. Salazar testified that Anglo county commissioners did not represent the Hispanic community because "[t]o me, representative government means having somebody that looks like you, that understands what you do; and we didn't have that." She also testified that her family does not interact much with Anglo residents. Notably, her extended family includes several individuals who testified on behalf of the Government: her husband, Arnold Salazar; her sister-in-law, Lillian Gomez; her brother-in-law, Pedro Gomez; and her husband's first cousin, Henry Herrera.

[30] According to Ms. Real-Lindsay, she and others left the Democratic Central Committee and the party following a dispute with a group led by Mr. Gomez. The controversy concerned how much emphasis the Democratic party should give to the ethnic identification of its candidates.

[31] Section 2 of the Voting Rights Act, 42 U.S.C. § 1973, provides:

(a) No voting qualification or prerequisite to voting or standard, practice, or procedure shall be imposed or applied by any State or political subdivision in a manner which results in a denial or abridgement of the right of any citizen of the United States to vote on account of race or color, or in contravention of the guarantees set forth in section 1973b(f)(2) of this title, as provided in subsection (b) of this section.

(b) A violation of subsection (a) of this section is established if, based on the totality of circumstances, it is shown that the political processes leading to nomination or election in the State or political subdivision are not equally open to participation by members of a class of citizens protected by subsection (a) of this section in that its members have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice. The extent to which members of a protected class have been elected to office in the State or political subdivision is one circumstance which may be considered: Provided, That nothing in this section establishes a right to have members of a protected class elected in numbers equal to their proportion in the population.

[32] Applying this framework, the Tenth Circuit Court of Appeals considered Section 2 challenges to voting practices in the San Luis Valley on two occasions. In Sanchez v. Bond, 875 F.2d 1488 (10th Cir.1989), the at-large election of county commissioners in Saguache County was challenged. The circuit court affirmed the trial court's conclusion that there was no vote dilution upon findings that Saguache County Hispanic voters were not politically cohesive and that Anglo voters did not usually vote as a bloc to defeat candidates preferred by Hispanic voters. In Sanchez v. Colorado, 97 F.3d 1303 (10th Cir.1996), the circuit court considered the configuration of Colorado House District 60, which included Alamosa, Saguache, Conejos, Costilla, Mineral, Rio Grande, Las Animas, and Huerfano Counties and found impermissible vote dilution. No Hispanic candidate had been elected to the state house from the district in 56 years, Hispanic residents in Saguache County experienced voter registration problems, Hispanic individuals in the San Luis Valley tended to be more socioeconomically disadvantaged than Anglo individuals, and elections were ethnically polarized.

[33] S.REP. 97-417, 1982 U.S.C.C.A.N. 177 at 206-07.

[34] In Sanchez v. Colorado, Hispanic residents of voting age comprised 50.03% of the illustrative district.

[35] The Defendants prefer Dr. Weber's illustrative district which has an Hispanic share of the voting age population of 54.1%, but less than 50% of the registered voters. Arguably, even this district would satisfy precondition one.

[36] In Sanchez v. Colorado, the trial court mistakenly used multivariate analysis, concluding that there were a number of explanations for election outcomes. Reliance upon multivariate analysis, which addresses why voters cast votes for particular candidates, is reserved for consideration as part of the "totality of the circumstances."

[37] This method yields a hypothetical homogeneous precinct. It was used because there were no actual precincts composed of only Hispanic voters, or even precincts with an Hispanic majority. The error rate in this method is affected by its extrapolative nature (see note 38), the limited sample (16 precincts through 1993; 8 precincts thereafter), the fact that absentee ballots were not accounted for except with respect to the 2002 elections and the surname bias.

[38] Although the use of bivariate ecological regression analysis and the concomitant creation of a hypothetical homogeneous precinct are customary evidence considered with regard to this precondition, the persuasiveness of this method in a vote dilution case, particularly this case, seems doubtful. Other courts have expressed similar reservations. See Lewis v. Alamance County, North Carolina, 99 F.3d 600, 605 n. 3 (4th Cir.1996); Overton v. City of Austin, 871 F.2d 529, 539 n. 12 (5th Cir.1989).

The analysis is used to determine how minority voters voted. In reality, absent exit polls, no one really knows how any voter, much less how a group of voters, voted. Thus, the statistician makes an assumption — here, that Hispanic voters prefer Hispanic candidates. Unfortunately, the assumption guarantees the statistical result. Put simply, in this context, bivariate ecological regression analysis does not test a hypothesis; it is an exercise in circular reasoning.

For example: Assume a precinct has 100 voters, 60 Anglo and 40 Hispanic. There is an election involving two candidates, one Anglo and one Hispanic, where all registered voters vote. The Anglo candidate receives 60 votes and the Hispanic candidate receives 40 votes. From these facts one might surmise, but certainly cannot prove, that the Hispanic voters voted for the Hispanic candidate. It is equally possible that all, or half, or some number of the Hispanic voters crossed over and voted for the Anglo candidate and a corresponding number of Anglo voters crossed over to vote for the Hispanic candidate. The situation becomes even more confused if only 75 of the eligible voters voted. Which voters were Anglo; which were Hispanic? Perhaps all of the Hispanic voters voted, but only 35 of the Anglo voters did. In such event the comparative ratio is no longer 60/40 (Anglo/Hispanic), but instead 40/35 (Hispanic/Anglo). If the Anglo candidate wins with 60 percent of the vote, without knowing how many Hispanic voters voted (either in number or percentage) and how they voted, one cannot determine how many persons within each ethnic group supported each candidate. Sadly, the accuracy of the analysis is not improved by plotting more precincts because the analysis for each precinct is similarly flawed.

The bivariate ecological regression and homogeneous precinct analyses only operate in the context of the assumptions Dr. Engstrom and Dr. Weber made. First, they assumed that Hispanic voters preferred Hispanic candidates, then (unless there was a list of the names of the voters who voted in a given election) they reasoned that the percentage of voters within each ethnic group who actually voted matched the percentage of voters of that ethnic group who could have voted, that all voters of similar ethnic background voted the same way, or at least in a constant ratio, and in a hypothetical ethnically homogenous precinct, all voters would vote for the same candidate. The assumption that Hispanic voters prefer Hispanic candidates is the very fact which bivariate ecological regression analysis is designed to test. By making this assumption on the front end, the analysis invariably yields that conclusion — when an Hispanic candidate loses, the Hispanic candidate of choice has lost.

One might be able to compensate for the flaws in this statistical method by relying upon anecdotal evidence, but here the anecdotal evidence is mixed. Marguerite Salazar expressed the view that only an Hispanic candidate could adequately represent her or other Hispanic residents; thus presumably, she always prefers an Hispanic candidate. But not all Hispanic voters apparently feel the same way. A number of Hispanic witnesses actively supported Anglo candidates when they could have supported an Hispanic candidate, for example, Mr. Zimmerman (county commissioner — general) and Ms. Bobicki (county commissioner — primary).

[39] Elections involving only Anglo candidates can be relevant when analyzing whether there is ethnically polarized voting if it can be determined that one of the Anglo candidates was preferred by Hispanic voters. Sanchez v. Colorado, 97 F.3d at 1320. A court may consider several factors when determining whether an Anglo candidate is the candidate of choice for a minority group, including but not limited to: (1) whether members of the minority group sponsored the candidate; (2) whether members of the minority group contributed financially to the candidate; (3) whether the candidate focused on issues of interest to the minority group; (4) whether the candidate campaigned in predominantly minority neighborhoods or before minority audiences; (5) the degree of minority turnout for the election; and (6) the Anglo candidate's ties to the minority community. Id. at 1321. Such races might have been relevant, but the Defendants failed to lay the accessory foundation for consideration of Dr. Weber's opinions with regard to them.

Clearly, if all elections had been between Anglo candidates there could be a Section 2 violation if Hispanic candidates had been denied access to the process. Sanchez v. Colorado, 97 F.3d at 1321. But a mix of elections, some between Hispanic candidates, some between Anglo candidates, and some with both Hispanic and Anglo candidates suggests substantial political participation by Hispanic residents and voter preferences that extend beyond ethnic identity. Both minority and majority candidates must factor into their campaign strategies the practical knowledge that all voters, whether members of a minority or majority group, must seek out common political ground with others in order to advance a political agenda. See Johnson v. DeGrandy, 512 U.S. 997, 1020, 114 S.Ct. 2647, 129 L.Ed.2d 775 (1994) (cautioning that although Anglo "candidates may not represent perfection to every minority voter, ... minority voters are not immune from the obligation to pull, haul, and trade to find common political ground, the virtue of which is not to be slighted in applying a statute meant to hasten the waning of racism in American politics.").

[40] Gingles, 478 U.S. at 76, 106 S.Ct. 2752.

[41] Although not in evidence, examples which demonstrate this common-sense approach are John F. Kennedy (who de-emphasized his Catholic faith and Irish-American heritage), William J. Clinton (who minimized his underprivileged background growing up in Arkansas), and Colorado Attorney General Ken Salazar (who portrayed himself as an "nth" generation Coloradan and did not focus his campaign on ethnicity).

[42] The Government's suggestion that a Republican Hispanic candidate is not representative of Hispanic voters appears convenient and disingenuous. First, it contends that only Hispanic candidates can represent Hispanic voters. Then it argues that only Democratic candidates can do so. The evidence is to the contrary. Hispanic voters belong to both parties and support Anglo as well as Hispanic candidates.

[43] The only election data analyzed for primary elections was for Democratic primaries. The number of Hispanic voters in the party was not presented, thus the Court cannot determine whether Hispanic voters were the majority or minority, or whether Democratic primary elections are indicative of Hispanic voting preferences. The Court notes that the Democratic primary elections of Pasqual Lee in 1972, Walt Gomez in 1978 and 1988, and Marguerite Salazar in 1992 were unopposed. At least in those elections, there was no Anglo bloc vote.

[44] Senate factor 1 addresses whether there has been official discrimination against Hispanic individuals which affected their ability to vote, register to vote, or participate in the political process. See S.REP. 97-417, 1982 U.S.C.C.A.N. 177 at 206. Senate factor 3 concerns whether there are voting procedures that enhance the opportunity for discrimination against Hispanic individuals. Id. Senate factor 4 concerns whether members of the minority group were denied access to any candidate slating process. Id.

[45] Mary Sparrow purportedly advised Jorge Amaya not to run for county commissioner because he was "Mexican." He ignored her advice and won. Pedro Gomez had a personal dispute with Robert Zimmerman in which heated, ethnic comments were exchanged. However, Mr. Gomez declined to run, not because of such comments, but "because he felt that the Anglo majority would not support him." Charles Griego, who announced his desire to be nominated by the Democratic party in 1996, overheard Anglo Democratic delegates making crude ethnic remarks after a caucus. A controversy on the Democratic Central Committee in 1990 concerned a disagreement among Hispanic members as to whether Hispanic candidates should include ethnic appeals in their campaign strategies.

[46] See S.REP. 97-417, 1982 U.S.C.C.A.N. 177 at 206 n. 114 ("The courts have recognized that disproportionate educational, employment, income level and living conditions arising from past discrimination tend to depress minority political participation.").

[47] Henry Herrera.

[48] See Testimony of Lillian Gomez.

[49] See n. 8, supra.

[50] The Hispanic unemployment rate was 10.5%, as compared with a non-Hispanic unemployment rate of 6%.

[51] As of the 2000 Census, 17.9% of Hispanic residents and 31.0% of Anglo residents had incomes above $50,000.

[52] See Findings of Fact at ¶ III.B.

[53] See note 29, supra.

[54] It should be noted, however, that Walt Gomez and Jorge Amaya moved from Alamosa County 15 and 25 years ago, respectively. Accordingly, their testimony is limited to its historical context.

[55] All experts believe this differential is unusually high.

[56] Dr. Engstrom found that voter participation rates vary from election to election. In some elections, registered Hispanic voters turn out to a greater degree than registered, non-Hispanic voters. The contrary is also true in other elections.

[57] S.REP. 97-417, 1982 U.S.C.C.A.N. 177 at 207.

[58] Senate factor 2 asks whether elections tend to be ethnically polarized. See S.REP. 97-417, 1982 U.S.C.C.A.N. 177 at 206.

[59] Senate factor 7 asks whether minorities have been elected to office. S.REP. 97-417, 1982 U.S.C.C.A.N. 177 at 207.

[60] Senate factor 6 asks whether overt or subtle ethnic appeals were made by the candidates in their campaigns. See S.REP. 97-417, 1982 U.S.C.C.A.N. 177 at 206.

[61] As one commissioner observed, virtually every trip to the grocery store involves conversations about county business.