Earl OLD PERSON; Carol Juneau; Joe MacDonald; Jeannine Padilla, Plaintiffs-Appellants, v. Bob BROWN, Secretary of State for the State of Montana; Judy Martz, Governor of the State of Montana, Defendants-Appellees, 312 F.3d 1036

OPINION GOULD, Circuit Judge. We enter again the turbulent and at times turbid waters of voting rights litigation. Four American Indian plaintiffs appeal the district court’s judgment in favor of defendant officials of the State of Montana on the plaintiffs’ vote-dilution claim under

section 2 of the Voting Rights Act, 42 U.S.C. § 1973. The district court held that the plaintiffs had standing to allege Native American vote dilution only with respect to the House and Senate Districts where they resided. Also, the court held that the plaintiffs had not shown vote dilution in the House and Senate Districts where they resided. Finally and alternatively, the district court held that, even if the plaintiffs had shown vote dilution, their claim failed because of the state’s imminent redistricting and because any remedy would impermissibly disrupt the 2003 elections.

We have jurisdiction pursuant to 28 Ú.S.C. § 1291. Concluding that the district Court did not clearly err in determining that the totality of circumstances did not establish vote dilution in the districts where plaintiffs resided, we affirm on that ground. I The four Indian plaintiffs, Earl Old Person, Carol Juneau, Joe MacDonald, and Jeannine Padilla, are before us again in appeal of the district court’s rejection after bench trial of their vote-dilution challenge to Montana’s districting scheme under

section 2 of the Voting Rights Act. The plaintiffs live on the Flathead and Blackfeet Indian Reservations, which are located in the four-county area including Flathead, Lake, Glacier, and Pondera Counties in northwest Montana. . Defendants Bob Brown and Judy Martz are officials of the State of Montana, with responsibilities relating to elections. Montana is divided into House Districts and Senate Districts. Each Senate District is composed of two House Districts; there are 100 House Districts and 50 Senate Districts. One representative is elected from each district.

The current plan contains five majority-Indian House Districts and one majority Indian Senate District. State Senators serve four-year terms, whereas members of the state House of Representatives serve two-year terms. According to the 1990 census, on which Montana’s current districting plan was based, Native Americans comprise about 6% of the state’s population. The 2000 census indicates that the population of the state has increased but that the Native American population has increased in greater proportion than the white population. Montana will automatically redistrict under its laws in 2003.

In the first trial, the district court entered judgment for the defendants after a bench trial. The district court rejected the plaintiffs’ claims 1) that Montana’s 1992 redistricting plan (“the 1992 plan”) had been adopted for a discriminatory purpose and 2) that the 1992 plan impermissi-bly diluted the voting power of Native Americans. Old Person I, 230 F.3d at 1117. We affirmed the district court’s rejection of the first claim relating to discriminatory legislative motive but reversed the district court’s judgment on the second claim of vote dilution.

We held that the district court erred in relying on the electoral success of Indians in majority-Indian House Districts when evaluating the extent of white bloc voting, and in concluding that Native Americans were proportionally represented as a result of the 1992 plan. Id. We remanded, directing the district court to evaluate further whether vote dilution had occurred under the totality of the circumstances. On remand, the district court' heard additional evidence relating to the 1998 and 2000 elections. Only four out of the eleven plaintiffs who had been before us remained in the case.

These four five in House Districts (“HD”) 73 and 85, and Senate Districts (“SD”) 37 and 43. The district court held that the plaintiffs had limited standing and to prevail had' to show, but did not show, vote dilution in the specific legislative districts where they resided. Also, the district court held that even if the plaintiffs had shown vote dilution, their claim failed because of the unavailability of an adequate remedy. The district court reasoned that the state would be redistricting in 2003, and the 2002 elections (the last elections to be conducted under the 1992 plan) were fast approaching.

This appeal by the plaintiffs followed. II We set forth the standard of review in Old Person I: We review for clear error the district court’s findings of fact, including its ultimate finding whether, in the totality of circumstances, vote dilution exists in violation of § 2. Thornburg v. Gingles, 478 U.S. 30, 78-79, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986); Smith v. Salt River Project Agric. Improvement & Power Dist., 109 F.3d 586, 591(9th Cir.1997).

We retain the power, however, “ ‘to correct errors of law, including those that may infect a so-called mixed finding of law and fact, or a finding of fact that is predicated on a misunderstanding of the governing rule of law.’ ” Gingles, 478 U.S. at 79, 106 S.Ct. 2752; Salt River, 109 F.3d at 591. 230 F.3d at 1119.

The “clear error” standard generally requires that “[i]f the [trial court’s] account of the evidence is plausible in light of the record viewed in its entirety, the court of appeals may not reverse it even though convinced that had it been sitting as the trier of fact, it would have weighed the evidence differently.” Phoenix Eng’g & Supply Inc. v. Universal Elec. Co., 104 F.3d 1137, 1141 (9th Cir.1997) (quoting Anderson v. Bessemer City, 470 U.S. 564, 573-74, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985)).

In Gingles, the Supreme Court stressed that a finding of vote dilution should be reviewed under the clearly-erroneous test because the examination of a vote-dilution claim “is peculiarly dependent upon the facts of each case[ ] ... and requires an intensely local appraisal of the design and impact of the contested electoral mechanisms.” Gingles, 478 U.S. at 79, 106 S.Ct. 2752 (quotation marks and citations omitted).

The Supreme Court has also stated that a district court’s factual findings in a vote-dilution case represent “ ‘a blend of history and an intensely local appraisal of the design and impact of the ... [voting] district in the light of past and present reality, political and otherwise.’” Rogers v. Lodge, 458 U.S. 613, 622, 102 S.Ct. 3272, 73 L.Ed.2d 1012 (1982) (quoting White v. Regester, 412 U.S. 755, 769-70, 93 S.Ct. 2332, 37 L.Ed.2d 314 (1973)).

III Old Person I is the law of the case. “Under the ‘law of the case’ doctrine, a court is ordinarily precluded from reexamining an issue previously decided by the same court, or a higher court, in the same case.” Richardson v. United States, 841 F.2d 993, 996 (9th Cir.1988). The law of the case is a discretionary doctrine, which is founded upon the sound public policy that litigation must come to an end.

An appellate court cannot efficiently perform its duty to provide expeditious justice to all if a question once considered and decided by it were to be litigated anew in the same case upon any and every subsequent appeal. This doctrine also serves to maintain consistency. Jeffries v. Wood, 114 F.3d 1484, 1489 (9th Cir.1997) (en banc) (citations and quotation marks omitted).

The “law of the case” doctrine is subject to the following exceptions: “(1) the decision is clearly erroneous and its enforcement would work a manifest injustice, (2) intervening controlling authority makes reconsideration appropriate, or (3) substantially different evidence was adduced at a subsequent trial.” Id. (citations, quotation marks, and footnote omitted). Here, none of the three exceptions applies. We are bound by the opinion of the prior panel as the law of the case.

Also we have no discretion to depart from precedential aspects of our prior decision in Old Person I, under the general law-of-the circuit rule. See Ross Island Sand & Gravel v. Matson, 226 F.3d 1015, 1018 (9th Cir.2000) (“Since Newton remains the law of the circuit ... we are without authority to overrule its directives.”). IV The district court concluded that the plaintiffs lacked standing to allege dilution beyond the House and Senate Districts where they reside.' We need not reach that issue.

As our analysis makes clear, even if the plaintiffs had standing to allege vote dilution outside of their own legislative districts, that would not affect our holding that, under the totality of the circumstances, it was not clearly erroneous for the district court to determine that there was no vote dilution. V To evaluate the district court’s findings, we must address the elements of a vote-dilution claim and the type of evidence that is relevant in assessing it. In 1982, Congress amended

section 2 of the Voting Rights Act to eliminate any intent requirement with respect to vote-dilution claims.

Section 2 as amended is set forth in 42 U.S.C. § 1973: (

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, ... as provided in subsection (b). (

b) A violation of subsection (

a) 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) 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. 42 U.S.C. § 1973. In Thornburg v. Gingles, 478 U.S. 30, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986), the Supreme Court held that three preconditions must be satisfied before a plaintiff can proceed to a vote-dilution challenge: First, the minority group must be able to demonstrate that it is sufficiently large and geographically compact to constitute a majority in a single-member district.

If it is not, as would be the case in a substantially integrated district, the multimember form of the district cannot be responsible for minority voters’ inability to elect its candidates. Second, the minority group must be able to show that it is politically cohesive. If the minority group is not politically cohesive, it cannot be said that the selection of a multimember electoral structure thwarts distinctive minority group interests.

Third, the minority must be able to demonstrate that the white majority votes sufficiently as a bloc to enable it — in the absence of special circumstances, such as the minority candidate running unopposed-usually to defeat the minority’s preferred candidate. Gingles, 478 U.S. at 50-51, 106 S.Ct. 2752 (citations and footnotes omitted). Although Gingles involved a challenge to multi-member voting districts, the Court in Johnson v. De Grandy, 512 U.S. 997, 1006-07, 114 S.Ct. 2647, 129 L.Ed.2d 775 (1994), held that plaintiffs seeking to challenge a single-member districting scheme under

section 2 of the Voting Rights Act must meet the three threshold requirements laid out in Gingles. Also, De Gran-dy reasoned that, especially in challenges to single-member districts, the Gingles preconditions must not be viewed as sufficient alone to support a finding of dilution. Id. at 1012, 106 S.Ct. 2752. After a plaintiff has satisfied the Gingles preconditions, a court resolves a vote-dilution claim that a minority group has had less opportunity to participate in the political process, and to elect representatives of their choice, by evaluating the “totality of the circumstances,” as is required by the statute in

section 2(b). Such an evaluation, in turn, is to be guided by the factors identified in the Senate Judiciary Committee Majority Report, which accompanied the bill amending

section 2 in 1982.

These are: 1. the extent of any history of official discrimination in the state or political subdivision that touched the right of the members of the minority group to register, to vote, or otherwise to participate in the democratic process; 2. the extent to which voting in the elections of the state or political subdivision is racially polarized; 3. the extent to which the state or political subdivision has used unusually large election districts, majority vote requirements, anti-single shot provisions, or other voting practices or procedures that may enhance the opportunity for discrimination against the minority group; 4. if there is a candidate slating process, whether the members of the minority group have been denied access to that process; 5. the extent to which members of the minority group in the state or political subdivision bear the effects of discrimination in such areas as education, employment and health, which hinder their ability to participate effectively in the political process; 6. whether political campaigns have been characterized by overt or subtle racial appeals; 7. the extent to which members of the minority group have been elected to public office in the jurisdiction.

Additional factors that in some cases have had probative value as part of plaintiffs’ evidence to establish a violation are: whether there is a significant lack of responsiveness on the part of elected officials to the particularized needs of the members of the minority group. whether the policy underlying the state or political subdivision’s use of such voting qualification, pre-requisite to voting, or standard, practice or procedure is tenuous. S.Rep. No. 97-417, at 28-29 (1982), reprinted in 1982 U.S.C.C.A.N. 177, 206-07.

See also Gingles, 478 U.S. at 36-87 n. 4, 106 S.Ct. 2752 (citations and quotation marks omitted). De Grandy added “proportionality” to the factors included in the totality of circumstances. De Grandy, 512 U.S. at 1013-14, 114 S.Ct. 2647. The Supreme Court explained that “proportionality” looks to “the political or electoral power of minority voters” rather than to “the success of minority candidates.” Id. at 1014 n. 11, 114 S.Ct. 2647. This established law guides our review of the district court’s finding that vote dilution did not occur based on the totality of the circumstances.

VI A Relying on the “law of the case” doctrine, the district court readily adopted Old Person I’s conclusion that the Gingles preconditions had been satisfied. The district court concluded that a contrary determination was not necessary to avoid a manifest injustice. We agree. The determination that all three threshold preconditions in Gingles were met is not clearly erroneous. We next consider the factors set out in Gingles and De Grandy as having probative value in evaluating a vote-dilution claim to the extent they were generally not contested. We defer to a later

section our discussion of the more controversial issues. In Old Person I, we said that “[t]here was a history of discrimination by the federal government and the State of Montana from the 1860s until as recently as 1971. American Indians have a lower socio-eco-nomic status than whites in Montana; these social and economic factors hinder the ability of American Indians in Montana to participate fully in the political process.” Old Person I, 230 F.3d at 1129.

We also said that elections in some of the districts at issue were characterized by racial polarization, and that there had been overt or subtle racial appeals in two recent elections. Id. The district court adopted and gave weight to these conclusions, all of which are in the plaintiffs’ favor. We consider each in turn.

The defendants argued that recent events showed progress in reducing discrimination against Native Americans in Montana, but the district court properly held that an inquiry into the history of discrimination against a minority group “focuses on a prior history of discrimination” and that here, our holding “constitutes the law of the case.” The district court accepted our determination that some elections in the districts at issue were racially polarized, and the court found that the evidence adduced on remand also supported this finding.

The district court relied on both the plaintiffs’ and defendants’ statistical experts in reaching this conclusion. Nothing in the record shows that the district court’s conclusion about racial polarization was clearly erroneous. There was no new evidence concerning Native Americans’ socio-economic status in Montana, and we had previously upheld the district court’s finding that American Indians have a lower socio-economic status than whites in Montana, hindering the ability of Indians there to participate fully in the political process. Old Person I, 230 F.3d at 1129.

The district court noted that no evidence had been presented showing that a particular change of district lines would cause Indians to “fare any better.” The absence of a proposed remedy is not a full answer to analysis of this factor. Socio-economic status hindering election participation is merely one factor in a total circumstances analysis of whether vote dilution has occurred. Whether or not new boundaries are suggested, this factor weighs in favor of a finding of vote dilution.

The district court’s reasoning shows tacit recognition of this, along with the not unsurprising observation that a redrawing of district lines would not fundamentally alter socio-economic status of Indians in Montana. Similarly, there was no new evidence concerning the use of racial appeals in elections. We had previously upheld the trial court’s finding that there were racial appeals in at least two recent elections in Lake County, which is within the plaintiffs’ legislative districts. See Old Person I, 230 F.3d at 1129. This factor, too, weighs in favor of a finding of vote dilution.

The district court also considered factors that weigh against a finding of vote dilution. The district court noted that our prior holding upheld the trial court’s finding that Montana had not used voting practices that may enhance discrimination, id., and that the parties had presented no new evidence on this factor. The district court recited that, as we noted in Old Person I, Montana has no slating process, and again, there was no contrary evidence; we must conclude that Native Americans have not been denied access to such a process. Old Person I, 230 F.3d at 1129.

The district court adhered to its prior conclusion, also previously affirmed by us, that Montana officials are “generally responsive” to the heeds of Native American voters. Again, the district court said that no evidence introduced on remand would lead it to change the prior finding of responsiveness. But the district court, echoing our prior views, thought this of “limited relevance” under a result-based (rather than intent-based) test.' Old Person I, 230 F.3d at 1129 n. 14.

The district court also found, in accord with our prior determination upholding the district court’s prior finding, that the policy underlying the creation of the existing district boundaries was not tenuous. On this point, again, nothing new was presented. All these findings weigh against vote dilution. There was no error in any of these findings, for all were the law of the ease, and none fell within the limited exceptions to that doctrine, particularly absent supervening evidence to the contrary. B We next consider the factor of “proportionality,” which De Grandy taught was relevant.

This issue generated some controversy in the district court, and because of unsettled law, it poses challenges in analysis. As the Supreme Court has explained, “proportionality,” a word not expressly used in the operative language of

Section 2, involves a comparison between 1) the percentage share of legislative districts in which the population of the protected class has a majority and 2) the protected class’s percentage share of the “relevant population.” De Grandy, 512 U.S. at 1014 n. 11. The district court, doubtless affected by its ruling on standing, limited the frame of reference for “proportionality” to the legislative districts where the plaintiffs reside, and it looked at the number of Indian-preferred candidates who had been elected in those and adjacent districts.

The district court did not evaluate the share of districts in which the protected class of Indians constitutes a majority, as contrasted with the Indians’ overall share of population in the broader relevant area. Though the Supreme Court in De Gran-dy and we in Old Person I were not crystal clear in explaining how to assess the relevant geographic area for assessing “proportionality,” and the district court’s reasoning was understandable, we now conclude that the district court’s method of assessing proportionality was not sufficiently precise and we seek to clarify our law on the calculation of proportionality.

The Supreme Court has suggested that proportionality is' an aid to evaluate “equality of opportunity, not a guarantee of electoral success for minority-preferred candidates.” Id. at 1014 n. 11, 114 S.Ct. 2647. We have not determined what geographical frame of reference should be used to analyze proportionality, whether it is by district, county, or state, and the Supreme Court has not resolved this issue. In view of our disposition, we do not determine all pertinent issues today.

However, we consider whether the frame of reference for proportionality may be limited to the legislative districts where the plaintiffs reside. While in this case it might favor the state to limit the frame of reference to the House and Senate Districts where the plaintiffs reside, it might do the opposite and promote manipulative litigation if we were to adopt a rule that limits the proportionality analysis in this manner.

To look at proportionality only for the districts where plaintiffs reside would encourage litigation with selection of particular plaintiffs living in districts where the statewide minority does not have a majority. If success in those cases led to redistricting to enhance minority voting power, there would be a resulting injustice to other voters because the districting scheme would have minimized consideration of districts where the statewide minority had effective majority control.

As the De Grandy Court explained, it would be absurd to suggest that the failure of a districting scheme to provide a minority group with effective political power [significantly] above its numerical strength indicates a denial of equal participation in the political process. Failure to maximize cannot be the measure of § 2. De Grandy, 512 U.S. at 1017, 114 S.Ct. 2647 (footnote omitted).

As the frame of reference expands from one legislative district or a set of legislative districts to a county or a set of counties or to the entire state, it becomes correspondingly more difficult to select plaintiffs in clusters that misrepresent their group’s share of the overall population. We conclude that limiting the frame of reference to the plaintiffs’ legislative districts would allow for an inaccurate proportionality calculus that may interfere with the goals of the Voting Rights Act.

To hold otherwise would be to promote a result that the Sixth Circuit has correctly deemed to be impermissible under- the Voting Rights Act — namely, to “enable future litigants to carvé up successively smaller areas of the State until they are able to maximize the number of majority-minority legislative districts.” Rural W. Tenn. African-Am. Affairs Council v. Sundquist, 209 F.3d 835, 844 (6th Cir.2000) (citing De Grandy, 512 U.S. at 1017, 114 S.Ct. 2647). As Rural West shows, it does not follow that a statewide frame of reference is always preferable to one that is limited to a county or several counties.

However, we do not need to resolve the question of when it is appropriate to focus on a county rather than the entire state. It is enough to conclude that a frame of reference that is limited to certain legislative districts is impermissibly narrow. Here, the plaintiffs argue that proportionality must be determined either by looking to the entire state or to the four-county region that includes the Flathead and Blackfeet Indian Reservations.

Our holding that proportionality analysis could not here be limited to the districts of the plaintiffs’ residence does not require us to choose between the state or the four counties as frame of reference. In either case, there is a lack of proportionality.

The district court, focusing on the four legislative districts where the plaintiffs reside, concluded that because “Indian-preferred candidates have been elected to the Montana legislature from both of the House Districts at issue and from one of the two Senate Districts at issue in this case, ... the proportionality factor is satisfied.” The district court should have used a broader frame of reference, and instead of considering the number of Indian-preferred candidates who had been elected, the district court should have considered the Indians’ share of majority-minority districts.

Because that share is not proportional under either a four-county or a statewide frame of reference, the proportionality factor weighs in favor of a finding of vote dilution. C The district court also considered, and gave considerable weight to, the extent to which members of the minority group have been elected to public office. The district court found that Indians make up 13.49% of the population in the four-county area that includes the Flathead and Blackfeet Indian Reservations, and that 2 out of 20, or 10%, of the legislators representing those districts are Indian.

The district court concluded that “proportional representation of American Indians exists.” While we do not look at “proportionality” in this sense, it is clear that the factors to be considered under

section 2 in assessing vote dilution include the success of Indian or Indian-preferred candidates in elections. While the district court looked solely to the plaintiffs’ own legislative districts to analyze “proportionality” in the De Grandy sense, the court looked more broadly to the four-county area to determine representation in public office by Indian candidates.

Both proportionality and success in elections are to be considered as part of the “totality of the circumstances,” and both help us to understand the ability of members of the minority to participate in the political process and to elect representatives of their choice. It necessarily follows that the same frame of reference should be used for both inquiries. The Supreme Court has not expressly announced such a rule, and neither have we, in our prior decisions.

However, courts that have analyzed “proportionality” in the De Grandy sense have been consistent in using the same frame of reference for that factor and for the factors set forth in Gingles.

See, e.g., De Grandy, 512 U.S. at 1022, 114 S.Ct. 2647(analyzing proportionality only for Dade County, and not statewide, where “the plaintiffs ... passed up the opportunity to frame their dilution claim in statewide terms” and instead applied the Gingles factors to “Hispanics in the Dade County area”); Rural West, 209 F.3d at 840-41 (using the same six-county frame of reference to examine proportionality and all of the Gingles factors); Solomon v. Liberty County Comm’rs, 221 F.3d 1218, 1220 (11th Cir.2000) (using the same county as a frame of reference in analyzing all factors); African-Am.

Voting Rights Legal Defense Fund, Inc. v. Villa, 54 F.3d 1345, 1354-55 (8th Cir.1995) (using all 28 wards of the city of St. Louis as the frame of reference for analyzing both proportionality and the Gingles preconditions). In Old Person I, we used a statewide frame of reference to consider success in elections, and we concluded that “the number of American Indians elected to the state legislature was ‘roughly proportional’ to the Annerican Indians’ share of the voting age population in Montana.” Old Person I, 230 F.3d at 1129-30.

In the first trial, the district court made extensive findings on success of Indian-preferred candidates, which findings were upheld on appeal in Old Person I. On remand, the district court adhered to its prior conclusions about Indians’ success in elections, and also further considered evidence concerning the election of a Native American candidate to HD 73, discussed below. There was no error in the district court’s conclusion; however, as noted above, the court’s evidence of election success was drawn from the four-county area, not from the entire state.

This was appropriate .if the same four-county frame of reference is used in the proportionality analysis. As set forth above, the district court reasoned that percentages of election success by Indian-preferred candidates bore some proportion to Indian numbers in the four counties. Neither we nor the Supreme Court has held that the same degree of correlation required by De Grandy in analyzing proportionality is also required in measuring the extent to which members of the minority group have been elected to public office.

But while these two considerations may not involve exactly the same standard, a comparison between share of population and share of elected officers'is an obvious and helpful measure of success. It was not clear error for the district court to conclude that where Indians constituted 13.49% of the relevant population and held 10% of the elected offices in the four-county area, the factor of success in elections favored defendants. Indeed, we do not read

section 2 of the Voting Rights Act to require any precise mathematical calculation of percentages of success in election. Such calculations may be misleading or give undue weight to what is only one of many factors listed in Gingles. What is important is that the district court consider the success in elections of the protected class. Neither the statute, nor the Senate report pertinent to its

interpretation, nor the Supreme Court’s teachings in Gingles or De Grandy require any particular form for the district court’s assessment of election success. In making its finding about election success, the district court stressed evidence concerning the election of Joey Jayne, a tribal member and an Indian-preferred candidate, in HD 73 in the 2000 election. Plaintiffs argue that this evidence has little weight because Jayne ran against a third-party candidate who came close to winning. But elections are not horseshoes, and having a close win is not the same as losing an election.

The factor designated by the Gingles Court, success in elections, favors defendants when the protected class has had success. That determination is not the sole basis for evaluating the minority group’s political power; it is one of the factors that make up the totality of the circumstances. The district court properly concluded that while Jayne’s election does not show that there is no bloc voting in the four districts at issue, her election could be considered in determining the extent to which minority candidates have been elected to office.

D The district court found that factors indicating vote dilution included the history of official discrimination in the state; racial polarization of elections; socio-economic factors hindering political participation; and use of racial appeals in elections.

Factors weighing in the other direction included the absence of voting practices that would enhance discrimination; absence of denial of access to a slating process; state officials’ responsiveness to the needs of Indian voters; - the absence of a tenuous policy underlying the creation of the existing district boundaries; and the election of Native Americans to public office. Lack of proportionality, which the district court included on the second list, in our analysis should have been added to the factors weighing in favor of vote dilution.

After weighing all pertinent factors, the district court concluded that the plaintiffs had not met their burden of establishing vote dilution. The district court’s analysis of proportionality was erroneous, as we see it, because the proportionality analysis had to be made for a broader area than the districts where the plaintiffs reside. Of course, that analysis would only be one factor in the overall assessment, based on total circumstances, of whether there was vote dilution in the districts where plaintiffs resided. The court’s assessment of success in elections by Indian candidates was not clearly erroneous.

We must face the ultimate question as to whether our disagreement on proportionality analysis requires reversal. It is not an easy task for a district court to assess vote dilution. The language of the statute speaks in terms of equal opportunity for protected minority classes to participate in the electoral process. The consideration of a

section 2(

b) vote dilution claim by the statute’s terms is to be made based on the “totality of circumstances.” The Supreme Court in Gingles set forth a long line of factors relevant to the assessment, based on the Senate’s report underlying the 1982 amendments. Initially, as explained in Gingles, this did not include proportionality, but the Court in De Grcmdy added proportionality as a factor, though without defining its contours precisely. The standards that we today derive from our assessment of De Grandy and other relevant considerations are nowhere set forth in explicit statutory language.

Surely, it would be helpful if Congress or the Supreme Court were to revisit the thorny issue of vote dilution, but absent clarification, district courts and we must proceed as carefully as possible based on settled grounds. There are two things that are absolutely clear in the Congress’s directive of

section 2 of the Voting Rights Act and the Supreme Court’s guidance on it. One is that an assessment of vote dilution, an insidious form of indirect discrimination, must be made based on the “totality of circumstances.” Whatever factors are to be considered, and indeed they are many and varied as explained by the Supreme Court, none is talismanic, none alone has controlling weight, none provides safe harbor, and none yields per se violation.

The other point that has been made perfectly clear by the Supreme Court is that in this unusual sphere of law, where legal' principle may necessarily interact with the state or local legislature’s sense of political practicality, the ultimate legal question whether vote dilution has occurred is to be treated as factual, and reviewed not de novo but rather for clear error.

The Supreme Court’s direction that we apply this standard, usually reserved for factual decisions based on credibility of witnesses and the like, to a question with at least mixed legal and factual aspects, must surely reflect the Supreme Court’s decision that the congressional scheme will work best if the inherently factual issues of vote dilution are resolved by district courts close to the scene, and that the district court’s decisions be given particular deference. This is one of those not unusual cases where our decision is controlled by the proper standard of review.

On one side of the scale lies a history of official discrimination, the presence of racially polarized elections, the presence of socio-economic factors limiting Indians’ political participation, the use of racial appeals in elections, and disproportionality. On the other side of the scale we see the absence of discriminatory voting practices, the viable policy underlying the existing district boundaries, the success of Indians in elections, and officials’ responsiveness to Native Americans’ needs.

We have fully considered the legal issues presented and the detailed factual record with which the district court grappled. We cannot say that the district court’s determination that there was no vote dilution, considered in the totality of circumstances, was clearly erroneous. YII The district court erred in alternatively holding that the plaintiffs’ claims failed, even if they showed vote dilution, because of the district court’s view that there was no adequate remedy for the alleged Voting Rights Act violation.

We conclude that the integrity of the Voting Rights Act requires that a substantive assessment be made, even if the only practical relief available is declaratory relief after an election is held. The plaintiffs submitted several proposed redistricting plans. The defendants dispute the adequacy of the plaintiffs’ plans, but the district court did not rule on the plaintiffs’ proposal, instead holding that the proximity of the next election cycle and of the state’s own redistricting precluded any remedy.

The Eleventh Circuit has derived a remedy requirement from the Supreme Court’s Gingles opinion and other Supreme Court precedent. See Negron v. City of Miami Beach, 113 F.3d 1563, 1571 (11th Cir.1997) (en banc). As the court explained in Nipper v. Smith, 39 F.3d 1494 (11th Cir.1994): The absence of an available remedy is not only relevant at the remedial stage of the litigation, but also precludes, under the totality of the circumstances inquiry, a finding of liability.

As we note above, a court cannot determine whether the voting strength of a minority group has been impermissibly diluted without having some alternative electoral structure in mind for comparison. Thus, where there is no objective and workable standard for choosing a reasonable benchmark by which to evaluate a challenged voting practice, it follows that the voting practice cannot be challenged as dilutive under § 2. Id. at 1533 (citations and quotation marks omitted).

Even if we were to adopt the Eleventh Circuit’s view of the remedy requirement, the district court appears to have improperly extended it by focusing, not on whether the plaintiffs provided an adequate redistricting plan, but on how disruptive redistricting would be. We think it correct, as the plaintiffs point out, that redistricting could not be accomplished before the next pending elections. We also reject as unfeasible the plaintiffs’ contention that the terms of the next state legislators could be shortened to allow another election without vote dilution.

In-junctive relief of such a nature would be extremely disruptive to states and inconsistent with the relations between the federal government and the states. But to say that the plaintiffs should lose if they have an otherwise valid claim because of disruption would subject them to an impossible burden, and one that would defeat the purpose of the Voting Rights Act. If a scheme that controls elections could escape scrutiny on the ground that it cannot be corrected in time, the effect would be to ignore a problem that is “capable of repetition, yet evading review.” Moore v.

Ogilvie, 394 U.S. 814, 816, 89 S.Ct. 1493, 23 L.Ed.2d 1 (1969) (quoting Southern Pac. Terminal Co. v. Interstate Commerce Comm’n, 219 U.S. 498, 515, 31 S.Ct. 279, 55 L.Ed. 310 (1911)). Although there will be no further proceedings in this case because of our affir-mance of the district court’s conclusion that there is no Voting Rights Act violation, had there been a finding of violation the district court could have given declaratory relief explaining the nature of the violation and how it might be avoided in the future.

Because the state has plans to engage in redistricting on a regular basis, any problem could then have been corrected by the future redistricting and a ruling from the district court on the merits may aid those charged with redistricting duties. VIII We affirm the district court’s ruling dismissing the Voting Rights Act claim based on the conclusion that there was no vote dilution. . AFFIRMED. . This case is before us for review for the second time. Old Person v. Cooney, 230 F.3d 1113 (9th Cir.2000) ("Old Person I"). . In Thornburg v.

Gingles, 478 U.S. 30, 47, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986), the Supreme Court described a vote-dilution claim after the 1982 amendments: "The essence of a § 2 claim is that a certain electoral law, practice, or structure interacts with social and historical conditions to cause an inequality in the opportunities enjoyed by black and white voters to elect their preferred representatives.” .

The De Grandy Court also explained the difficulties in proving vote dilution based on single-member districts: Plaintiffs challenging single-member districts may claim, not total submergence, but partial submergence; not the chance for some electoral success in place of none, but the chance for more success in place of some. When the question thus comes down to the reasonableness of drawing a series of district lines in one combination of places rather than another, judgments about inequality may become closer calls.

As facts beyond the ambit of the three Gin-gles factors loom correspondingly larger, factfinders cannot rest uncritically on assumptions about the force of the Gingles factors in pointing to dilution. De Grandy, 512 U.S. at 1012-13, 114 S.Ct. 2647. . Normally the protected class is one that is a minority in a relevant geographic area broader than a particular legislative district. .

It must be kept in mind that “proportionality” in this sense is only one factor to be considered in assessing the totality of circumstances to determine if unlawful vote dilution has created an "unequal political and electoral opportunity” for a protected class. De Grandy, 512 U.S. at 1022, 114 S.Ct. 2647. Just as the Supreme Court in De Grandy made clear that “proportionality” of opportunity cannot be a “safe harbor” precluding

Section 2 liability which turns on total circumstances, id. at 1018-22, 114 S.Ct. 2647, so too a showing of lack of proportionality is but one factor in the total circumstances analysis. It could not be correct to treat proportionality as but one factor if exculpatory while viewing it as a per se violation if inculpatory. In either case, the Supreme Court has made clear that "the degree of probative value assigned to proportionality may vary with other facts.

No single statistic provides courts with a shortcut to determine whether a set of single-member districts unlawfully dilutes minority voting strength.” Id. at 1020-21, 114 S.Ct. 2647. We also address the significance of that factor in the overall assessment of vote dilution based on totality of circumstances, when we address infra whether the finding of no vote dilution was clearly erroneous. .If the frame of reference were limited in this manner, it would include only four legislative districts. One of those districts — HD 85 — is majority-Indian.

Thus the share of legislative districts in which the population of the protected class has a majority is 25%. Old Person I shows that Indians constitute approximately 34% of the population in all four districts. See Old Person I, 230 F.3d at 1119. Within the limits of a four-district frame of reference, this correlation might indicate proportionality under De Grandy. .Consider the case of a minority group that makes up 21% of the population in a county with ten districts, with 100 people in each district.

If the minority group has a 54% majority in two districts, then 20% of the districts in the county are "majority-minority” (by this we mean a district in which the minority group has a numerical majority), and there is proportionality under De Grandy. If the minority group also constitutes 25% of the population in another four districts in the county, and plaintiffs alleged dilution there, they might show a lack of proportionality by limiting the focus to those four districts, which might be a factor in causing those districts to be redrawn to create another majority-minority district.

If that occurred, the minority group would have a majority in 30% of the districts in the county, while representing only 21% of the population. Those seeking such an outcome might select plaintiffs in a manipulative way to limit the frame of reference and distort the court's view of proportionality. . In limiting its focus to the legislative districts where the plaintiffs live, the district court also relied in part on Rural West Tennessee African-American Affairs Council v. Sundquist, 209 F.3d 835 (6th Cir.2000), which we had cited in Old Person I. See Old Person I, 230 F.3d at 1129 n. 15.

In Rural West, the plaintiffs alleged vote dilution in six rural counties in western Tennessee, but Tennessee argued that the entire state was the proper frame of reference. The court limited its focus to the counties where the plaintiffs lived, for "neither over-proportionality in one area of the State nor substantial proportionality in the State as a whole should ordinarily be used to offset a problem of vote dilution in one discrete area of the state.” Rural West, 209 F.3d at 843.

The court concluded that because the presence or absence of statewide proportionality could not determine the vote-dilution claim of plaintiffs in a particular locality, a more limited frame of reference was appropriate. The Rural West court refused to use an over-broad frame of reference that could conceal evidence of vote dilution in a specific area, but also did not use an overly narrow frame of reference that could exaggerate such evidence. .

In Old Person I, we held that the factor of proportionality weighed in favor of a finding of vote dilution if a statewide frame of reference was used, because Native Americans constituted 4.8% of the voting age population in Montana and 6% of the population in the state, but they had majorities in 2% of the Senate Districts and 4% of the total number of House and Senate Districts. Old Person I, 230 F.3d at 1129-30. The district court in the first trial considered the entire state in its proportionality analysis.

On review, we noted: Finally, with respect to proportionality, the district court found that the number of legislative districts in which American Indians constitute an effective majority is "roughly proportional to the American Indians’ respective share of the voting age population in Montana.” ' Old Person I, 230 F.3d at 1129.

We found error in the proportionality assessment because, although Indians constituted 4.8% of the voting-age population of Montana, and 6% of the state’s general population, only 2% of the Senate Districts were majority Indian) and only 4% of the combined House and Senate Districts were majority-Indian. Id. at 1129-30.

On remand, the district court found that "statewide, the gap between the number of majority-minority districts to minority members’ share of the relevant population has increased in Montana [since the decision in Old Person /].” A focus on the four counties that include the Flathead and Blackfeet Indian Reservations would not change this result. The district court found that the Indian population in those four counties is 13.49%.

Among the 13 House Districts and 7 Senate Districts that include some portion of those four counties, there is one majority-Indian House District (HD 85) and there are no majority-Indian Senate Districts. Thus 7.69% of the House Districts are majority-Indian, 0% of the Senate Districts are majority-Indian, and 5% of the total number of legislative districts are majority-Indian. We calculate proportionality by dividing the percentage of majority-minority legislative districts by the minority's percentage share of the population. See De Grandy, 512 U.S. at 1023 n. 19, 114 S.Ct. 2647.

Here, proportionality is 55.9%, which we deem not reasonably proportional. Both a four-county and a statewide frame of reference would show a lack of proportionality. . The parties disagree about who can be counted as an "Indian-preferred” candidate. That question arose because in considering proportionality in the four legislative districts where the plaintiffs'reside, the district court included a candidate that it regarded as "Indian-preferred,” although the plaintiffs disputed that characterization.

Since we conclude that the district court used the wrong frame of reference and should have considered the number of majority-Indian districts, not the number of Indian-preferred candidates who were elected, we do not need to make a determination about which candidates are minority-preferred. . The district court’s observation about "proportional” representation by those in public office, a fair share of persons elected favored by the Indians, is of course an assessment different from the “proportionality” inquiry mandated by De Grandy.

We also note that the statute’s proviso expressly recites that there is no right to proportional representation: "[N]othing in this

section establishes a right to have members of a protected class elected in numbers equal to their proportion in the population.” 42 U.S.C. § 1973(b). . The Supreme Court has required consideration of “the extent to which members of the minority group have been elected to public office in the jurisdiction.” Gingles, 478 U.S. at 37, 106 S.Ct. 2752 (emphasis added). Taken literally, this would allow for consideration only of minority candidates, and not of non-minority candidates who are preferred by the relevant minority.

We believe such a view is unduly restrictive and does not reflect the realities of political behavior. That view does not, however, require us to resolve the dispute about who counts as an "Indian-preferred” candidate, see note 10 supra, because the same result would obtain here regardless of how that dispute is resolved. In finding that Indian voters had election success, the district court considered only Indian candidates, and not Indian-preferred candidates who were not Indian. Had the district court also considered the latter category, that would only have added further support for the court's conclusion. .

Gingles noted that in some cases, the election of a minority candidate does not negate a finding that there is bloc voting by a majority, because the candidate’s success can be explained by “special circumstances, such as the absence of an opponent, incumbency, or the utilization of bullet voting.” Gingles, 478 U.S. at 57, 106 S.Ct. 2752. Here, it can be argued that Jayne's election was a "special circumstance” entitled to little weight. Jayne was elected by a margin of 50.7% against a third-party opponent who received 49.3% of the vote.

But a theory that Jayne's election should be disregarded as “special circumstances” is speculative and we cannot substitute our views for those of the district court. Even if a "special circumstance” has relevance to discount bloc voting, that is not controlling on how many minority candidates have been elected. That Jayne owes her success, in part, to running against a third-party candidate does not mean that her election should be ignored when determining minority candidates' success in elections.

This factor falls within the fact-finding province of the district court charged with making the final finding on vote dilution. And of course success of Indian candidates is not a safe harbor or total defense, but again only one factor in a total circumstances analysis. See De Grandy, 512 U.S. at 1019-22, 114 S.Ct. 2647. . It is perhaps not wholly insignificant that for decades Supreme Court precedent held that the question of legislative reapportionment based on equal protection challenges was not justiciable, a principle overruled in the landmark case of Baker v.

Carr, 369 U.S. 186, 82 S.Ct. 691, 7 L.Ed.2d 663 (1961), which permitted an equal protection challenge to legislative districts. Of course, in the Voting Rights Act Congress has laid squarely at the door of the federal courts the problem of assessing legislative districting and other election practices that may dilute rights of voters, particularly those of minority voters who rightly should not be shut out of the American electoral process. It is relatively easy to assess the issues and find judicial resolution in cases of intentional discriminatory practices.

It is more of a challenge, and quite difficult, to assess issues of indirect impact of facially neutral policies that are urged to dilute voters' power at the polls. . The clearly erroneous standard requires us to view the evidence in its entirety, and does not allow us to reverse the district court simply because we would have weighed the evidence differently if we had been sitting as trier of fact. Phoenix Eng'g & Supply Inc. v. Universal Elec. Co., 104 F.3d 1137, 1141 (9th Cir.1997) (quoting Anderson v. Bessemer City, 470 U.S. 564, 573-74, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985)). .

The Senate Report accompanying the 1982 amendment to 42 U.S.C. § 1973 notes that the purpose of the Voting Rights Act was "not only to correct an active history of discrimination, ... but also to deal with the accumulation of discrimination.... The bill would attempt to do something about accumulated wrongs and the continuance of the wrongs.” S. Rep. No. 97-417, at 5 (1982), reprinted in 1982 U.S.C.C.A.N. at 182 (quoting 111 Cong. Rec. 8,295 (1965) (Statement of Sen. Jarvis)).